York Jessop

You could bake shadows onto any size texture, if you used a 1024 texture with no baking on it, it would cost exactly the same as a 1024 texture with baking on it. I understand what your saying that to get better baking results you would have to use a higher resolution texture but when you compare say a house that has no baking on it to with the same house with baking the difference is astounding at no extra cost. If you were to enable shadows and lighting on the viewer, your 'costs' would go way beyond texture sizes.

Baking is something I just spent the last 2 days trying to understand too, in terms of actually doing it. After many many hours of fiddling with lighting, making lightmaps and all sorts of other technical hoo haa I finally understand how to do it! (go me) The purpose of baking objects is to add shadows to textures, instead of using the viewer shadows (which requires alot of computer powah) it allows you to give objects or environments (like a room with furniture in it) a much more realistic appearance by having a fixed shadow on the floor under a piece of furniture or by making corners in a room alittle bit darker than say next to a window. The great thing about meshes is that the 3D programs you use can do this automatically and give you a UVMap with these shadows already baked onto them, all you have to do from this point onwards is texture your UVMap. Heres a very quick example of non baked vrs baked: And heres one I refined to be better and more subtle And this is exactly what the texture/image itself looks like:

Thats a very good point about V2, I have no doubt there will be some who will not want to use it, but they are going to have to get used to the idea that V1.X.X viewers will not be supported for much longer. Its like those of us who refused to upgrade to Windows Vista/7 (I was one) and stuck with XP because its what I knew, but I had to take the plunge and now im as comfortable with Win7 than I was with XP. V2 is the same, its not perfect but it works and it gets easier to use as you get used to it not to mention when people start seeing mesh rice everywhere in the unsupported viewers (which looks like crap) they might consider the best option to be to just take the plunge. In regards to viewer shadows, I cant help but feel your contradicting yourself here, realtime shadows are a feature of viewer 2, but they also require a pretty hefty computer to render, many current games and 3D environments use baked shadows because they reduce the amount of resources that are need to render everything else, SL is more affected by this than most other 3D environments because there is an added process of streaming aswell as rendering. Baking a shadow add no addition cost to anything, if they are subtile enough you can even use them pretty well with shadows enabled! Regardless, baking shadows turns a product from meh to wow. People want realism and ambience, i thought being an RPer you would have wanted this too?

They did the same with Land Tier fees, didnt stop them and no doubt they will go up again at some point in the future. Imagine if they charged current land tier prices back when SL started up, they would never have got off the ground due to the collective pointing and laughing from teh whole internets. Meshes need some serious incentives to get off the ground, lower introductory upload costs should be an option (aswell as a blog clearly stating such a thing)

Ashasekayi Ra wrote: @Vivienne: Amen to that. However, I do think that Ishtara has a point that it will take some time for mesh to be viable from a commerce standpoint. I think it may take six months to a year to grow significantly and be adopted gridwide. I think like anything thats new and shiny, people will adopt meshes fairly quickly, especially when you see the quality that is possible with meshes compared to anything else thats been seen so far. Just yesterday I was skimming through some prefabs on the marketplace, I saw one for L$16000 and was gobsmacked by the price, but it had 5 star reviews and everyone who had bought it loved it, so I went in world for a look myself and while it was nice the only thing I could think was that a mesh version would be so superior in everyway. One of the main features of the house was how the shadow baking had been done and while it was okay - decent it was slightly pixelated and there were discrepancies in the 'fake' light sources. With meshes you can have programs like maya and blender bake shadows, occlusion and ambience from precise calculated light sources giving you fantastic lightmap to recreate some pretty stunning scenes, in addition instead of just standard boring shadows you can use nifty little options like ray trace shadows which give shadows a much more realistic effect. To give you an example:

Charlar Linden wrote: On the topic of upload fees, I like your point (Drongle made it as well) about people uploading for personal use or without a definite sales opportunity. Also someone pointed out that they want/need to do a lot of work in-world and that will require multiple uploads. We need to be cognizant of that with our upload fee. Preview ability was possibly the hardest compromise I made. We all know that being able to see things and manipulate them in context of the land and other objects is very very important to efficient content creation. I'm going to push for that functionality as soon as possible, but in the mean time you need to use Aditi early and often. As for upload cost, based on all the feedback we determined that we don't want the baseline amount to drive people to work around it by grouping uploads. Lowering the baseline during our rollout has also given us a lot of useful feedback. I want to set the expectation that upload costs will rise from the deploy amount, but we're not sure yet how much. Your well-reasoned feedback is welcome. Spew just means I have to scroll more and might miss a useful nugget. ;-) We're writing up some documentation to help explain the new terminology and what it means to them. This information might go out in the form of a wiki, or blog post, or perhaps I'll just walk up to each resident and tell each of them personally. Regardless, in the past you all have offered some good insight into how to organize it or what types of residents to focus on, so, again, I'd like to hear your thoughts. Charlar ps: It's probably not worth going into MMO specifics - and it's not germaine to this thread so I shouldn't have even commented on it. Thanks for the reply Charlar and sorry for the spew hehe. It would be nice to see clear cut documentation on the fee's and the reasons behind the costs. Your right about comparing MMO's it isn't the best example to use and I also understand the reasons why there needs to be upload costs but I like many others are concerned about 2 major aspects of meshes, PE and upload costs, its the 2 determining factors of how viable meshes are. As I mentioned above, the process from idea to implementation with meshes is a time consuming activity which Im hoping will pay off and provide some small rewards for all that work and in truth I have no idea how its going to work out for me personally, It may be a raving success where I can make enough sales to cover my in world costs and possibly own my own sim or it may be a total failure where the only thing I inject is time and money. Upload costs will add to the already high costs of tier rents, sales tax, and initial L$ investment. Im worried I may not be able to compete although it seems not many people are actually making meshes or how well it is going to be recieved by the rest of the population. As such shouldn't the current upload costs remain as they are for a few months AFTER it has been 100% deployed so that we can all have a better understanding of how its going to pan out?

A mesh file is miniscule too Charlar, not to mention MMO companies spend millions making content for thier userbase to actually have an environment to play in, in adddition they provide new content, usually for free unless its major content like an expansion, yet most expansion packs are tiny in cost to potential upload costs here and i wont even go into land fees. Then you have file hosting companies who let you upload 100MBs for free, mesh files are usually a few hundred bytes for goodness sake. However, its good news that it wont go up to L$150 baseline but that wont make any difference what so ever if mesh costs go up, the chances that people will be hitting the baseline whatever it is will be a small minority of creators, most mesh uploads will probably cost more than that anyways. Current prices right now are acceptable, ive been paying on average L$80 to upload my small/medium house, i expect if prices stayed as they are, i would on average be paying L$150-200 per mesh, that is also acceptable, but you start going into the L$300 and above im gonna start thinking twice about uploading anything, especially if it doesnt sell, this is the risk you need to consider, its one thing uploading a mesh with a guarenteed sale to cover costs, its a completely other situation if those meshes never sell or dont sell enough, you start to weigh up the amount of time and effort your putting into making a mesh (which is a considerable amount of time) compared to the potential of it even being worth it, or in a worst case scenario, actually costs the creator to create for nothing.

Ok here is why not using the 3 lowest LODs is a bad idea The following pics show: Red house/triangle has only the highest LOD, other 3 are set to lowest settings (4tris) (35PE) Green house has 1 highest LOD and 1 High LOD with other 2 set to lowest settings (4tris) (50PE) Blue house has 1 highest, 1 high and 1 low LOD with lowest LOD set to lowest settings (4tris) (50PE) (<--best option imo) Pink house has 2 highest, 1 low and 1 lowest (79PE) As you can see in that last pic, I had to be practially on the steps before it wasnt a triangle, ask yourself this, if you were looking for a house to buy, which one would you be happy with?

weird, seem like PE scaling has been turned off on the beta grid EDIT wow well this is really weird, if all 4 LODs are the same model, scaling doesnt affect your PE (possibly just my house) however when i used my 4 different LODs the scaling went weird, at max scale it was 78PE (same as standard size) but if i make it half the standard size i get 79PE lol, then the PE drops significantly as it gets smaller.

So far I must admit ive been pretty happy with medium scaled meshes, my house is 25x35m and its an exact copy of a prim house, the prim house is 168 prims, the mesh model is 80PE, this is with custom LODs and a custom physics shape so I really cant complain, but I have no idea whats a full 64x64 mesh would come in as so its hard to judge.

LL doesnt compare scuplts and meshes, sculpties were a mistake 3 years ago but because of the LL policy of not breaking legacy content they will be a permenant mistake, they are intending not the make the same mistake with meshes, infact they are probably going way overboard with them in terms of PE and especially L$ upload costs, I was just explaining how LOds are calculated from the explanations the Lindens have given us. If you pretend scuplties dont exist like LL does, then it makes perfect sense

Im not sure how the calculation works when it determines scale vrs PE but i think it should go something like this... 0m-20 x1 cost 20m-32m x1.25 cost 32m-64m x1.5 cost

Moo Spyker wrote: that is my point though. what is the point to use the lower LOD when it charges you more L$ and more PE? its pointless. the PE should be based off only the highest LOD. Because I will not add 4 LOD's for a over double increase in prim count... Because if you intend to sell your mesh models then your customers will expect quality work with proper LODS, not some crappy model thats only visable if you have a top notch computer capable of rendering the highest LOD from a decent distance, there was a post by a Linden recently that had some demographics about what kind of computers residents have, 33% of them use computers that are basically 6 yrs old, something like 5% use top end computers, the question would be, do you want to sell your models to 5% of the population or 100% of the population. PE is important there is no doubt about that but quality and actually being able to see your mesh from more than 10foot away is probably more important. In terms of high PE and L$ upload costs based on the lower LOD is because this LOD will be the first to render and from the furthest away. Many mesh makers actually use a billboard for the lowest LOD (a 2D plane) this helps alot on the PE costs and doesnt cause too many problems for people with old computers as the next LOD up will render fairly quickly and from a decent distance anyways.

Rusalka Writer wrote: Anyone trying to avoid mesh should never have joined SL, or any other virtual world. Mesh is everywhere. Prims, sculpts, everything has an underlying mesh to it. The mesh program going through in SL just gives creators more control over the shape of meshes and the appearance of objects. With luck and skill, the mesh objects generated will make SL more efficient and interesting for everybody. I really have to agree here, even the ground is made from a mesh, to say that meshes will force anyone to leave SL is naive. In regards to the difference between a scuplt and a mesh, a scuplt is a cylinder that has been unwrapped, but it will always be a cylinder and it will always have a sphere for a physics shape, a mesh has no such restrictions and can have a physics shape that is as complex as the mesh model itself.

The highest LOD has the least effect on the PE, the lowest LOD has the highest effect, the problem creators would have if they used the minimum possible tris for the 3 lower LODS is anyone who has thier object settings on low-medium will most probably never see the highest LOD until they were right on top of it, making thier purchased mesh model a very expensive triangle. Its in a merchants best interests to use decent LODs so that thier product will appeal to everyone and not just a small percentage of residents who have super computers.

Alisha Matova wrote: I wish I was more familiar with Maya. Ruling out lod and viewer issues does point at your modeler/exporter. Could there be some loose verts? Or maybe a camera or light sneeking into your .dae? Does Maya have any "export only selected" tricks... If not, though it sounds like a pain and dangerous, try selecting Every face or vert you know you want to export, then use select inverse(to select everything but your model) and delete. This will catch any loose verts, cameras, lights, etc. No idea if this is a viable answer for maya, but is a schnazzy Blender trick. Ive no idea what the problem was but it is possible the face generation tool in maya may have caused the issue. When your using Maya to make faces between border edges you can use a very snazzy tool called Append to Polygon, which when clicked highlights the border edges, you just click on each edge until you have filled in the gaps, however you can basically click a heap of edges and it will generate the tri's automatically. This time I did the tri's manually, as in only clicked 2-3 edges each time i used the tool and repeated the process until the ceiling and floor were filled in. It has worked, all the overlapping textures have now gone! No disfunctional tri's in sight!

That was a great suggestion, i just reuploaded the model using the highest lod as all 4 lods to test, but unfortunately they are still there, its like the upload has generated 3 double faces between certain verts, very strange, my only other idea is to delete those verts in maya and rebuild it and see if that works, I was just wondering if there was something causing it or if its possibly a bug in Maya or the OPEN collada export. My main concerns it having the issue repeat itself in other meshes i make. EDIT: Decided to delete the faces around the problem areas and remake the faces, thanksfully its worked! And even better is i didnt have to remove the verts, so its an easy quick fix, still no idea what caused it but the solution is pretty painless Thanks anyways guys

Thanks for the reply, in the maya screenshot i posted, face normals are already turned on and pointing the the correct direction. Also, those faces in the screenshot are doublesided, they show through on the other side which is what causes the textures to flicker.

Yes of course, i know if you did uploads individually it would cost a huge amount more based on the L$150 min, I meant if you uploaded a 15000 prim mesh scene (highly unlikely i know) but even then its a good point, $60 would be the lowest cost if you actually did that. Which is just rediculas. If LL does this with the L$150 min and L$1 per 1PE it will be one of the most expensive upload costs on the internet lol, lets face it, 100's of websites let you upload and store your files for free, if you subscribe to an MMO for $15 per month, you effectively store ALOT more bytes than any mesh ever would.

Im currently running through my first mesh and so far its going very very well, however Im having a slight problem I just cant seem to find an solution for. This is the problem: The light gray tri's are basically faces that overlap the actual ceiling faces (they actually seem to be doublesided faces), ive checked in maya and there are no faces there, or double verts or face normals the wrong way, they just dont exist, but in SL they do, the same problem is present on the floor in a slightly different place, the problem it causes is these tri's cause texture flickering. This is the same view from inside maya: Floor: Anyone have any ideas what could be causing it?

Well I'd like to appoligise for this post for I have a strong opinion on charging content creators for profits..... Really dont understand why it would need to be L$150 minimum, I know it was like this before but if it goes live as that its going to cost silly money to upload stuff, the cost of uploading a mesh shouldnt even be a significant factor, it should be a minor afterthought to the complexities of creating meshes. The lower limit should remain at L$10 but increase the scaling costs 'slightly' with a max and i mean serious max (like 2000PE max) should be L$500. At L$1 per PE thats just insane, if you decided to upload meshes to fill an entire region, LL wants $60 of your money for a privilege of adding content to SL, the only reason why anyone is here at all. What I really dont get is Linden seemingly unable to comprehend that new mesh content will stimulate the ingame economy, providing extra incentives for new residents to join, old residents to return and all residents to stay, this results in more marketplace sales (sales tax) more L$ buying (transaction fee) more land rented (sim tier fees) and more customers for everyone, so why are these factors not taken into account when some genius decided that mesh uploads should cost content creators some real $$$. They say its to cover costs, pffft .. its to make a cheap buck on the backs of content creators.