The infamous match has apparently lost its claim to fame

[Chic Aeon]

When the Firestorm toolset came out awhile ago (end of Feb 2018 my blog reminds me) there were several threads of the OMG what heavy mesh you have variety -- who knew?  No names of course but many of us will remember the infamous match stick and match stick head :D.  I doubted that could be beaten but there is a new king of the hill it seems. There may be others, I just look at the land impact and go "no way" and delete.  

 

So here is my new find for the heaviest mesh I have seen. This is not as small as the matchstick but it is way less than half a meter in any direction.  

A lot of Bellisserians lament over our "measly" 351 land impact points. Me?  I have found them more than adequate and just today did some rearranging and redecorating and added quite a few things to my kitchen and dining room AND retrieved some LIs (I currently have 11 free). 

 

Anyway, for those folks wondering what to do to get more for their LI buck, the answer is fairly simple, buy well-made GAME ASSET MESH. 

Here is my screenshot which tells the story if NOT the product

Edited June 5, 2019 by Chic Aeon
spelling

[Coffee Pancake]

WOW

 

[Sasy Scarborough]

I have seen things that when I tried to see how dense, were so dense I could not see...two more posts after this one woohooo

[Kyrah Abattoir]

I love it when the tricount is so far from the complexity. When you use SL default settings you almost never get a complexity that is below your tri count.

So we all know what's going on here at the lowest lod.

[steeljane42]

Here's one shelf (it's not even usable or has any opening/sliding parts, just a kinky static prop in pretty average size, around 40li I believe) that I wanted to get before I've seen the numbers. And I can't say that I care about it too much, I do have quite a few decor/furniture items between 100k and 200k triangles, a couple around 250k each. But some things are just too much, like that one, so I just don't get those, despite liking them. My region is far from the "low-end PC friendly" ones, but I do actually use it for more than just taking static pictures where it wouldn't matter, so I do my best to avoid such extremes.

P.S.

There's also a very cute "food set" (to avoid names) at fameshed right now with this kind of crazy numbers (props with just 1-2li, but together they are over 440k triangles). Yep, also one of those examples that I'd get otherwise, but did not, purely because it has inadequate amount of triangles to what they are. I have not worse looking food props with just 8-15k triangles that look good even zoomed up fully.

 

[Kyrah Abattoir]

"Quality"

[OptimoMaximo]

8 hours ago, Kyrah Abattoir said:

"Quality"

You're forgetting a keyword... "High Quality"... and i'd add "Life-like" 😁

[MaKaylaa93]

Do ppl just not know they shouldn't make these kinds of thing?  I'd never thought before I started reading this forum that I should even check numbers before purchase.  Is it ignorance or apathy?

[Coffee Pancake]

1 hour ago, MaKaylaa93 said:

Do ppl just not know they shouldn't make these kinds of thing?  I'd never thought before I started reading this forum that I should even check numbers before purchase.  Is it ignorance or apathy?

It's a combination of 

• I found / ripped this model, I don't care.
• I bought this model, I don't have the skill to make it suitable for SL.
• I exported this model from a specialist utility application, I don't have the skill to make it suitable for SL because it's a totally different skill set and it would add significant time to my workflow and .. 
• This model started life as a 3d sculpt, what are these 'tri-angles' and 'vert-eye-cees' you speak of ?
• This is an attachment, it should be full quality always so I don't even have to try ... 
• This is for indoor use / gatcha, the high detail model will be a bomb, all the others will be set to zero. See it's 1li ! That's good, right?!
• This was made for an event, with tight deadlines, think yourself lucky it even comes with a monochrome texture.
• Blender hard. Blender tutorials hard.

 

[Kyrah Abattoir]

You forgot the "yardstick of good work".

AKA: It sells well.

[Coffee Pancake]

Oh yes, the market will decide ! because consumers always know what they are buying before they buy it, and have the data & skills to make objective decisions.

[Wulfie Reanimator]

THIS 👏 NEEDS 👏 TO 👏 STOP 👏

Edited June 9, 2019 by Wulfie Reanimator

[Coffee Pancake]

You're never going to be able to stop someone linking 56 ++ objects together and cooking up a monster

[Wulfie Reanimator]

11 minutes ago, CoffeeDujour said:

You're never going to be able to stop someone linking 56 ++ objects together and cooking up a monster

Certainly, but that doesn't mean I have to be accepting of it. (And I won't.)

For context, that was a single, unfurnished mesh skybox with materials and normals (272 faces, 453 textures).

People shouldn't do it. They will, but they shouldn't. It's absolutely horrible and a crime against limited resources.

[Chic Aeon]

15 minutes ago, Wulfie Reanimator said:

Certainly, but that doesn't mean I have to be accepting of it. (And I won't.)

For context, that was a single, unfurnished mesh skybox with materials and normals (272 faces, 453 textures).

People shouldn't do it. They will, but they shouldn't. It's absolutely horrible and a crime against limited resources.

I think that is is VERY likely -- looking at the statistics -- that the skybox was made with prims and then exported to DAE and NOT CLEANED UP let alone optimized.   That is a completely different story than "heavy mesh" simply because it really isn't a 3D model at all, just a hack.   

 

I absolutely agree that it shouldn't happen.  It is terribly messy and I have said my piece on that many times.   The folks doing that "export to mesh" thing don't know what they are doing so in their cases it is just lack of knowledge. In the cases of many well known brands,  there are other reasons --- none which are acceptable to me. So I simply don't buy and don't use those products -- and my sim runs much better for the lack of them :D.  

 

Understand your frustration though. 

 

[Wulfie Reanimator]

35 minutes ago, Chic Aeon said:

I think that is is VERY likely -- looking at the statistics -- that the skybox was made with prims and then exported to DAE and NOT CLEANED UP let alone optimized.   That is a completely different story than "heavy mesh" simply because it really isn't a 3D model at all, just a hack.   

I absolutely agree that it shouldn't happen.  It is terribly messy and I have said my piece on that many times.   The folks doing that "export to mesh" thing don't know what they are doing so in their cases it is just lack of knowledge. In the cases of many well known brands,  there are other reasons --- none which are acceptable to me. So I simply don't buy and don't use those products -- and my sim runs much better for the lack of them :D.  

Understand your frustration though. 

I'm aware of the prim-to-mesh process, but the skybox didn't look like the end-result of that. The construction was consistent, there was no "prim-like" wireframe like you would get by exporting, it included shapes that are not possible with prims and a clearly separate physics shape from the model itself. There was definitely some kind of skill on display there.

But I digress, it does not matter what the exact reasons are for creating a product like that. I would totally understand if it was a bunch of someone's personal objects linked together for convenience, but that wasn't the case here. It was an "as-is" commercial product, that's not okay under any circumstances and that was the point I was trying to make in my first, in my opinion obviously exaggerated post.

Edited June 9, 2019 by Wulfie Reanimator

[Penny Patton]

On 6/9/2019 at 12:19 AM, Wulfie Reanimator said:

People shouldn't do it. They will, but they shouldn't. It's absolutely horrible and a crime against limited resources.

People will never stop, and the problem will continue to get exponentially worse, unless LL finally puts a foot down on excessive resource use. I understand that they don't want to destroy existing content or upset the userbase, but there are ways they can approach the issue while minimizing the negative fallout. Like introducing new resource caps but linking them to new features, or content created after a specific date, etc. Then let new content phase out the old content naturally, like it always does.

 Communicating best and worst practices to the userbase, through like an official content creation blog, and encouraging better made content through contests and free advertising, would help smooth the transition as well.

 

[Penny Patton]

Actually, I'll add to this. LL should do the above, but it will take time to both implement and then more time before we see the benefits. So, in the meantime, here's what LL should be making a priority:

Communication and education.

When I'm experiencing texture thrashing, SL should pop up a message explaining why that may be. We already get messages telling us when HUD attachments use too many textures, take that further by tying cause to effect. Most people ignore the "your HUD attachments may be using too many textures", because they have no idea what that means. So how about when the  viewer sees that a lot of texture thrashing is occurring it pops up a message:

You seem to be experiencing a lot of texture thrashing. Here are the probable causes for this problem.

• Your HUD attachments appear to be using 52MB of textures. Click here to see the texture use of individual HUD attachments and try to bring that under 10MB.
• Your avatar appears to be using 278MB. Click here to see the texture use of  your individual avatar attachments and try to bring that under 100MB.
• Your surroundings are using an excessive amount of texture memory. Click here for a list of objects you own and their memory use. Try to reduce the amount of texture memory as much as possible.
• 5 of the rendered avatars around you are using over 100MB of texture memory each. Click here for a list of avatars that may be impacting performance. Click here to open preferences and adjust your rendering caps.

Similar messages could appear for triangle counts, script use, etc. All of the common SL performance woes and their causes. Also I'm totally pulling these numbers out of the air as examples, LL should put some thought into this and figure out the most appropriate goals.

LL should add texture memory and triangle counts to marketplace listings, and encourage content creators to use it by encouraging everyone else to look for it. Listings that don't include such information should come with a warning about the performance pitfalls of content that uses excessive resources, and a link on how to inspect their attachments and rezzed objects.

On top of this direct approach to explaining performance issues to the userbase, LL should be working on that content creation blog right now. Loads of tutorials showing content creators how to make better looking content using fewer resources. Holding regular contests to reward well made content. Pushing the importance of optimization every way they can, in ways that make sense to the average SL user.

If they do this now, then however they propose to deal with performance killing content in the future, they should be able to introduce those changes at least a little more easily.

Edited June 13, 2019 by Penny Patton

[Coffee Pancake]

The solution can never involve shoving technical terms and scary numbers in people's faces, or even targeting end consumers. Eyes glaze over and the numbers take on a touchy-feely quality and an entirely useless personal meaning for that individual. Showing avatars ARC is a great example of why this fails.

Throwing around numbers of MB is equally useless. Unless they have a technical interest, their only prior exposure will be in a retail "bigger is better" context. You can't expect people who turn on a 'magic wonder box' that lets them play games, including SL to even care. Ask someone what specific component they could change in their own computer to make SL better and they will either pick the GPU or have no idea.

4 hours ago, Penny Patton said:

people ignore the "your HUD attachments may be using too many textures", because they have no idea what that means. So how about when the  viewer sees that a lot of texture thrashing is occurring it pops up a message:

 

So rather than a specific message about items on your own HUD, you think the situation will improve if the viewer could start naming and shaming nearby objects and people.

 

[Penny Patton]

The main goal is to just raise a general awareness of what causes performance problems and it's not that different than understanding "Land Impact". If you have another way, a better way, of conveying this information to users then great. If not, well then this is still the best idea on the table, isn't it?

ARC fails because it's both too far removed from what actually causes performance issues (You can have a very low ARC score and yet be the single biggest framerate killer on the grid) and it entirely fails to communicate what that ARC score is supposed to mean, as in how it is affecting you, the user. It is not tied to any specific performance problems in a way people understand.

ARC also has no teeth. It needs to be like Land Impact. A set pool of points people have to keep under. Better communication regarding the problem is just one part of the solution. Introducing (slowly and deliberately so as to avoid the content apocalypse people tend to assume will happen whenever the topic of resource caps come up) better LI calculations and some sort of avatar version of LI, is another part.

31 minutes ago, CoffeeDujour said:

So rather than a specific message about items on your own HUD, you think the situation will improve if the viewer could start naming and shaming nearby objects and people.

First, no. I'm saying the information about your own HUD, avatar, what-have-you, should be more specific and more clear. But you knew that already because it's what you complained about in your first two paragraphs.

As for "naming and shaming", you might have a point. Maybe. LL needs to tackle this problem somehow, if it's better to focus on what each individual user owns, then maybe they should do that. But if the environment or other avatars are the source of your problems, then people have a right to know what the problem is and how to fix it. We shouldn't be able to create framerate killing avatars in the first place, but that ship sailed long ago, didn't it? It will take a lot more to fix it now.

But let's say we avoid "naming and shaming", fine then. Users should still get a message saying if the environment or avatars around them are the problem and be directed to some solutions, such as adjusting your ARC rendering cap, avatar imposters settings, object detail settings, etc.

 That said, people complained that the grid would be torn asunder from the "witch hunts" brought on by ARC when it was first introduced and yet that never happened. So maybe, just maybe, this concern is a bit overblown. But, hey, that's why it's good to discuss these things.

Edited June 13, 2019 by Penny Patton

[Kyrah Abattoir]

I'mho just put a hard limit and you get a helpful little progress bar that tells you how far from the limit you are.

Done.

4 minutes ago, Penny Patton said:

That said, people complained that the grid would be torn asunder from the "witch hunts" brought on by ARC when it was first introduced and yet that never happened. So maybe, just maybe, this concern is a bit overblown. But, hey, that's why it's good to discuss these things.

If ARC hadn't been completely defanged...

Edited June 13, 2019 by Kyrah Abattoir

[Alyona Su]

I will go ahead and presume this is among the reasons why Linden Lab was absolute *against* the ability to import the mesh into SL for *years and years* until someone finally cried "Uncle" to appease the incessant shouts by some creators. It's not only what this thread is about, but also region-designers who are over-meshing places. Frame rates are massively affected (and I have a pretty powerful machine) - and the moment I cross a border into a better-designed region, frame rates instantly skyrocket, turn around to look back at the mesh-jungle and ...yeah.

I'll take good old-fashioned, well-built PRIM anything before I take Sculpty, and I'll take well-built Sculpty anything before I take mesh if I can help it. And this thread basically validates my belief in that process. I'm not a technonerd, most of this stuff flies altitudes over my head, but I'm able to make sense of the gist.

Shameless plug: Thank you Catznip viewer, performance outshines every other and makes the mesh-cesspool of the SL grid survivable. Srsly.

Edited June 13, 2019 by Alyona Su

[Kyrah Abattoir]

18 minutes ago, Alyona Su said:

I will go ahead and presume this is among the reasons why Linden Lab was absolute *against* the ability to import the mesh into SL for *years and years* until someone finally cried "Uncle" to appease the incessant shouts by some creators.

It also limited the copyright issues of the grid too since content in SL could only be made in SL (excluding sound/tex/anims obviously)

[Coffee Pancake]

1 hour ago, Penny Patton said:

That said, people complained that the grid would be torn asunder from the "witch hunts" brought on by ARC when it was first introduced and yet that never happened. So maybe, just maybe, this concern is a bit overblown. But, hey, that's why it's good to discuss these things.

People are lazy, LL didn't make add a script function so we could repeat the script memory debacle all over again.

If this were my call, I would explore junking the built in mesh upload decomposition, base Li off the raw top detail mesh & scale, and enforcing in viewer LOD creation following a logarithmic drop off target, unless the creator could manually specify equivalent or lower triangle model. In combination, change the way mesh assets are referenced grid wide enabling a creator to add hand made LOD meshes to an original uploaded asset and have that change instantly reflected in all instances of that mesh in use. 

The problem can't be placed on the shoulders of end users in the hope creators will take note. They wont, their workflows are not conducive to the kind of content SL needs and handing out pitch forks only gets us poking each other in the eye (while remaining fantastically dressed).

 

[Penny Patton]

1 hour ago, CoffeeDujour said:

If this were my call, I would explore junking the built in mesh upload decomposition, base Li off the raw top detail mesh & scale, and enforcing in viewer LOD creation following a logarithmic drop off target, unless the creator could manually specify equivalent or lower triangle model. In combination, change the way mesh assets are referenced grid wide enabling a creator to add hand made LOD meshes to an original uploaded asset and have that change instantly reflected in all instances of that mesh in use. 

That doesn't sound like a bad idea, if feasible. I'm not sure I'd tie the object's LI to it's original scale. People would just upload things smaller in that case to get lower LI, then make it huge again after uploading.

1 hour ago, CoffeeDujour said:

The problem can't be placed on the shoulders of end users in the hope creators will take note. They wont, their workflows are not conducive to the kind of content SL need

I agree. I've always said as much. The problem is, you're effectively suggesting information be withheld from creators because it will also be available to the end users and that might hurt someone's feelings. Not a good strategy. LL needs to be forthright about this kind of information. Not everyone may be able to make use of it, but the information needs to be out there for those who can. And if LL can make it instructional so more people understand it, all the better. Even if we limit the scope of it towards content the user owns, and the settings they need to adjust to deal with content they do not own, that's better than nothing, which is what we currently have.

 It's also worth noting that we still have sims that check people for script use. It has not resulted in the end of the world. The outrage and indignation over these things tends to be short lived.