Problem creating oblong sculpties

Chosen Few · December 11, 2011

Josh Susanto wrote:

Whatever the problem was, it's apparently now gone. I reinstalled blender, and the buttons actually seem to work now. Thanks for persuading me to do that.

Glad to hear it's working. Now it's time to learn to use it!

The Machinimatrix Blender Trail is fantastic place to start, so you know.

I think we can dismiss a lot of the points we were going back and forth on now, since they're no longer relevant. I'll skip to the ones that have to do with Blender not working.

Josh Susanto wrote:

So you advocate use of free mesh tools you don't use over free sculpt tools you also don't use on the basis of your extensive experience with the costly mesh tools you actually do use? Maybe people should also ask the Pope for gynecology referrals?

Two things:

First, I don't need to be an active user of Blender to understand what its capabilities are. The things it does are absolutely standard. They're the same things that every other 3D modeling platform in its class does. While the exact step by step work flows obviously vary from program to program, the principles are universal across all. The end results are exactly the same, no matter what the specific program.

Second, this was never about specific tools. It was about the medium in general. Trying to pretend otherwise at this late juncture is just plain silly.

Since you brought it up, though, I might invite you to consider that the most popular sculpty tool in use today is Blender. So there's really no logic at all in saying "free mesh tools over free sculpt tools". They're one in the same.

The tool I use, Maya, is what sculpties were invented on in the first place, and it remains a very popular choice as well, for both sculpties and arbitrary mesehs.

The fact that there happen to be programs that can only do one or the other, and that you happen to be using one of them, is circumstantial, and not at all relevant to the real topics we've been discussing.

Josh Susanto wrote:

I'm not seeing any need to install or run Python separately. I assume it's installed, but how would I know if it wasn't?

Start Blender, and take a look at the console window. If you have the right version of Phython installed, it will say "Checking for installed Python... got it!" If Python is not present, or if you have the wrong version installed, it will say "Checking for installed Python... No installed Python found."

Josh Susanto wrote:

Every minute spent trying to debug something that can only be debugged from the other end is a minute I can't spend building (or having some kind of potentially productive argument on this forum).

And every instance in which you don't try to overcome temporary obstacles is a lifetime of income you miss out on.

Small step back, giant leap forward. It's a cliche because it's true.

Thomas Edison was once quoted in saying he had to find 10,000 ways a light bulb doesn't work in order to find the one way it does work. He also had to install the thing free of charge in an office building before anyone would even look at it. Imagine if he'd just said, "every minute I spend testing a filament that doesn't light up is a minute I can't spend doing something that'll makes me quicker money."

Should Albert Einstein have said, "Every minute I spend working on physics equations is a minute I can't spend doing my current job as a clerk"?

Michael Jordan was cut from his high school basketball team. Imagine if he had said, "Leaning to play basketball is harder than I thought. Maybe I should just go flip hamburgers instead."

It doesn't stop there. Eli Whitney was laughed at when he first showed his cotton gin. Barthelemy Thimonnier, the inventor of the sewing machine, was almost killed by an enraged mob of tailors who burned down his factory. Cyrus McCormick had to try for fourteen years before he was able to get people to use his reaper. George Westinghouse was publicly considered a fool for stating that he could stop a train with air. Samuel Morse had to plead before ten Congresses before they would even look at his telegraph.

If any of those people had had your same shortsighted attitude, the world we live in wouldn't even exist.

Josh Susanto wrote:

Building is important to me. Watching tutorials that skip over the part about whether something will actualy work at all (andy why) is not important to me.

You really think every tutorial on every subject should be capable of troubleshooting every tool covered? If that's the case, then I guess you'll never let yourself learn anything from any tutorial, ever. That's a shame.

Josh Susanto wrote:

Focusing on potential long term challenges while ignoring immediate problems is not necessarily a better idea than doing the reverse. Step B is potentially important. But not if you have no way to get past step A.

There's ALWAYS a way to get past step A. Immediate problems and long term challeges are interdependent. Solving one usually means ending up in a better position to solve the other as well. Neither should be ignored.

Josh Susanto wrote:

I didn't have to take part in anything to learn how to make sculpties. Every button in Sculptypaint 092 does exactly what it says it does, except that surface images load reversed left to right (easy to work around that). Lunapic and Irfanview are similarly transparent. No lessons needed. No tutorials. No endless pages of FAQ's that don't answer the damn question.

Every button in every program we've discussed does exactly what it says it does, too. You just seem to think that if you don't immediately understand what you're looking at, then any effort to grow into understanding it must be a waste of time. That's tragic.

Josh Susanto wrote:

My failures so far are due to a gap between what people say and what they seem to think they say.

I can't agree with that.

Consider this. You've asked several questions about Blender so far, and I've been able to answer most all of them in detail for you, even though I don't use the program actively. How do you think it is that I'm able to do that? The answer is twofold. First, I simply exercise more intellectual curiosity than you seem to want to allow yourself. Second, unlike you, I assume that what people say in tutorials and such is in fact what they actually mean.

For example, when you asked the question about Python installation just now, I remembered that I had seen something about that in a tutorial video I watched a while back. I didn't remember exactly what it said, of course, but I did remember it said SOMETHING on the subject. So, I went back and watched it again just now, and then relayed the relevant information to you here. There was no gap between what the author said, and what he thought he was saying, nor was there any such gap in my repeating it. The information is the information.

Why did I watch that video in the first place? I was curious about how a few things were done in Blender, since I don't use it. No other reason.

If I, as a non-user, can so casually find such information without even half trying, then there's absolutely no reason that you as a potential active user can't do the same. You just have to be inquisitive enough to actually do it instead of sitting around complaining about not doing it.

Josh Susanto wrote:

My successes are due to reducing as much as possible any dependency on anyone to say anything to me.

I think you know that's not true. Otherwise you wouldn't be here asking questions.

Consider what's happened just in the couple of days since this discussion began. You had been previously unable to get Blender running, but because we talked about it, now it's working. You hadn't previously known about several other viable options that you might like in case Blender isn't for you, but because I told you about them, now you do. You didn't know the Python thing, but because you asked, now you know. Seems to me, you're now leaps and bounds ahead of where you were before, precisely because someone is saying things to you.

Josh Susanto wrote:

You're doing a lot better than most. Congratulations.

I do better than most at a few things, and I do worse than many at a lot of things. The same is true of everyone. We all have our particular areas of expertise. It just happens that the digital arts is an area upon which I decided to focus my efforts over the past several years, because I wanted to make a career out of it. If it weren't this, I would have found something else upon which to concentrate the same energies.

You could do just as well at this, if you were to make the same commitment to learning how that I did. Anyone could. Not everyone will, of course, but everyone could.

Josh Susanto wrote:

I don't need classes. I need applications that do what they say they do.

They all do what they say they do. But the user does bare responsibility to learn how to work them, and that's what classes and other avenues of learning are for.

If it were really true that the apps don't do what they say they do, how would you explain the fact that the rest of us get the results we get from them? Are we all just pretending?

Look, the only difference between those of us who make these things work for us and you, is we've put in the time and effort to learn how to use them, while you thus far have refused to.

Josh Susanto wrote:

It's a social approach I have tried periodically on principle. So far what you're seeing here is an example of this approach being a lot more productive than usual.

If it's not usually productive, why keep doing it?

Josh Susanto wrote:

Otherwise, I can see what the buttons do by pressing them. If they do anything.

That's really not how it works, Josh. I can promise you if you just randomly push buttons in a full-featured 3D platform like Blender (or Maya or Max or any of the other comparable programs that are out there), you won't accomplish a thing. There are need-to-know concepts, principles, and practices that you have no idea about right now, that cannot be discovered merely by pressing buttons.

After reading this statement from you, I'm beginning to question whether it was really the case that the buttons you pressed before actually didn't work. The functions behind most of them can only be applied to selected objects, and some of them can only be applied to certain types of objects in certain types of situations. If you have nothing selected, most of the buttons will appear to do nothing.

Josh Susanto wrote:

My attitude isn't what kept those buttons from doing anything when I clicked on them.

It very well might have been, if it was misperception, rather than malfunction, that was happening. But either way, It's certainly true that your attitue is what prevented the problem from being solved until I made a point of getting you to solve it in spite of yourself.

Josh Susanto wrote:

Forums are for software that works. Not for software that doesn't work.

They're for teaching you how to get it to work if you're unable to do so on your own.

Josh Susanto wrote:

With Direct Delivery still "just around the corner" as it was in August, can you really dismiss my total attitude on these fora as mere pessimism?

I'm not getting the reference. What does "direct delivery" mean in this context?

Whatever it is, the answer is yes, your attitude is pessimistic. The problem with pessimistic people is they always think they're being realistic, not pessimistic. That's a hard rut to get out of.

Josh Susanto wrote:

I'm the guy who told all of you that Brooke would lie to you, then told you that she was lying, and then proved that she had lied. If you're not seeing may posts on the Merchants forum attributed to Brooke recently, I hardly think that's due to my mere pessimism.

I have no idea who this Brooke character might be, nor do I much care. Whoever she is, and whatever your dealings with her have been, none of it has anything to do with what we're talking about here.

Josh Susanto wrote:

There has been no conspiracy around Blender. It's simply been a matter of the huge number of people who were unable to use it mostly logging off and forgetting about it rather than speaking up, much as did the thousands of people Rod mentions not completing the registration process for SL accounts. .

They didn't speak up, yet you somehow magically just know they exist?

Again, please don't present assumptions as fact.

Josh Susanto wrote:

I still maintain that millions of others did not use it successfully, but just didn't make as much noise about it here or elsewhere as did people for whom it worked just fine. When most people crash their whole system with a piece of free software, they don't keep trying to get it to work. They uninstall it, eradicate any trace of it from their hard drive, and move on to something else.

Oh, now it's millions, huh? You know this how?

Look, there's no question that not every program is the ideal choice for every prospective user. Sure, plenty of people who have tried Blender have opted not to stick with it, for all kinds of reasons. But the exact same can be said of every single program in the world. It's just a given.

But none of those other people are you. The only thing relevant to this discussion is the fact that YOU were having trouble getting it to work. Now that that's no longer the case, I do hope you'll commit to learning it properly.

Josh Susanto wrote:

The solution is working software. Not positive visualization or deep breathing exercises.

Yet you were unable to get working software to work for you. No Josh, the solution is learning how to solve whatever problem might be preventing it from working properly, if and when it's not.

Josh Susanto wrote:

Ridiculed. Yes. People say they are not deterred by the threat of ridicule, but they mostly are.

Where is this supposed "threat of ridicule" even coming from? Have you ever seen anyone on this forum, or the Blender forums, get ridiculed simply for asking how to do something? The whole point of these forums is to ask for help. They serve no other purpose.

Even if someone did get ridiculed by some random idiot on a forum, why should they care? If somebody is truly so insecure that the fear of an imagined potential insult from a total stranger who knows nothing about them is enough to prevent them from proceeding, then they've got much bigger problems than whether or not they can make 3D content.

Josh Susanto wrote:

You mean this page?

Apparently I should have tested the link before submitting it. Sorry about that. Direct access to specific pages in the manual does not work, for some reason. You have to walk though from the main help page to get there. I'll explain the steps:

1. Go to http://support.google.com/sketchup/ . (That link does work. I tested in two different browswers just now.)

2. Under where it says "Learn more about usng SketchUp", click "Learn the basics".

3. Click "SketchUp Getting Started Guide".

From there, you'll see links to all the various instructional topics for getting started with the program.

Josh Susanto wrote:

I have given myself a chance to make stuff that other people are not making by using tools that other people are not using. If I'd stuck with the program you describe, I wouldn't have been able to produce much of anything. How is that better?

Had you managed to get it working back then, and had you learned to use it, you'd have done many great things with it by now. There's nothing you can do in SculptyPaint that you can't also do in Blender (or any other modeling program that can export sculpt maps). So at the very least, you would have been able to do the exact same things you've already done.

You also would have been able to do other things that you haven't been able to do thus far, because SculptyPaint can't do them. That opportunity is now right in front of you at this point, since you do have Blender working, finally.

Chosen Few · December 11, 2011

Josh Susanto wrote:

What I do know is that the #1 reason cited (to me) for not building things is that users can't make adequate sense of the tools available to them.

That's to be expected. It's not for everyone. If it were, everyone would do it.

I firmly believe everyone CAN do it, but it will alway be the case that most people won't. Different strokes...

Josh Susanto wrote:

my only serious complaints would be the mirror imaging issue and the fact that it doesn't load sculpts out at 256x256 but is limited to 128x128.

Be glad that it doesn't. 256x256 images cannot be uploaded losslessly to SL, only 128x128 and smaller. If I were you, I'd be upset that it's not outputting 64x64. As you know, 64x64 and 128x128 both produce the same resolution sculpt mesh. 75% of the data in a 128x128 sculpt map is just wasted.

Josh Susanto wrote:

What you're missing is that some people at LL have been heavily invested, for a long time in the maximum apparent success rate of mesh in SL. You may also recall that Qarl, the guy who brought sculpts to SL, was fired under rather mysterious circumstances. There clearly is or has been a political aspect the sculpt/mesh question within LL, even if Qarl was not involved. Someone at LL still stands to benefit by making the sculpt supply dry up, one way or another.

That's just ridiculous, man. I know you love your conspiracy theories, but please do try to apply at least a modicum of logic when inventing them.

Nobody cares whether you use mesh or sculpties or prims. There are simply three options now, whereas before, there were only two. It's as simple as that. You're free to use or not use any or all of them at any time, and in any combination you want.

I can assure you, nobody's paycheck is in any way connected to which medium you choose to employ for any given model. There's simply no way it could be. That's just not how reality works.

I cannot understand for the life of me how you could possibly think anyone could "benefit by making the sculpt dry up". The fact that they've finally allowed us to use something better doesn't mean there's any hidden agenda. Put away the tin foil hat. It's clearly hampering your ability to use common sense and basic intelligence.

As for Qarl, he's a friend of mine, and although I'm not about to speak for him about why he was let go, what I can tell you is he played a vital role in bringing mesh to SL. He'd be the first one to tell you sculpties were added simply as a stopgap measure, just to give SL some mesh-like capabilities without having to redesign anything, while they were still trying to figure out how to make full blown mesh support work.

Sorry to have to be the one to tell you that reality is so sensible and straight forward. I know it must be a huge inconvenience for you. No doubt your wild nonsensical imaginings are far more fun for you in your own mind.

Josh Susanto wrote:

it's difficult for me to imagine anything much easier to use than Sculptypaint 092. As I've said, the buttons all do what they say they do. Nothing is at all msyterious. All I had to do to develop ease of use was to press the damn buttons a couple of times and see what they did.

I'm not talking about ease of button pushing. A button is a button. What I'm talking about is the level of direct control you have over every aspect of what you're doing.

Because you have no experience with this yet, you're not in a place to understand what I mean. Once you do learn what you've been missing all this time, you'll be amazed just how blinded you were before.

Josh Susanto wrote:

Except that the rest of the analogy is that when someone asks for advice on running shoes, they're always told how much more athletic they could be if they bought a uniform and took karate classes for 10 years.

That's not what happened here, Josh, and you are seriously starting to piss me off by continuously telling me I was in the wrong for making a simple FYI suggestion. Providing the best possible information is the only right thing to do, and I'm disgusted with you that you would try to stand in the way of that.

IT IS A FACT that the kinds of items the OP was asking about can be more easily and more efficiently made as arbitrary meshes than as sculpties. That's not an insult to sculpties. It's not an over-hyping of mesh. It's not furthering the agenda of any secret society. It's just a plain and simple statement of reality, nothing more, nothing less.

The fact that you keep denying this only makes you look like a total buffoon who has no idea what you're talking about (because you really don't). Once you've become a skilled mesh modeler yourself, you'll have a leg to stand on in these discussions, and you can make all the comparative assessments you want, from an educated point of view. Until then, all you're doing is making stuff up, and I really wish you'd cut it out. It benefits absolutely no one, least of all yourself.

Josh Susanto wrote:

Sculpty vehicles act they're made of spheroids.

That's why you should always use at least one regular prim in a sculpty vehicle. For most cars, you just need two transparent cubes to serve as a physics lattice.

Josh Susanto wrote:

I can understand people being willing to make only to replicate the design of specific RL vehicles. Otherwise, if the idea is to make a car, why not just make a car?

Your thinking is so painfully limited on this. It's not just about replicating RL vehicles, although that's certainly one use. There are all kinds of wholly original designs all over the grid. Your reduction of sculpty vehicle artists to nothing more than human Xerox machines is insulting.

Josh Susanto wrote:

If many of them are going to be used, I can see a serious mathematical advantage to using meshes, yes. Assuming that they consist of many straight lines and flat surfaces.

Another example of limited thinking. The mathematical advantage you mention applies as much to the one as it does to the many. That's just simple logic. And the benefits go well beyond just straight lines and flat surfaces.

Josh Susanto wrote:

Oblongs have become somewhat synonymous with that, but there are plenty of other good uses for them. Ponk Bing's plants, for example, I think must be at least about as efficient as mesh, if not more efficient. .

How many times do I have to remind you that an arbitrary mesh can be made identically to the sculpty? So, even if sculpty geometry happened to be the most efficient structure for a given model, it still couldn't be more efficient than an arbitrary mesh, because the arbitrary mesh could always be made exactly the same way.

In nearly all cases, sculpty geometry is an extremely inefficient way to create any given shape. Therefore, arbitrary mesh, which does not have the sculpty restrictions, can almost always be more efficient.

These are the facts. Continuing to deny it is ridiculous.

Josh Susanto wrote:

It's a pervasive theme in mesh/sculpt discussions, even if it is never directly stated.

The "pervasive theme" is in your own mind. All appearances are you've got this imaginary theory about secret agendas, and nothing will dissuade you from it, so you inaccurately and unfairly project it onto the things the rest of us say.

In reality, I've never seen anyone suggest that sculpties no longer have any place at all. There's little point any more in using them for most of the things they've been traditionally used for, but that's hardly the same thing as their being completely useless.

Josh Susanto wrote:

If sculpts are a waste of data, then the implication is that they should be replaced by something not as wasteful.

And how exactly is that a problem? ANYTHING that is a waste of data should be replaced by something less wasteful. Again, this is simple logic. Why is logic such a problem for you?

Josh Susanto wrote:

Since there's a belief by many that sculpts are inevitably less data-efficient than mesh, it may be construed that these people would have to agree that sculpts should all be eventually replaced.

It's not a belief. It's a quantified fact. In nearly all cases a sculpty model is inherently less data-efficient than a well made, equivalently shaped, arbitrary mesh model. No one disputes that (except you, apparently).

As for sculpties being "replaced", I assume you mean deprecated. That will never happen. LL will not do anything that would destroy existing content. As we've discussed before, sculpties are here to stay.

Josh Susanto wrote:

As to my own work, if someone can get a more effcient result with a mesh version, I think they should do that.

I'd say you should be the one to do it.

Josh Susanto wrote:

I'm just not persuaded that this is necessarily always the case as some have seemed to obliquely suggest.

It's not always going to be the case, and no one has suggested as much. It will be the the case quite often, though, and you'd do well to make peace with that fact.

Josh Susanto wrote:

I think that depends not only upon the type of item, but on the specific design of the item. No?

And we're back to the semantics game. Exactly how many ways do you think there are to design a planked floor, Josh?

There's not a window design or a picket fence design or a plank floor design imaginable that could not be made more efficiently as an arbitrary mesh than as a sculpty.

Josh Susanto wrote:

It has actually been an additional deterrent to me to get into mesh. I was actually more excited about mesh before it was deployed, and I've become less and less excited about it since the deployment.

It's a shame you let the irrelevant get to you like that. I don't let anyone else make my business decisions for me, especially not my customers. My business strategy has always been if you build it, they will come. I decide what my customers are going to buy, not them. Sitting around waiting for your market to create itself is a formula for going out of business. The most successful entrepreneurs are always those who create their own market.

All those great men I mentioned earlier, Edison, Westinghouse, Whitney, Morse, they didn't wait for the public to demand things of them. They saw new ways to create, and they made it happen. Public demand inevitably followed.

The same thing will happen with mesh models in SL. Right now, they're still at a disadvantage in that not all viewers support them. But once that's no longer a factor, the market landscape will be very different. With few exceptions, customers aren't going to care what an item is made out of. They're just going to care that it looks good and performs well.

Josh Susanto wrote:

I suppose it could be that, but I don't think so. The vibe has been more along the lines of "I like this thing, but please tell me it's not mesh so I won't be afraid to make it even bigger." People tend to like my stuff big for some reason.

That's interesting, and I see their point, certainly. But you as you've already stated, the kinds of items you make are a bit different from most. As I've told you before, you've carved out a nice niche for yourself, and you've put the sculpty as a medium to a very practical use.

Drongle McMahon · December 11, 2011

Intrigued as I am by this discussion, I thought I would do an experiment to improve my understanding of some of the points raised. So I took a wall texture (my own photograph, manipulated in Gimp to make it tileable) and applied a simple version of what I unserstand to be Josh's technique to generate the sculpt map to fit the texture, all in Gimp. This took all of ten minutes, even though it was the first time I had tried this.

The texture was 512x512 (I used the largest version deliberately), so first I imported the sculpt map at that size, the size it was made at. Then I simply downsampled it without interpolation to 64x64. The results of these two sculpt maps are shown at the top of the picture. They are not quite identical because the sampling in Gimp doesn't use the same pixels as the sampling during generation of the sculpty mesh. However, this could easily be made identical just by shifting the pixels before sampling. As it is, I think it adquately illustrates Chosen's point that map sizes greater than 64x64 are superfluous.

Then I imported the 64x64 sculpt map into Blender (2.49 + Primstar), re-exported it as a Collada file, uploaded that (with a box physics shape) and applied the texture. The mesh version, on the left, and the sculpty version, on the right, are compared in the lower part of the picture. Not much difference in appearance, but even at this relatively modest size, 3.2 x 3.2 x 0.75, the mesh has a LI of 4, while the sculpty is, of course, 1.

Looking at the mesh in Blender, it is perfectly clear that it would be quite involved to match it to the texture from scratch, and that it would be much more work than making it using Josh's approach. The increase in the LI also means that any customer would choose the sculpt before the mesh. I have not compared the LOD behaviour in detail. The mesh used the auto LOD and their behaviour was similar. The mesh physics shape is somewhat superior.

The triangle numbers, and therefore the rendering load, are identical. It may well be that some optimisation of th mesh could reduce this, maybe to a half, without significantly changing the appearance. That would reduce the download cost by a similar amount. This would be more work that generating the mesh this way though.

So for this particular kind of application, Josh's arguments in favour of sculpties seem to be quite reasonable, except that he could reduce network load by using subsampling of the final sculpt maps. That having been said, I totally agree with Chosen that for the vast majority of objects, which are not constructed in this way, the same arguments are not applicable.

Josh Susanto · December 11, 2011

I'll answer Chosen eventually.

In the meantime, I think I should explain that my thinking behind the 128x128 maps is simply that if someone wants to export them and edit them, more data is better than less data.

I, myself, have often cut up sculpt maps at 128 to get a piece of the same shape without resolution loss.

If someone wants to push the resolution of a particular shape in order to get a rounder curves from some angle, they need the data to be available already because they can't very easily create all of it correctly just by pressing the smooth button on the their editor.

People sometimes flatten my columns, I'm told, and I have suggested to people that they may also wish to rotate the sculpt map on the green axis to get the best flattening angle. Both of these things are bound to go better if there is "extra" data, rather than just enough.

Drongle McMahon · December 11, 2011

Yes. I suppose you could provide both, 64 x 64 for efficiency and larger for editing, but then that would cost you more in upload costs. Hmm.

Chosen Few · December 11, 2011

Cool experiment, Drongle. Thanks much, by the way, for introducing more good to this thread than had previously been in it.

I just did a similar experiment of my own, utilizing a slightly different technique, to get the same kind of result. Let me first explain what I did, and then I have a few comments/suggestions/questions on some of your points.

To crete the mesh in the above picture, I did the following:

1. I whipped up the seamless texture very quickly, in Filter Forge. This took less than a minute, including render time.

2. Instead of generating a sculpt map from the texture like you did with yours, I generated just a standard 8-bit grayscale bump map. This took just two extra clicks in Filter Forge, since it happens the program is already set up for that. Had I been working from a photograph or some other pre-existing image, it just would have been an extra minute or so to generate the bump map in Crazy Bump or Photoshop.

3. In Maya, I created a plane with 32x32 quads (identical to a sculpty plane). I then applied the diffuse texture to the color channel, and applied the grayscale bump map to the displacement channel. (Yes, the displacement channel, not the bump channel. This is super important.) This step took about 30 seconds in total.

4. Next, I converted the displacement to polygons. Maya makes this super easy to do, just two clicks and it's done. (I would assume Blender has a similar option?) This part was about 20 seconds, including the time it took for the conversion to calculate. At this point, the LOD1 model was done.

5. To create the other three LOD's, I simply reduced amount of divisions on the original plane, and spit out new displacement-to-polygon conversions. For LOD2, I did 16x16 quads. LOD's 3 and 4 are 8x8 and 4x4 respectively. For the physics mesh, I reduced the divisions all the way to 1x1, to yield just a single quad.

The end result, is a texture-generated mesh, just like the kind of texture-generated sculpties Josh likes to do. The whole thing, from start to finish, took less than five minutes.

Drongle McMahon wrote:

Not much difference in appearance, but even at this relatively modest size, 3.2 x 3.2 x 0.75, the mesh has a LI of 4, while the sculpty is, of course, 1.

At the same size, mine has the exact same LI as yours, which is to be expected, since it's the same number of triangles. The full specs are 3.6 dowload, 0.2 physics, 0.5 server, 684 display.

As a sculpty, the numbers are 2.5 download, 1.8 physics, 0.5 server, and 861 display.

This, of course, confirms what we already knew, which is that for objects like these, the sculpty is indeed the better choice in some aspects, while the mesh is the better choice in other aspects. With this, as with all things, it's a question of which particular aspects are most important, for the particular project. If you absolutely need the amount of polygons that are in a sculpty, and you absolutely need the land impact to reamain at 1 at all sizes, then the sculpty is of course the way to go. However, if you want faster rendering, and appropriately accurate physics, then the mesh is the way to go.

Josh mentioned seems to feel his customers value the "one prim fits all sizes" feature of the sculpty, more than anthing other factor. If that's indeed the case, then it only makes sense that he continue doing what he's doing.

But that's not the final word on the subject, of course. We do have to consider, case by case, whether objects like this really need so many polys in them. If I start with the 16x16 version instead of the 32x32 version for LOD1, use the 8x8 for LOD2, the 4x4 for LOD3, and the single quad for LOD4, the land impact at that same size comes out to 1 instead of 4, and the model looks every bit as good.

Full specs in that case are 0.6 download, 0.2 physics, 0.5 server, 434 display. That beats the sculpty by miles in every measurable way (except the server weight, of course, which is identical), and it's still not even as well optimized as it could be if I really wanted to make a project of it.

From that, we can undisputably conclude that for any object of this type that does not need to keep the land impact the same across all possible sizes, there's simply no justification to be found for using a 32x32 quad sculpty.

Drongle McMahon wrote:

Looking at the mesh in Blender, it is perfectly clear that it would be quite involved to match it to the texture from scratch, and that it would be much more work than making it using Josh's approach.

Well sure, if you're placing every vertex by hand. Had I manipulated all 1000+ vertices one at a time, it would have taken me hours instead of minutes. And that's just for one LOD, let alone all four.

I can't imagine anyone in their right mind choosing to work that way for this type of model. This is precisely the sort of thing that displacement maps are for. That's why Josh found that sculpties naturally work so well for this, since as you and I both know, sculpt maps are just multi-axis displacement maps.

Drongle McMahon wrote:

The increase in the LI also means that any customer would choose the sculpt before the mesh.

I don't know that it's quite fair to say ANY customer. There will be those who will be savvy enough to weigh all the factors.

Right now, the vast majority are singularly focused on land impact, because that's all they've ever known in SL. But as time goes on, people will osmose the rest of the picture. The full set of weights will become part of the cultural consciousness of SL, just as every new factor that has ever been introduced always eventually has. It will take some time, but it WILL happen. It has to.

Drongle McMahon wrote:

I have not compared the LOD behaviour in detail. The mesh used the auto LOD and their behaviour was similar.

I deliberately created my LOD's to mimic sculpty behavior very closely, just for the sake of doing as fair of base comparison as possible. But for real projects, the mesh LOD's can obviously be tweaked in different ways, depending on the specifics of each model. Any given detail that might get obliterated by sculpty's unchangeable LOD reduction could be preserved in the arbitrary mesh's LOD's, while keeping the poly count at each level the same as, or lower than, that of each sculpty LOD.

The texture I used actually makes for a pretty good example of this principle. Some of the rock formations stick out quite far from the surface, while others are almost flat. The flatter areas could be reduced to just a handful of polys in the lower LOD's, while leaving the higher amplitude areas completely intact.

For a more uniform example, like the one you made, that gets a little harder to pull off, but still, there's a lot you can do to make it both look better and cost less with a smart reduction vs. a dumb reduction.

(By the way, before any of our resident conspiracy theorists gets bent out of shape, thinking I just called sculpties dumb, relax; I did nothing of the kind. Those are actual terms. A smart reduction is one that is done with a focus on preserving important details, while letting less important ones go. A dumb reduction is one that is done uniformly by the math alone, without any mind toward the real details of the model. All sculpties, by definition, employ dumb reduction, as do any mesh models which have had their LOD's auto-generated by the uploader.)

Drongle McMahon wrote:

The triangle numbers, and therefore the rendering load, are identical.

The viewer doesn't seem to think so, at least for mine. The sculpty's display weight is higher than that of the mesh. I'm not sure precisely why. Did you get different results than I did?

Drongle McMahon wrote:

It may well be that some optimisation of th mesh could reduce this, maybe to a half, without significantly changing the appearance. That would reduce the download cost by a similar amount. This would be more work that generating the mesh this way though.

I gotta say yes and no on this one. The yes part is that optimizing the mesh absolutely can significantly reduce the costs, across the board, with little or no discernible difference in appearance, as I illustrated above. That wasn't even a particularly good optimization, but it still was a huge boon to the costs, and it didn't affect the appearance at all.

The no part is it doesn't have to be more work. There are lots of simple ways to optimize that require little or no extra effort. It's more a question of starting off the right way than anything else. As long as you don't box yourself into a corner, work-flow wise, there's always plenty of room to work smart, not hard.

Drongle McMahon wrote:

So for this particular kind of application, Josh's arguments in favour of sculpties seem to be quite reasonable, except that he could reduce network load by using subsampling of the final sculpt maps. That having been said, I totally agree with Chosen that for the vast majority of objects, which are not constructed in this way, the same arguments are not applicable.

Some of Josh's arguments are perfectly reasonable, yes. As I've said before, I applaud him for having found such an effective way to put sculpties to use. Certain other of his arguments, on the other hand, are... well, I've already responded to those, so no need to get into it again.

Chosen Few · December 11, 2011

Josh Susanto wrote:

I'll answer Chosen eventually.

Take your time.

Most of what we've been bickering about is off topic, anyway. You and I together seem to have a particular talent for that, when we get into it.

Drongle has now steered the conversation in a much more interesting, and much more beneficial, direction.

Josh Susanto wrote:

Yes. I suppose you could provide both, 64 x 64 for efficiency and larger for editing, but then that would cost you more in upload costs. Hmm.

I think that's a fine idea.

While we're on the subject of offering optons, I also think that if you were to provide mesh versions of your products, made similarly to how Drongle and I have done it here, and packaged them with your sculpty versions, you'd be providing your customers an amazing service. Those who want the uniform land impacts can use the sculpty versions, and those who are more concerned about display weight and such could use the mesh versions. It would be a total win-win for everybody, and if you do it the way I described, it would only take you a few extra minutes per model.

The only obstacle right now, of coure, is that you don't yet know how to do what we did here. But you could learn relatively quickly. It's all basic beginner level stuff, nothing advanced.

Drongle McMahon · December 12, 2011

You are right about doing it with a displacement map. I never used that in Blender, but I guess it would be just as easy as the sculpty-to-mesh route, maybe easier.

"The sculpty's display weight is higher than that of the mesh. ... Did you get different results than I did?"

The render weight for the sculpty was about 1.5 times that for the mesh, but I regard this as an artefact of the calculation. What's actually sent to opengl to be rendered is essentially identical in all respects. I think most of the discrepancy comes from the fact that the mesh render cost uses an estimate of the triangle count based on the actual data sizes at each LOD which includes the variable effects of compression, while the sculpty cost presumably uses the actual triangle count. I must look at a 4x4 sculpty and see.....

Arrggghhh! No. The render weights for a 8 x 8 sculptmap and that wall, at the same size, are 872 and 886, respectively. Something very wrong there! The wall has 64 times as many triangles! I'll have to do a jira for that!

eta: https://jira.secondlife.com/browse/VWR-27827

Josh Susanto · December 12, 2011

>Yes. I suppose you could provide both, 64 x 64 for efficiency and larger for editing, but then that would cost you more in upload costs. Hmm.

At my prices, it seems like it should be worth someone's trouble to export, edit, and reload any data they actually want to use.

I also have a lot of 1024 surfaces that someone might want to 512 or smaller.

It's always easier to remove data than to add it back in, so I generally give people more rather than less.

One exception is that way I make seamless textures, some of them actually look better at 512 than at 1024 (yes, I have thought about this a lot).

Josh Susanto · December 12, 2011

>The Machinimatrix Blender Trail is fantastic place to start, so you know.

I'll start there. Thanks.

>Eli Whitney was laughed at when he first showed his cotton gin.

He's actually in my family. He did invent the interchangeable rifle, but the cotton gin, he merely had to patent for a woman in the family of a friend of mine because, as a woman, she couldn't get the patent at that time. Funny you should choose such an example.

>You really think every tutorial on every subject should be capable of troubleshooting every tool covered?

No. I think one of them should explain how to make it work when it doesn't work. If there was one, I didn't find it.

>There's ALWAYS a way to get past step A.

Understood. I actually have a reputation as creative problem solver. My solution when step A doesn't have a solution is not to keep hitting it, though, but to go around it and try to come back at some point if there's an advantage. That's what I've done here. I have had music students who say they chose to forget about music practice while their instrument is in the shop and they can't get another one during that time. I have always told them they can do something to improve themselves musically during that time, and I give them specific examples. What I do not do is to tell them to go to the shop every day and ask "is it ready yet?" and then go home and forget about it. Blender was not ready for me, and when it was ready, nobody told me, which is understandable.

>Every button in every program we've discussed does exactly what it says it does, too.

So far, I have found that not to be true of the import button on Blender. It clearly does something, though. When I find out what, I'll let you know.

>How do you think it is that I'm able to do that?

Others are able. But they don't do it. You are a kind of anomaly in answering, just as I am a kind of anomaly in asking. If this is conversation that 2 people could have had here months ago, why didn't it happen?

>For example, when you asked the question about Python installation just now, I remembered that I had seen something about that in a tutorial video I watched a while back.

What's the video, and why didn't the Blender website provide a conspicuous link to it if it's necessary in order to get their program to work? There was a gap between what was stated on the Blender website and what was necessary to say. How can you say there was not?

>I think you know that's not true. Otherwise you wouldn't be here asking questions.

I ask a lot of questions, yes. One reason is that most of them never get satisfactory answers, so it's sort of a numbers game with me. A decade after getting my MA, I'm still waiting for anyone in the Schenkerian analysis community to tell me how what their models model differs from the emphatic structure of a mnemonically optimized performance. I had to start asking this question because when I simply asked what their models model, they would just say "the structure", but declined to explain what the nature of "the structure" was, telling me that that should be intuitively obvious. What's intuitively obvious to me is that if they can't say whether or not their models model a specific thing which I have asked whether they model, they probably don't really model anything in particular (and now you know why I'm not working as a music theory professor).

>You could do just as well at this, if you were to make the same commitment to learning how that I did.

My commitment is to results. I lost a decade or more of my life to learning $hit.

>If it were really true that the apps don't do what they say they do, how would you explain the fact that the rest of us get the results we get from them? Are we all just pretending?

In the Blender example, the results are not strictly from Blender, but from Blender in combintion with other things unstated. It's like selling someone a car without mentioning they have to buy the keys separately. I understand that Blender is "free", but after a buggy program costs me a certain amount of my time, I can no longer consider it to be "free".

>If it's not usually productive, why keep doing it?

Scientific principles. If someone else is getting a different result by doing something I've done, I am eventually compelled to figure out why. Often, the explanation is that they aren't doing things exactly the way they describe them.

>That's really not how it works, Josh.

That's how Sculptypaint 092 works.

>After reading this statement from you, I'm beginning to question whether it was really the case that the buttons you pressed before actually didn't work.

Earlier, none of them worked at all. Now, most of them at least do something, even if I'm not yet sure what they do.

>But either way, It's certainly true that your attitue is what prevented the problem from being solved until I made a point of getting you to solve it in spite of yourself.

I don't think that's accurate. I tried to use Blender at least a dozen times obver a period of months before I finally decided that practically anything else would be a less counterproductive use of my time. And comparing the results I've got so far by using other things and trying and failing to get Blender to do anything, there's an abundance of evidence to support my conclusion.

>They're for teaching you how to get it to work if you're unable to do so on your own.

Then they failed in my case. I mentioned repeatedly that Blender did not work for me, and this is the first time that anyone has done anything about it.

>I'm not getting the reference. What does "direct delivery" mean in this context?

>Whatever it is, the answer is yes, your attitude is pessimistic. The problem with pessimistic people is they always think they're being realistic, not pessimistic. That's a hard rut to get out of.

Saying that there was still no Direct Delivery on 1 December is not pessimism. It's an historical fact. How is (insert actual deployment date of Direct Delivery here) "just around the corner" from August, as I was repeatedly told by "optimistic" people who used DD as a way to dismiss my complaints about aspects of existing Magi Box function? If it was "pessimism" that made me focus on solving my own Magic Box problems (I did) rather than just ignore them and wait for DD, then "pessimism" has actually been working pretty well for me here. I'm not pessimistic about my own problem solving ability. I'm only pessimistic about other people getting correct the stuff that they tell me, or each other. I'm the guy who removed his own kidney stone with a home-made tool after getting jerked around for a week by his insurance referral system. I also sued my last US boss without a lawyer and was awarded more than asked for. It took over year to prevail, and I did so from a different continent. But I did it. This tends to be the way I eventually get things done. Waiting for someone else to fix problems just because it's their job to fix them for me is expecting things to work the way they allegedly work, rather than how they actually work. And I have learned to expect otherwise, at least when it comes to dealing with my fellow Americans.

>I have no idea who this Brooke character might be, nor do I much care. Whoever she is, and whatever your dealings with her have been, none of it has anything to do with what we're talking about here.

That's pretty weird. It has a lot to do with the basic communication dynamics in this culture, here. People do mostly mean very well, but meaning well just doesn't cut it when people get out of their depth.

>They didn't speak up, yet you somehow magically just know they exist?

They talked about it when I asked. One reason I couldn't get more help with Blender is that people I would have assumed were using it were instead using cracked versions of other 3D modeling software because they found Blender to be extremely frustrating in various ways. And these include some pretty significant builders. People whose work I'm confident that you've seen.

>Yet you were unable to get working software to work for you.

It wasn't working software. It was software that worked in conjuction with something else not provided. A car that will work perfectly if it has an engine is not a working car if it does not have an engine. In fact, calling such a thing a car at all can be a little bit confusing to people who want to drive something.

>Where is this supposed "threat of ridicule" even coming from?

Mostly from their own minds. That I realize that explains why I'm even here. But there's plenty of ridicule in the merchants forum, for example. And other places. That you've been able to ignore it to the point that it doesn't even easily come to mind is probably to your credit.

>Apparently I should have tested the link before submitting it.

Now you're getting the picture.

The next link you provided works, but I've been looking at stuff that looks like the first link for a year. If I can't even trust Google to provide support links for its own apps, who can I trust?

Just you, apparently. If this stuff is not obvious to me, you can be pretty sure it's also not obvious to plenty of other people. And yet that first link still exists to still be posted for me by you.

>Had you managed to get it working back then, and had you learned to use it, you'd have done many great things with it by now.

Sure. And if I'd just guessed some correct lottery numbers and been optimistic enough to play them, I could have just hired someone to solve the problem for me a long time ago, too.

I appreciate your help a lot. But it's ironic that in order to get something fixed that someone somewhere has already been paid to get correct for a year or more, I have to come on a public forum and gently antagonize someone I know is really on my own side. Don't you think it's at all ironic?

Josh Susanto · December 12, 2011

>Be glad that it doesn't. 256x256 images cannot be uploaded losslessly to SL, only 128x128 and smaller. If I were you, I'd be upset that it's not outputting 64x64. As you know, 64x64 and 128x128 both produce the same resolution sculpt mesh. 75% of the data in a 128x128 sculpt map is just wasted.

The 256 option would at least allow me to archive a 256 version to make available to people off-world. It would also allow me to cut even smaller pieces out of the thing which have useful degrees of resolution. I realize that 512 can't happen with Sculptypaint models because the shape data is also color data, but I still blow the 128's up to 512 to put layers of gray relief data onto them. I can explain why, but I'm not sure anyone else would benefit from such ideas.

(What you're missing is that some people at LL have been heavily invested, for a long time in the maximum apparent success rate of mesh in SL. You may also recall that Qarl, the guy who brought sculpts to SL, was fired under rather mysterious circumstances. There clearly is or has been a political aspect the sculpt/mesh question within LL, even if Qarl was not involved. Someone at LL still stands to benefit by making the sculpt supply dry up, one way or another.)

>That's just ridiculous, man. I know you love your conspiracy theories, but please do try to apply at least a modicum of logic when inventing them.

Do you have a better theory? The correct response to a bad theory is a better theory.

>Nobody cares whether you use mesh or sculpties or prims.

Then why should LL bother to mislead people about the real benefits of each?

>I can assure you, nobody's paycheck is in any way connected to which medium you choose to employ for any given model. There's simply no way it could be. That's just not how reality works.

Someone got paid to bring mesh technology who wouldn't have got paid simply to maintain support for sculpt technology. That, in itself, is no reason to imagina a conspiracy. But it does contradict what you are saying.

>I cannot understand for the life of me how you could possibly think anyone could "benefit by making the sculpt dry up". The fact that they've finally allowed us to use something better doesn't mean there's any hidden agenda. Put away the tin foil hat. It's clearly hampering your ability to use common sense and basic intelligence.

I am wearing the same tinfoil hat I was wearing when I told people Brooke would lie to them. If you want me to stop acting paranoid, step one should be to get LL to stop fulfilling my paraoid visions.

>As for Qarl, he's a friend of mine, and although I'm not about to speak for him about why he was let go, what I can tell you is he played a vital role in bringing mesh to SL. He'd be the first one to tell you sculpties were added simply as a stopgap measure, just to give SL some mesh-like capabilities without having to redesign anything, while they were still trying to figure out how to make full blown mesh support work.

Much better. What happened to flexi-sculpts? Were they shelved simply because mesh was "right around the corner"?

>Sorry to have to be the one to tell you that reality is so sensible and straight forward. I know it must be a huge inconvenience for you. No doubt your wild nonsensical imaginings are far more fun for you in your own mind.

After seeing the personnel files of hundreds of doctors, I can tell you that reality is nowhere near as straight forward as most people think. After that job, I have started to wonder whether I was really paranoid enough in the first place. If you're happy in your panglossian paradise, pray that you never get the jobs I got or see the things I've seen.

(it's difficult for me to imagine anything much easier to use than Sculptypaint 092. As I've said, the buttons all do what they say they do. Nothing is at all msyterious. All I had to do to develop ease of use was to press the damn buttons a couple of times and see what they did.)

>I'm not talking about ease of button pushing. A button is a button. What I'm talking about is the level of direct control you have over every aspect of what you're doing.

There is no direct control, There is a mouse and a keypad. If you're using a glove, you might have mentioned that by now.

>Because you have no experience with this yet, you're not in a place to understand what I mean. Once you do learn what you've been missing all this time, you'll be amazed just how blinded you were before.

Possibly. I'm looking into it..

>That's not what happened here, Josh, and you are seriously starting to piss me off by continuously telling me I was in the wrong for making a simple FYI suggestion. Providing the best possible information is the only right thing to do, and I'm disgusted with you that you would try to stand in the way of that.

I retract the imputation of any ulterior motive, and I apologize for same.

>That's why you should always use at least one regular prim in a sculpty vehicle. For most cars, you just need two transparent cubes to serve as a physics lattice.

But wouldn't the whole thing be more efficient if it didn't use sculpties at all?

(I can understand people being willing to make only to replicate the design of specific RL vehicles. Otherwise, if the idea is to make a car, why not just make a car?)

>Your thinking is so painfully limited on this. It's not just about replicating RL vehicles, although that's certainly one use. There are all kinds of wholly original designs all over the grid. Your reduction of sculpty vehicle artists to nothing more than human Xerox machines is insulting.

Sorry, I didn't mean it that way.

What I meant was that using sculpts for cars when they are not absolutely necessary is... not absolutely necessary.

Regular prims can provide an array of shapes almost as wide as what I see being used for sculpts on cars. If I knew how to export regular prims I've variously contorted, I might be spending as much time texturing them as I do sculpting out textures. What they will not do is provide precise analogues to specific shapes on specific RL vehicles.

>Another example of limited thinking. The mathematical advantage you mention applies as much to the one as it does to the many. That's just simple logic. And the benefits go well beyond just straight lines and flat surfaces.

I think I get what you're saying here. But there is an economy of scale in terms of the time involved to produce something. To spend twice as long to make something that will only be used in one place in order to save a fraction of a prim might not be great math. But it also could be. Maybe.

(Oblongs have become somewhat synonymous with that, but there are plenty of other good uses for them. Ponk Bing's plants, for example, I think must be at least about as efficient as mesh, if not more efficient.)

>How many times do I have to remind you that an arbitrary mesh can be made identically to the sculpty? So, even if sculpty geometry happened to be the most efficient structure for a given model, it still couldn't be more efficient than an arbitrary mesh, because the arbitrary mesh could always be made exactly the same way.

That should be true, of course. But for some reason, I keep being told different things about this. I do intend to test it for myself. But as you can imagine, I don't necessarily trust the metrics as provided with metric tools provided by the same people who are providing the mesh technology.

>In nearly all cases, sculpty geometry is an extremely inefficient way to create any given shape. Therefore, arbitrary mesh, which does not have the sculpty restrictions, can almost always be more efficient.

>These are the facts. Continuing to deny it is ridiculous.

I don't deny that, given every imaginable shape, mesh should be a better way to produce most of them. What I deny is that it's clear that this actually is true for a lot of common shapes. The person I trust most actually has given up on mesh; not because it's bad for modeling, but because the math has turned out not to be what it should be.

>In reality, I've never seen anyone suggest that sculpties no longer have any place at all.

It doesn't need to be said. If there's no reason to continue using them, then they will gradually disappear, and that's fine. The persistent theme is not that sculpts need to go. The persistent theme is that there's no reason to continue using them; something of which I'm not convinced just by some numbers that even mesh users cannot agree add up correctly.

>And how exactly is that a problem? ANYTHING that is a waste of data should be replaced by something less wasteful. Again, this is simple logic. Why is logic such a problem for you?

Eliminating waste is not a problem for me. If there's a mesh version that's more efficient and just as good in other ways, that's what should be used. But a lack of consistency in what's reported in terms of the demands of meshes invites the question of whether it can really be stated with certainty that meshes are as reliably more efficient as LL wants us to believe.

>It's not a belief. It's a quantified fact. In nearly all cases a sculpty model is inherently less data-efficient than a well made, equivalently shaped, arbitrary mesh model. No one disputes that (except you, apparently).

It should be true, I would think. But there's enough lack of clarity about how great the benefit really is, that the whole idea that there is a uniform benefit comes into question. Mesh physics, for example, are assuredly more complex than sculpt physics. If sculpt physics are adequate for some purposes, then what's the point of putting extra demands on the system by applying mesh physics?

>As for sculpties being "replaced", I assume you mean deprecated. That will never happen. LL will not do anything that would destroy existing content. As we've discussed before, sculpties are here to stay.

Even I don't think there's necessarily any reason they need to be here to stay. If they are lost by attritiion to mesh equivalents, that's an acceptable process to me, assuming that the meshes really are better. I just have my doubts about something being better simply by virtue of being mesh.

(As to my own work, if someone can get a more effcient result with a mesh version, I think they should do that.)

>I'd say you should be the one to do it.

It's a matter of specialization of labor. There are plenty of people who know how to use Blender technically who apprently have no idea how to make at least some of the stuff that I make. Once I see them getting rich by converting my objects, maybe I'll get more excited about it.

>And we're back to the semantics game. Exactly how many ways do you think there are to design a planked floor, Josh?

I can only think of about 3 offhand.

>There's not a window design or a picket fence design or a plank floor design imaginable that could not be made more efficiently as an arbitrary mesh than as a sculpty.

If the LI is essentially the same, though, why not use a prim oven one already has, rather than spend weeks or months trying to decipher new software?

>It's a shame you let the irrelevant get to you like that. I don't let anyone else make my business decisions for me, especially not my customers.

Yeah. Customers don't matter. Phuk'em.

>My business strategy has always been if you build it, they will come. I decide what my customers are going to buy, not them. Sitting around waiting for your market to create itself is a formula for going out of business. The most successful entrepreneurs are always those who create their own market.

The formula that has worked for McDonalds is to do both. Keep providing new stuff to see if it's more profitable, but also refine the existing product line to approach some optimum cost/revenue intercept. I build what they ask for if I can. When they don't ask for anything, I build whatever I can.

>All those great men I mentioned earlier, Edison, Westinghouse, Whitney, Morse, they didn't wait for the public to demand things of them. They saw new ways to create, and they made it happen. Public demand inevitably followed.

At one point Edison specifically decided not to prioritize the invention of things for which there was no commercial demand. This is a point I make with some of my student when they study Edison. I pull a Rubik's cube out of my pocket and ask them how unprofitable has been this thing which Edison would not have bothered to invent (yes, I see the irony).

>The same thing will happen with mesh models in SL. Right now, they're still at a disadvantage in that not all viewers support them. But once that's no longer a factor, the market landscape will be very different. With few exceptions, customers aren't going to care what an item is made out of. They're just going to care that it looks good and performs well.

Agreed. I also think when they have more of a chance to compare meshes directly with sculpt equivalents, the response will continue to be mixed, and in new, unanticipated ways.

(I suppose it could be that, but I don't think so. The vibe has been more along the lines of "I like this thing, but please tell me it's not mesh so I won't be afraid to make it even bigger." People tend to like my stuff big for some reason.)

>That's interesting, and I see their point, certainly. But you as you've already stated, the kinds of items you make are a bit different from most. As I've told you before, you've carved out a nice niche for yourself, and you've put the sculpty as a medium to a very practical use.

Thanks for acknowledging this.

Josh Susanto · December 12, 2011

>The whole thing, from start to finish, took less than five minutes.

That sounds about right for such a project in my case, if even a bit slow.

OTOH, the time I didn't have to spend figuring out how to do it the first time is pretty substantial.

My best-selling products are of essentially this kind. The reason I don't have more of them is that not every surface image is equally commercially viable.

Josh Susanto · December 12, 2011

> I think most of the discrepancy comes from the fact that the mesh render cost uses an estimate of the triangle count based on the actual data sizes at each LOD which includes the variable effects of compression, while the sculpty cost presumably uses the actual triangle count.

... and the tip of the iceberg emerges...

>Arrggghhh! No.

Arrggghhh- YES.

Whether the discrepancies favor mesh or sculpt, they're always just a little bit weird, aren't they?

Drongle McMahon · December 12, 2011

Well, this one is heavily biased against sculpties, as it doesn't reward small sculpt maps, which make less triangles, and probably it doesn't exclude culled degenerate triangles which don't get sent to opengl. I am surprised to see you apparently weclcoming it.

Josh Susanto · December 12, 2011

I do not have a strong opinion yet, either way about the net result, because I still don't really trust the math either way.

So, until there's a clear problem to be solved by converting all my products to mesh, I'm not going to be very quick about it.

Another reason I won't be quick is that I just watched 4 videos right on the Blender website and I still don't know how to get data into or out of Blender.

Considering that, without these 2 functions, nothing else is going to be very productive, I suppose I should be puzzled that they didn't put that in either the 1st or 2nd video.

The reason I'm not puzzled is that if you look at the Blender interface as compared the interface for Sculptypaint 092, they're clearly designed by people who think very differently from each other.

That is: Sculptypaint 092 is designed by someone who wants me to move data into and out of the thing, whereas Blender seems to be designed by someone who wants me to go repeatedly to the Blender website and watch videos.

The next video is supposed to be about animation.

So I'm to understand that knowing how to animate is more important than knowing how to get even simpler kinds of data in or out?

Wow.

Josh Susanto · December 12, 2011

Sketchup looks like it might be easier to use than Blender, but it also seems to be intended primarily for things much more geometric than what I normally produce. I'm actually looking forward to that as a creative challenge.

OTOH, I'm also past my 4th Google Sketchup video and there has also been no mention of how to get data in or out of it, so I'm guessing I'll also have to learn how to use it to animate before I'm allowed to produce anything with it.

No?

Kwakkelde Kwak · December 12, 2011

ehm Josh, you seem like a smart enough guy to not start with "basic tutorial 1" then "2 " then..etc... if you want to learn how to do something specific as importing an object. I'm not all that familiar with sketchup, but if you google "import into sketchup" you'll have better chances of finding what you are looking for than by following all the tutorials. So to answer your question: no you do not have to be able to animate before you can import an object...or export for that matter.

Chosen Few · December 13, 2011

Sketchup is quite simplistic, yes. It's a no-frills extrusion modeler, or at least that's what it was last time I toyed with it. It may have evolved a bit since then.

I haven't been through all the videos, so I don't know in what order they present things. I would suggest making a point of getting at least somewhat familiar with all of it, though, even if you're not sure it will be necessary for what you're looking to do. More often than not, you'll find that things you don't think are relevant actually turn out to be.

You'll find extrusion modeling to be quite different from what you're used to with sculpties. It's actually somewhat fitting that you start with extrusions, because it's just about the oldest 3D modeling technique there is. If you've seen the original Tron, for example, those are all extrusion models.

Nowadays, extrusion is just one technique among many. When you get into a full featured platform like Blender, you've got just about every technique there is at your disposal. More singularly pointed programs like Sketchup can be good starters, but you'll eventually find yourself outgrowing them.

Chosen Few · December 13, 2011

Josh Susanto wrote:

He's actually in my family. He did invent the interchangeable rifle, but the cotton gin, he merely had to patent for a woman in the family of a friend of mine because, as a woman, she couldn't get the patent at that time. Funny you should choose such an example.

Very interesting. I did not know about the rifle, and I had no idea the cotton gin wasn't actually his own invention. It's cool that you're a relative.

Josh Susanto wrote:

No. I think one of them should explain how to make it work when it doesn't work. If there was one, I didn't find it.

I'll agree with you that in the past, Blender documentation has been sparse, and what was available has been awfully hard to follow. The vast majority of it appeared to have been written by engineers, for engineers. Lots of non-enginner people, including myself, had little stomach for Blender because of that. I used to complain about that all the time, in discussions on Blender.

It really wasn't until Gaia and the Machinimatrix team came along to produce such good tutorial videos that I changed my tune. I had always thought Blender was a good program suffering from a bad interface. But when Gaia and her team began making it make sense, in such an easily absorbable manner, I realized that the interface itself wasn't actually bad; it had just historically been a badly explained. Big difference there.

Whatever the problem you had been having in getting Blender to work in the past, the information you would have needed in order solve was out there, somewhere. But actually finding it, and understanding it once you had, might have been quite the challenge, depending on the time frame. It's a lot easier today than it was even a year ago.

In addition to great sources like Machinimatrix, YouTube is now full good tutorial videos, and there are now finally real books on Blender as well. But even with this wellspring of good information now flowing, I still don't know that any tutorial on how to make things with the program would be the right venue for presenting information on what to do if the program itself isn't working for you.

For comparison, a driver's ed class does have to assume you've already go a functioning vehicle. It's not going to teach you how to fix your car if it won't start.

That's why I keep saying asking for help is the best strategy when you can't get something to work, and you have no idea why. The chances of just stumbling across the missing information are remote.

Josh Susanto wrote:

My solution when step A doesn't have a solution is not to keep hitting it, though, but to go around it and try to come back at some point if there's an advantage. That's what I've done here.

That's often a wise strategy, and I don't fault you at all for having employed it in this case. What I did take issue with was your apparent stubborn insistence that this particular Step A was absolutely unworkable. Thankfully, we've now moved past that, and you're free to explore it at your leisure.

Josh Susanto wrote:

I have had music students who say they chose to forget about music practice while their instrument is in the shop and they can't get another one during that time. I have always told them they can do something to improve themselves musically during that time, and I give them specific examples.

Wise advice.

Josh Susanto wrote:

What I do not do is to tell them to go to the shop every day and ask "is it ready yet?" and then go home and forget about it. Blender was not ready for me, and when it was ready, nobody told me, which is understandable.

I don't think that's quite a fair comparison. In the music shop example, the work is totally out of the student's hands. There's nothing he or she can do directly, to facilitate the repair. That's not what we're talking about when it comes to correcting software malfunctions. In most cases, there's quite a lot the user can actively do to solve the problem.

Josh Susanto wrote:

So far, I have found that not to be true of the import button on Blender. It clearly does something, though. When I find out what, I'll let you know.

This comment intrigued me, so I went ahead and installed the latest version of Blender, just now. I don't see any button labeled "import". What I do see is the very standard File -> Import menu command, and that appears to work the same way as it does in every other program I've ever seen.

Where exactly is the button to which you're referring?

Josh Susanto wrote:

Others are able. But they don't do it. You are a kind of anomaly in answering, just as I am a kind of anomaly in asking. If this is conversation that 2 people could have had here months ago, why didn't it happen?

I'll admit that I tend to get a lot more involved in these things than most people. For whatever reason, I get enjoyment out of providing answers whenever I can. Not everyone does, obviously.

I also know that most people who read forums don't post. I'm not sure that clasifies those who do ask questions as "anomalous", though. It's interesting to consider.

As for why this conversation didn't happen months ago, I'm afraid the only answer I've got is because it didn't. It certainly could have, though. Not everyhting that can happen does (unless you believe in quantum reality, which, fun as it is for Star Trek episodes and whatnot, doesn't really help us here).

Josh Susanto wrote:

What's the video

There were two, actually (as well as many others that I have not watched, but which cover the same topic). The one I mentioned I'd seen previously was

. The other, which I found while searching again for that one was http://blog.machinimatrix.org/3d-creation/video-tutorials/installation/ .

Josh Susanto wrote:

why didn't the Blender website provide a conspicuous link to it if it's necessary in order to get their program to work?

A few answers on this one:

1. Python is not strictly necessary, just to get Blender to work. It is needed only if you intend to use scripts or plugins that depend on it.

2. The makers of Blender are hardly responsible to direct you to watch videos produced by third parties. They might not even be aware those particular videos exist. What they do have, very conspicuously on ther webiste, is an Education & Help section. Notice the very first link on that page is called Blender Quickstart.. That page in turn contains links to various tutorials, manuals, books, DVD's, etc.

3. It's open source software. You get what you get. The team that actively develops it is small, and they work for free. You're acting like they owe you something, when by rights, it is you who owes them. They've chosen of their own accord to give you the fruits of their labor, at absolutely no cost to you. You should be grateful for that, and you should not be so asinine as to complain that they didn't give you more.

4. I'm not them, and can't speak for them. If you want to know why they did what they did, and didn't do what they didn't do, you should ask them directly.

Josh Susanto wrote:

There was a gap between what was stated on the Blender website and what was necessary to say. How can you say there was not?

What's necessary to say for any open source application is, "Here's the program. Take it. Use it as you will. If you make any changes to it, share them." Anything beyond that is gravy.

You can call that a "gap" if you really want to. I call it "Be willing to do what it takes to learn, or just don't go there."

It's important to realize it took many years for the tutorial-producing community to accumulate enough experience, let alone produce enough material, to make Blender teachable to the masses with any degree of ease. Before that, if you wanted to use it, you had to be willing to dive into it yourself and figure it out.

That's the nature of open source projects. While commercial software companies like Autodesk employ technical writers with excellent communication skills to produce manuals and other educational materials, open source projects have to make due with the volunteers they have. Often the people trying to explain things are the very same engineers who built the software. That doesn't always go over since well, as engineers tend not to communicate in the same way as most non-engineers.

That's how it was with Blender for most of its history. People starting with the program today have no idea how good they have it.

Josh Susanto wrote:

I ask a lot of questions, yes. One reason is that most of them never get satisfactory answers, so it's sort of a numbers game with me.

Understandable, but that still doesn't seem to jive with your earlier statement that most of your successes come from NOT talking to people. It doesn't really matter, of course. I'm just saying, logical fallacy.

Josh Susanto wrote:

A decade after getting my MA, I'm still waiting for anyone in the Schenkerian analysis community to tell me how what their models model differs from the emphatic structure of a mnemonically optimized performance. I had to start asking this question because when I simply asked what their models model, they would just say "the structure", but declined to explain what the nature of "the structure" was, telling me that that should be intuitively obvious. What's intuitively obvious to me is that if they can't say whether or not their models model a specific thing which I have asked whether they model, they probably don't really model anything in particular (and now you know why I'm not working as a music theory professor).

On this, I'm in complete agreement with you. That crap is NOT music theory. At best it is for non-musicians who think they can make up for their lack of applicable emotional intelligence by focusing their intellectual intelligence into some giant mathematical equation that has nothing to do with anything. As an academic exercise, it's perhaps mildly interesting. It has no practical purpose, and it's therefore unsurprising that its proponents were unable to suggest one to you.

I'm not sure how this relates to what we've been talking about, though.

Josh Susanto wrote:

My commitment is to results. I lost a decade or more of my life to learning $hit..

The only way to get results is to learn.

Josh Susanto wrote:

In the Blender example, the results are not strictly from Blender, but from Blender in combintion with other things unstated. It's like selling someone a car without mentioning they have to buy the keys separately. I understand that Blender is "free", but after a buggy program costs me a certain amount of my time, I can no longer consider it to be "free".

I'd say it's more like selling a car to someone who doesn't know he's supposed to put gas in it. The buyer then insists the car is broken when it won't go. When someone tries to explain that cars require gas, the buyer then demands, "How come there wasn't a big sign at the dealership saying that?"

Josh Susanto wrote:

That's how Sculptypaint 092 works..

Apples and oranges. You might as well try to say that because someone might find one particular instrument intuitive to play, they should automatically be able conduct an orchestra.

Sculptypaint only does one thing. There's not a lot of room for misinterpretation. Blender does literally tens of thousands of things that Sculptypaint can't dream of. There just can be no comparison there.

If you want to discuss whether Blender is more or less intuitive than something in its class, like Maya, Max, Lightwave, etc., that's a conversation we can have. But trying to compare a full featured 3D platform to a one-note program, just because they both produce 3D content of some kind, just can't be done.

Josh Susanto wrote:

...Direct Delivery...Brooke...

Once again, I don't know what this direct delivery thing is all about, or who Brooke is. From what you've written here, I gather it's something you've been discussing somewhere on the commerce forums, that it has to do with some sort of upcoming improvement to the marketplace. I'd suggest you keep that conversation in its appropriate forum.

Josh Susanto wrote:

I'm the guy who removed his own kidney stone with a home-made tool after getting jerked around for a week by his insurance referral system.

Gruesome. I don't even want to begin to picture how that might have gone. I'm glad you didn't die.

Josh Susanto wrote:

I also sued my last US boss without a lawyer and was awarded more than asked for. It took over year to prevail, and I did so from a different continent. But I did it.

Nice work.

Josh Susanto wrote:

This tends to be the way I eventually get things done. Waiting for someone else to fix problems just because it's their job to fix them for me is expecting things to work the way they allegedly work, rather than how they actually work. And I have learned to expect otherwise, at least when it comes to dealing with my fellow Americans.

That's all well and good, exept when it comes to cases where things would work the way they allegedly work, if the user had a better understanding of whatever prerequisites allow it to do so, or cases in which the user lacks the experience to realize that it is indeed actually working. Both do happen.

Josh Susanto wrote:

They talked about it when I asked.

You asked millions of people? How long did that take?

Josh Susanto wrote:

One reason I couldn't get more help with Blender is that people I would have assumed were using it were instead using cracked versions of other 3D modeling software because they found Blender to be extremely frustrating in various ways. And these include some pretty significant builders. People whose work I'm confident that you've seen.

I don't want to go off on a tangent about cracked software. I'll just thank you for not using it yourself.

Josh Susanto wrote:

It wasn't working software. It was software that worked in conjuction with something else not provided.

Sure, the Blender download does not include a computer, an operating system, drivers, or any of the dozens of other things required to make any program work. It does, however provide what it, itself, needs to run, assuming you have all those other things already.

Back when it was componentized, what could be provided directly was right on the same page with the main download, and what couldn't supplied right there was linked.

If you missed something, you missed something. It's not a mark of shame. I don't see what's to be gained by continuing to deny it.

Josh Susanto wrote:

But there's plenty of ridicule in the merchants forum, for example. And other places.

That's a shame. I didn't think about the fact that negativity from other forums might well dissuade people from participating in this one. I probably should have. I see your point.

Josh Susanto wrote:

But there's plenty of ridicule in the merchants forum, for example. And other places.

I'm assuming the reason the first link didn't work is because of scripting, tied to your specific Google account login. The same thing would happen if I tried to give you a direct link to a message in my Gmail account. Although each message has its own URL, none can be accessed from anywhere except the main Gmail page.

No doubt, Google figures that if you need help with Sketchup, you'll go to the main support page, and follow the links from there.

Josh Susanto wrote:

I appreciate your help a lot. But it's ironic that in order to get something fixed that someone somewhere has already been paid to get correct for a year or more, I have to come on a public forum and gently antagonize someone I know is really on my own side. Don't you think it's at all ironic?

If we were talking about commercial software, I might agree with you, depending on the circumstances. But I don't think anyone anywhere is getting paid to diagnose your computer, and figure out why Blender might not be working on it (unless you're the one paying them to do that). The program is created by volunteers.

If you want them potentially to paid, click the Donate button on the website, and have at it. Encourage others to do the same.

Chosen Few · December 13, 2011

Josh Susanto wrote:

Do you have a better theory? The correct response to a bad theory is a better theory

Do I have a better conspiracy theory? Sure, aliens. It's always aliens. When they're not flying millions of lightyears just to stick a probe up some redeck's arse, they're taking away your sculpties. You know what a stolen sculpty is worth these days on Alpha Centauri Prime?

Oh, and Kennedy was killed by the phone company.

I don't need a theory, Josh, because I've already got the truth, mundane as it may be. It goes a little something like this...

Back in the day, a small startup company called Linden Lab wanted to sell servers. These weren't just any servers, mind you. They were the kind that are capable of running virtual world simulations. Well, in order to sell the servers, they needed something to demonstrate on them. So, they put together a little simulation they called Linden World.

It turned out there wasn't much buyer interest in the servers, but everybody thought the virtual world was pretty cool. So, that became the new focus of the company. Linden Lab was to be a software maker, not a hardware seller, and Linden World (eventually renamed to Second Life) was to be its product.

The company had its work cut out for it. This was late 90's/early 2000's, and the prospect of transmitting user-created 3D content across the Internet in real time was a daunting challenge, to say the least. One way to save a lot of headaches was to keep all the 3D objects entirely parametric, so that the only data that would need to be streamed was a small list of numerical values for each object, rather than a complete description of the geometry itself. Hence, content in Second Life was to be made of prims.

No one was exactly happy about that, of course. Obviously MUCH better content could be made utilizing standard mesh modeling techniques than by what amounts to playing with legos. But that was how it had to be, at least for the time being, until the state of all the relevant technologies would evolve to the point where it could be done right.

Fast forward a few years, and a guy named Qarl gets hired. In Qarl's previous job, while working on the Matrix films, he discovered a technique whereby 3D vertex coordinates could be mapped as channel values in an RGB image. When he began working on SL, he simply put two and two together. SL already knew how to import images and stream them to clients, and it already knew how to draw 3D geometry. All that was needed was a small wave of Qarl's magic programming wand, and now SL for the first time had a means of utilizing objects other than just parametric prims. Pretty slick.

These new objects, which we now call sculpties, were a really clever kluge, but they were far from ideal. Because vertices had to be mapped to pixels, and because images are rectangular, only perfectly rectangular topology could be utilized. Anything simpler was out, as was anything more complex. But again, it would have to do for the time being, until all the technologies would get to the point where proper 3D modeling could be incorporated.

Needless to say, a good portion of the community was super excited about this new option in SL. Even though it wasn't ideal, it was still going to be incredibly useful, a temendous improvement added to how things had been.

A very vocal subset of the population, though, was entirely opposed to it. They cried doom and gloom at every turn, about how evil sculpties were, and about how only very experienced 3D artists would be able to handle them, and it just wasn't fair, etc., etc., etc. There were even wild conspiracy theories about how LL would be taking prims away, because someone somewhere stood to benefit in completely inexplicable ways, should sculpties take over. Gotta love that hare-brained SL community. always a barrel of laughs.

But of course, prims didn't go away. Sculpties and prims have been happily coexisting for years now, and the conspiracy theorists have had no choice but to take aim elsewhere.

Fast forward another few years, and Qarl, along with a whole team of other graphics experts LL had hired for the task, finally pulled it off. SL was now going to be able to use standard 3D content (arbitrary mesh models), just like every other 3D platform in the world had been doing all along, and it was going to be able to stream it just as quickly as it already could stream prims and sculpties. It was about freakin' time!

Once again, the members of the community who understood what this meant were very excited. But just as before a vocal subset crawled out from under their rocks to scream about how evil this newfangled mesh thing was, and how only very experienced 3D artists would be able to use it, and it how it just wasn't fair. yada, yada, yada, And once again, there are wild conspiracy theories about how some mysterious 'smoking man' somewhere in a dark room wants sculpties to die because he somehow magically benefits from their demise, for reasons no one can possibly explain.

Just as before, the truth is nothing's going away. Mesh (which in a perfect world is what we would have been using all along), will coexist with sculpties and prims, no problem.

With the exception of Linden World, which was before my time, the above is what I witnessed first hand, in the eight years that I've been in SL. I didn't have to imagine or theorize any of it. I was there, and I saw it.

At the time I joined SL, there were only a few thousand residents, of whom only a few dozen were active. So, everybody knew everybody, and most of us knew most of the Lindens. We talked with them all the time, and so had lots of insight into, and input towards, how the platform was evolving. As the world grew, it eventually became impossible to know everybody, but quite a few of us who had made friends at LL maintained those relationships, as friends do, and we continued to talk.

I can PROMISE you, nobody cares whether you use sculpties or meshes or prims, for anything. The Lindens who worked so hard to bring us each of these things did so only so that we could have them at our disposal to use. There's nothing more to it than that. I know that's far more boring than imagining the CIA wants to abduct sculpty users, and sell their bodies for medical experiments, or that the king of Zamunda needs your sculpties in order to keep his goat herder son from marrying that peasant McDowel girl, but it is what it is.

Josh Susanto wrote:

Then why should LL bother to mislead people about the real benefits of each?

Nobody's misleading anyone about anything. If you're referring to the bug that Drongle discovered, it's most likely just that, a bug.

If anything, you should be happy about it, because it makes sculpties seem more render-efficient than they actually are. Also, the land impacts are artificially in sculptys' favor, which you've already state is a selling point of which you take full advantag. All signs are that if any particular object type is getting one hell of a promotional deal from LL, it's the sculpty.

Josh Susanto wrote:

Someone got paid to bring mesh technology who wouldn't have got paid simply to maintain support for sculpt technology. That, in itself, is no reason to imagina a conspiracy. But it does contradict what you are saying.

No, it doesn't contradict what I said at all. Yes, people got paid to bring mesh to SL, but no, those people's checks didn't depend on anyone actually using it once it was here. LL assumed (correctly) that it would be something people would want, so they hired a team to make it happen.

By the way, while I don't pretend to know all the details, appearances suggest that once the lion's share of the work was done, LL laid off those whose jobs were no longer considered vital (including Qarl), trimming the team down to just a skeleton crew. This, unfortunately, is commonplace among software companies.

Josh Susanto wrote:

What happened to flexi-sculpts? Were they shelved simply because mesh was "right around the corner"?

I could be wrong, but I don't think they ever got much past the "good idea" stage. As far as I know, they were never put into active development. If you want more information on the subject, you could attend a mesh team office hour, and ask.

Josh Susanto wrote:

After seeing the personnel files of hundreds of doctors, I can tell you that reality is nowhere near as straight forward as most people think. After that job, I have started to wonder whether I was really paranoid enough in the first place. If you're happy in your panglossian paradise, pray that you never get the jobs I got or see the things I've seen.

And here I thought I was kidding about selling people for medical experiments. I guess you know better.

Josh Susanto wrote:

There is no direct control, There is a mouse and a keypad. If you're using a glove, you might have mentioned that by now.

Way to deflect the point.

Josh Susanto wrote:

I retract the imputation of any ulterior motive, and I apologize for same.

Thank you.

Josh Susanto wrote:

But wouldn't the whole thing be more efficient if it didn't use sculpties at all?

Not necessarily. Vehicles have to be 31 prims or less, as you know. Not every desired shape for a vehicle can be made from so few. Plus prim builds are inherently geometrically inefficient, because they always have tons of hidden faces. A generic car shape can be made from just a few sculpty planes, and have few if any hidden faces.

Josh Susanto wrote:

Regular prims can provide an array of shapes almost as wide as what I see being used for sculpts on cars.

Regular prims can do anything, if you can use enough of them. If they were molecule size, and if you could use trillions of trillions of trillions of them, you could replicate the entire world, down the smallest detail. But with only 31, a lot of things are off limits.

Josh Susanto wrote:

I do intend to test it for myself. But as you can imagine, I don't necessarily trust the metrics as provided with metric tools provided by the same people who are providing the mesh technology.

So there's a secret agenda behind everything, even measuring devices?

Josh Susanto wrote:

I don't deny that, given every imaginable shape, mesh should be a better way to produce most of them. What I deny is that it's clear that this actually is true for a lot of common shapes. The person I trust most actually has given up on mesh; not because it's bad for modeling, but because the math has turned out not to be what it should be.

To what math are you referring, specifically? If it's just that your friend's meshes yielded higher land impacts than she might have liked, that's not an example of the math not being what it should be. Either she didn't do everything she could to model efficiently for a realtime environment, or she had inaccurate expectations of how things work. The latter would not be surprising, given that the impacts of sculpties have always been artificially lowered, while meshes will never be. I think a lot of people have been disappointed at the discovery of that fact.

Josh Susanto wrote:

It doesn't need to be said. If there's no reason to continue using them, then they will gradually disappear, and that's fine. The persistent theme is not that sculpts need to go. The persistent theme is that there's no reason to continue using them; something of which I'm not convinced just by some numbers that even mesh users cannot agree add up correctly.

Eliminating waste is not a problem for me. If there's a mesh version that's more efficient and just as good in other ways, that's what should be used. But a lack of consistency in what's reported in terms of the demands of meshes invites the question of whether it can really be stated with certainty that meshes are as reliably more efficient as LL wants us to believe.

If you have such a problem trusting the in-world numbers, then forget about them. Instead, make an honest assessment, based on the fundamental facts of the technology.

Fact: The more triangles in an object, the longer it takes to render.
Fact: An arbitrary mesh allows you to directly control how many triangles are in your model, while sculpties have fixed amounts.
Fact: Many of the shapes commonly made from sculpties do not need to have as many triangles in them as sculpty geometry forces them to have.
Indisputable conclusion: Many shapes commonly made from sculpties could be made far more efficiently as arbitrary meshes.

I don't need an in-world meter to tell me that a mesh plane made from two triangles is over a thousand times more efficient than one made from 2048 triangles.

Josh Susanto wrote:

It should be true, I would think. But there's enough lack of clarity about how great the benefit really is, that the whole idea that there is a uniform benefit comes into question. Mesh physics, for example, are assuredly more complex than sculpt physics. If sculpt physics are adequate for some purposes, then what's the point of putting extra demands on the system by applying mesh physics?

Your assumption is not accurate. Mesh physics CAN be more complex than sculpty physics, sure. But they certainly don't have to be. Just as with the visual shape, the physical shape of a mesh model can be made to be either more complicated or more simple than that of a sculpty.

The physics lattice of a sculpty is equivalent to that of a torus, which means it has 1152 collision faces. The physics lattice of an arbitrary mesh can have as few as one collision face, or as many as 130,000+. The efficiency or inefficiency is a function of decision making on the part of the creator.

In the example that I posted in response to Drongle, the physics mesh was just a flat plane with two triangles in it. That's over 500 times more efficient than they physics lattice of any of your sculpties, and it's more accurate to the visual shape.

Josh Susanto wrote:

I can only think of about 3 offhand.

And no doubt all three would be more efficiently made as meshes than as sculpties. Perhaps I should explain fully.

To go the sculpty route, you have to bend a plane into a square tube, "loaf pinch" the tube into sections, and then bend it into shape (like making a balloon animal), and you can only include so many planks. Then you have to waste all kinds of texture space, because although the pinched areas won't be rendered, they still take up as much UV canvas space as the unpinched parts. To get the maximum amount of planks from a single sculpty, you have to waste 20% the texture canvas. Finally, you have to surround it with an invisible cube to make it behave physically like a floor. Not only does that cube constitute an extra step in the construction process, it also increases lag, because the transparency requires an extra render pass.

If you're doing it as a mesh, on the other hand, you don't have to do any of that folding and pinching and bending. You can just create a single cube, scale it to the size of one plank, and then duplicate it to form all the others. You can even program the duplication variables so that all it takes is one click to spit out all the planks for the entire floor, all in their proper locations, and all at their correct sizes. You don't have to waste a single pixel of texture space, because the UV's can be set up any way you want. And of course, the physics can behave properly, so you don't have to bother surrounding the thing with a cube in-world, which means one less construction step, and one less render pass.

Josh Susanto wrote:

If the LI is essentially the same, though, why not use a prim oven one already has, rather than spend weeks or months trying to decipher new software?

As I said before, whatever amount of time you might spend learning the new software only has to be spent once, and then you can make all kind of things with it, forever. If your only goal is just to make one item, then you're right, don't spend weeks learning how to do it. If, however, your goal is what it should be, to make lots and lots of things, for years to come, then it's well worth a few weeks of your time to learn how.

Josh Susanto wrote:

Yeah. Customers don't matter. Phuk'em.

That's hardly what I meant, and you know it.

Josh Susanto wrote:

The formula that has worked for McDonalds is to do both. Keep providing new stuff to see if it's more profitable, but also refine the existing product line to approach some optimum cost/revenue intercept.

In used to tell all my marketing students this. McDonald's has the best marketing in the world. You cannot make food that bad, and sell that much of it, without good marketing. (Now if only they'd use their power for good instead of evil, wouldn't that be something! But that's another topic.)

Interesting story, when McDonald's first started up, people were flabbergasted at the lack of choice. "What do you mean I can only get lettuce, ketchup, and pickle on my burger? And what do you mean I can't decide whether I want it rare, medium, or well done?"

Nonetheless, McDonald's perservered, and their tale went on to become one of the greatest business success stories of all time. They told people what to buy, and people did.

Don't get me wrong; I'm not trying to say you shouldn't give your customers choices. McDonald's is obviously an extreme example. I'm just attempting to illustrate that it is you, far more than your customers, who has the power to control what they buy. Hopefully, you'll use that power for your customers' benefit, unlike McDonald's, who basically peddles death on a bun.

Josh Susanto wrote:

When they don't ask for anything, I build whatever I can.

Great.

Josh Susanto wrote:

Agreed. I also think when they have more of a chance to compare meshes directly with sculpt equivalents, the response will continue to be mixed, and in new, unanticipated ways.

It's gonna be fun ride.

Josh Susanto wrote:

Thanks for acknowledging this.

No problem. I'm only here to tell the truth as I see it.

Chosen Few · December 13, 2011

Josh Susanto wrote:

Another reason I won't be quick is that I just watched 4 videos right on the Blender website and I still don't know how to get data into or out of Blender.

I'll give you the same advice I give everyone when trying to learn this stuff for the first time. Don't approach it with "I want to do ______" in mind, whatever the blank happens to be. Forget all about whatever it is you think you want to do for now, and just concentrate on learning the program itself. Follow the beginner tutorials, allowing each lesson to build upon the last. Once you've done that, youll have a solid mastery of the basics of the program, and you'll be able to apply that knowledge to whatever it is you want to make.

In other words, don't put the cart before the horse. Try to focus too narrowly in the beginning, and it will only feel like a struggle the whole time. Broaden that focus, and the whole process will be much easier. The knowledge you thnk you're looking for is in there, along with all the other need-to-know information.

Josh Susanto wrote:

The reason I'm not puzzled is that if you look at the Blender interface as compared the interface for Sculptypaint 092, they're clearly designed by people who think very differently from each other.

Very true.

Josh Susanto wrote:

That is: Sculptypaint 092 is designed by someone who wants me to move data into and out of the thing, whereas Blender seems to be designed by someone who wants me to go repeatedly to the Blender website and watch videos.

Here's where your cycicsm may be starting to get in the way of your learning process. Blender is a MUCH bigger subject than sculptypaint. It's going to be a while before you've got a handle on the basics of it. The beginner videos are meant to walk you through all of the basics, in an order that is well absorbable. Let them.

If that particular video set isn't speaking your language, head on over the Machinimatrix. Thier stuff is produced by SL users, for SL users. I myself got a lot more out of watching their videos than any others.

Josh Susanto wrote:

So I'm to understand that knowing how to animate is more important than knowing how to get even simpler kinds of data in or out?

The ordering of the videos isn't in any way indicative of the relative importance of the topics they cover. All subjects are equally important, assuming your goal is to be able to do all that Blender has to offer. Some things do need to be presented before certain other things, though, in order for all of it to work.

That's not to say animation necessarily need to be covered before import/export, but do try to remember, the authors of those videos have no idea what in particular YOU are looking to do with the program. They have no way of tailoring it towards "Some guy who uses Second Life might want to replace some of his sculpty content with mesh content, so let's rearrange our entire Getting Started guide to focus on that." Their job is to provide the best, most universally applicable, overview of the whole thing that they can, so the widest possible range of users can get started with it.

Your focus is on SL, so of course, for you it makes sense that importing and exporting various types of data would be one of the first things you'd want to know how to do. For someone who wants to be film maker or a game artist, learning to animate would be much further up the immediate priority list.

With Blender, you're in the big leagues. It's not a one-trick pony like Sculptypaint. It has all kinds of different users.

So again, forget about whatever it is you think you want to know. Learn all the basics, and then apply that knowledge to your desired projects afterward.

Josh Susanto · December 13, 2011

>ehm Josh, you seem like a smart enough guy to not start with "basic tutorial 1" then "2 " then..etc... if you want to learn how to do something specific as importing an object. I'm not all that familiar with sketchup, but if you google "import into sketchup" you'll have better chances of finding what you are looking for than by following all the tutorials

Thanks.

1)

http://www.google.com.co/#hl=es&cp=20&gs_id=2d&xhr=t&q=import+into+sketchup&pf=p&sclient=psy-ab&source=hp&pbx=1&oq=import+into+sketchup&aq=0L&aqi=g-L1g-jvL3&aql=&gs_sm=&gs_upl=&bav=on.2,or.r_gc.r_pw.,cf.osb&fp=baed74643da2c96c&biw=1366&bih=667

Tge first 3 links don't much help.

2)

The 4th link says :"we will progess through the steps assuming you have a basic knowledge of sketchup tools"...

3)

Now I go back to Google Sketchup video #4, making a chair that no one has explained how to export.

As much as I make an art of complaining about things, I really am more interesting in making stuff. With new apps, I usually do skip tutorials and play with the interface to see how much I can learn without breaking the thing. THEN, if there are any tutorials I go through them both in order, and trying to skip ahead to what I know I will need.

It's been a year since I've even opened Sketchup, so I'm guessing I'll find the import and export tools if I look for them. But making a chair is more important? Really?

BTW: Does either Google or Blender employ anyone who knows how to pronounce "et cetera"?

Supposably, these videos are going to help me, though. Maybe I just need to axe the right questions.

Josh Susanto · December 13, 2011

>You'll find extrusion modeling to be quite different from what you're used to with sculpties.

Yes, that's what's good about it.

Or maybe.

What I'd like to be able to do is import some converted version of a sculpt product and subject it to some simple extrusion modification in order to make the geometric portions of the sculpt more of a benefit than a nuisance.

I certainly make a few sculpt spheroids, which I should think would be higher demand items if people understood the advantages, such as less inconsistent physics and the whole seam pulled to one single pucker being easier to hide, plus the ease of getting 3 appreciably different round rock planes by flattening the ones that have the pucker at a diagonal orientation to the bounding box.

But people seem not to want spheroids for some reason. Most planes are stupidly easy to produce (some viable images do require a few tricks), but what, exactly, to do with the frame has been a continuing question for me. Since I can't just get rid of the frame, I might like to take slightly more control over it than Sculptypaint easily allows. I say "easily" because there's quite a bit more control to be had if I really want it. But if the idea is ultimately to produce an architectural component with an photo-irregular face, I should be using a more architectural tool for the more architectural portions.

Josh Susanto · December 13, 2011

>Very interesting. I did not know about the rifle, and I had no idea the cotton gin wasn't actually his own invention. It's cool that you're a relative.

That was a long time ago. I didn't get any money out of it, and possibly not any genes, either. The story is that he put piles of rifle components on the White House lawn and asked the President to pick one component out of each pile at random to produce a rifle that would fire. And it did fire. This was the beginning of industrial machine tooling, as such. I think it also represent a shift in thinking from the simple question of component quality toward the question of component compatibility. Since that's a subtext issue in our current discussion, I thought I'd mention it.

>That's why I keep saying asking for help is the best strategy when you can't get something to work, and you have no idea why.

I ask people directly if they should know the answer and the relationship is such that it won't be strained by me asking repeatedly why the help didn't help. If there are simple answers, people who have them should have them. If there aren't, there aren't, or at least not yet while I'm asking.

>I don't think that's quite a fair comparison. In the music shop example, the work is totally out of the student's hands. There's nothing he or she can do directly, to facilitate the repair. That's not what we're talking about when it comes to correcting software malfunctions. In most cases, there's quite a lot the user can actively do to solve the problem.

Music students who have access to a keyboard instrument can usually learn to sing significant portions of wind instrument parts within a week, and they can often benefit more from that at least once in a while than simply moving fingers around while they push carbond dioxoide through a wooden or metal tube, so I try to see repair time as more of an opportunity than a crisis, at least in the short term.

I had a 30-day trial of that program that makes plane reliefs with the push of a button, but it ran out, of course. Had it not run out, I might have continued using that "superior" plane relief tool, rather than finding a better way to do plane reliefs by actively and deliberately controlling what goes into the relief axis from the RGB. I have now used things like edge detect function in some cases to get some of the relief data, and even separated that into postives and negatives by its RGB content. But doing it all with one button seems to be common among at least a lot of other people - that is; with one button or not at all. They've defined Step A as having software that does it with one button, and they've stuck to that.

>Where exactly is the button to which you're referring?

I shouldn't have said "button". If I don't know the proper term for such a thing, I tend to call it a button. When I completed the computer science portion of my music AA in 1990, everything was still being called a "button" for some reason, at least by my professors. In any case, the thing that says "import" allows me to put in the name of the .obj to be imported, but after that, I haven't found anything esle to do that produces any indication that anything has been imported.

>I also know that most people who read forums don't post. I'm not sure that clasifies those who do ask questions as "anomalous", though. It's interesting to consider.

From now on, if someone mentions a building challenge, I'll ask them to come here.

>There were two, actually (as well as many others that I have not watched, but which cover the same topic).

Interesting. Why can't the Blender people,themselves, make such a video, if they make so many others?

>1. Python is not strictly necessary, just to get Blender to work. It is needed only if you intend to use scripts or plugins that depend on it.

I didn't see any reason not to install it. Especially since, when someone asks "did you remember to install Python?", I can just say "yes", instead of getting into some kind of discussion about whether or not it's even necessary.

>2. The makers of Blender are hardly responsible to direct you to watch videos produced by third parties. They might not even be aware those particular videos exist. What they do have, very conspicuously on ther webiste, is an Education & Help section. Notice the very first link on that page is called Blender Quickstart.. That page in turn contains links to various tutorials, manuals, books, DVD's, etc.

Great. Where's the part that explains how to get data in or out? I'm still not seeing it? How many times will I have to click and where? If it's going to take all goddamned day, I might as well just ask someone, no?

3. It's open source software. You get what you get. The team that actively develops it is small, and they work for free. You're acting like they owe you something, when by rights, it is you who owes them. They've chosen of their own accord to give you the fruits of their labor, at absolutely no cost to you. You should be grateful for that, and you should not be so asinine as to complain that they didn't give you more.

I owe them nothing, because their service hasn't produced any value for me yet. OTOH, it has cost me hours of my life I'll never get back. Whatever their intentions, at this point, they owe me.

>4. I'm not them, and can't speak for them. If you want to know why they did what they did, and didn't do what they didn't do, you should ask them directly.

Good point. I'll do that.

>Understandable, but that still doesn't seem to jive with your earlier statement that most of your successes come from NOT talking to people. It doesn't really matter, of course. I'm just saying, logical fallacy.

I didn't say that I don't talk with people. I said that my successes come from not depending on them to get things right. When I interviewed for an intelligence position, the guys didn't even take the trouble to check a d*mn website to see whether what they told me was consistent with how the government was answering my questions elsewhere (it wasn't). Now you know why I'm not waterboarding people in Gitmo.

>On this, I'm in complete agreement with you.

Oddly, probably not complete. I use pitch class set theory, but only in conjunction with perception research. By definition, any 2 things that can be quantified or assigned parametric ordinal position have some kind of mathematical relationship to each other. But some mathematical relationships have extramathematical value and others do not. Being a good music analyst should consist of making good cases for and against extramathematical value in mathematical relationships between music data. If this were an accepted position in the community, I might still be part of it.

>The only way to get results is to learn.

OR the only way to learn is to get results.

>I'd say it's more like selling a car to someone who doesn't know he's supposed to put gas in it.

I know I'm supposed to put gas into it. Putting the gas receptacle apeture under rear passenger seat doesn't help me too much, though. And if the gas has to, instead, go into a completely different car, you can see why that might only be more confusing.

>You might as well try to say that because someone might find one particular instrument intuitive to play, they should automatically be able conduct an orchestra.

It's not a matter of intuition. It's a matter of things being labeled conspicuously according to function and being conspicuous about executing functions when applied.

>Gruesome. I don't even want to begin to picture how that might have gone. I'm glad you didn't die.

Thanks. I probably wouldn't have died. Their plan, apparently, would have been something more like waiting for me to develop something that could be covered as an emergency and actually treat that instead. Probably I'd not have lost more than half my **bleep** to the infection, although I suppose my bladder could potentially have burst as well. Inconvenient, but hardly an assured fatality.

>That's all well and good, exept when it comes to cases where things would work the way they allegedly work,

Blender's installation instructions were the allegation.

>You asked millions of people? How long did that take?

Sorry, no. Millions is an estimate, based on total global application exposure and the rate of complaints I've received from people who I know are absolutely capable of using programs of similar purpose.

>I don't want to go off on a tangent about cracked software. I'll just thank you for not using it yourself.

I've been given a version of zbrush that won't log in with the same password being used in Britain, a version of Photoshop that only works if I set my computer clock back ten years (disabling Second Life) and a version of Finale that will eat itself if I update it, and thus has bar line protocols essentially opposite to what I should think practically anyone would need them to be. More than half the operating systems, themselves, are pirated in this country. I dislike the idea of using cracked software if there is some other viable option, but my main reason for not using it is that it's generally no less buggy than what's free and legitimate.

>If you missed something, you missed something.

I did not miss something. The installation instructions missed something. If I had missed something, I would have missed it again this time. That it was not missed merely indicates that it no longer exists as similarly missable.

>No doubt, Google figures that if you need help with Sketchup, you'll go to the main support page, and follow the links from there.

Last year I had tried to find the help pages both from inside Sketchup, and from outside on a different browser. All I found was what you see in the first link. There are plenty of other really great URL's out there that I could use for all kinds of things, if I simply somehow pull them out of my a**. But until I somehow do that, or until someone else comes up with them for me, they may as well not exist.

>If we were talking about commercial software, I might agree with you, depending on the circumstances. But I don't think anyone anywhere is getting paid to diagnose your computer, and figure out why Blender might not be working on it (unless you're the one paying them to do that). The program is created by volunteers.

There is seriously no one getting paid anything? That's pretty weird. Even people at Wikipedia get paid.

As with other people who want donations, my policy remains the same; since I have no US accounts anymore and accounts in this country effectively require more income than I have, Second Life is the only place I can donate anything. If they'll take the trouble of creating any way to donate inside the SL economy, I'll both donate and suggest to others that they do so.

Otherwise, while I'm getting other people to cash out for me and wire the money for a commission, Wikipedia, Blender, the Long Now Foundation, and others all find it somehow too much work to create any way to receive donations inside Second Life. Go figure.

Kwakkelde Kwak · December 13, 2011

Josh Susanto wrote:

It's been a year since I've even opened Sketchup, so I'm guessing I'll find the import and export tools if I look for them.

You are complaining about a program without even opening it?! MY GOD! I haven't opened that program myself for years, but I am pretty sure the import and export functions are where they are in 95% or more of all programs. (File -> Export)

How do you try to learn a program? By watching all the tutorial videos without opening and trying out the program itself?

EDIT:

http://support.google.com/sketchup/bin/answer.py?hl=en&answer=114345

http://support.google.com/sketchup/bin/answer.py?answer=140409

Problem creating oblong sculpties

Recommended Posts

Link to comment

Share on other sites

Top Posters In This Topic

Popular Days

Top Posters In This Topic

Popular Days

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Please sign in to comment

Announcements

Linden Lab

Tilia

Second Life

Connect With Us

Partner With Us