Jump to content

Mesh Experimental Regions PE is MUCH better


Vivienne Daguerre
You are about to reply to a thread that has been inactive for 3360 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

I cannot speak about larger builds at this point, but I just rezzed my kitchen stuff in Mesh Sandbox 28 which is one of the experimental server regions after reading this from Drongle:

"Make sure you try them out on the Mesh Experimental server*, which has the diameter-instead-of-radius bug fixed. The cost rise at half the rate with increasing size, although they eventually reach the same plateau. It won't help with very big things.

*Sandboxes 20,21,27,28."

The prim equivalency was much more reasonable, and if mesh went to final release with those numbers, I would be content.

I have not rezzed out my townhouse yet or played with larger objects, so those having larger objects might want to try it out and post their results here for the benefit of the rest of us (please).

Link to post
Share on other sites

I uploaded my townhouse with the mesh walls on the experimental mesh region 28. I made the LODs all the same, for it is low poly, only 468 triangles. It was not behaving well LOD wise if I made any of those lower. Being a building, it must keep its shape at greater distances. The walls came in at 11 prims, which is acceptable. As a coalesced object, with the prim and sculpted parts (as well as moving parts) not linked to the mesh, it was an acceptable total of 93 prims.

I then linked the non-moving sculpted and prim parts to the mesh, and then the whole townhouse came in at 104 prims, which is better. I think it should be the same, but it is not as bad a difference as before. Given that, the savings of 11 prims is significant to the customer and I would package it with the mesh not linked to anything.

Given the original made of ordinary and sculpted prims on the main grid comes in at 149 prims, the mesh version is significantly better in terms of prim count.

Townhouses on Experimental Region.jpg

Link to post
Share on other sites

Doesnt sound too bad but I still think they need to be more forgiving on larger scaled objects, that townhouse I personally would be happy with it being around 50-70 prims without major LOD differences, its not furnished and its very basic in shape, meshes were suppose to allow us to have some details rather than flat square prims, when we start making our designs I think we are going to find that we have to spend 4x the amount of time trying to get it down to usable PEs at a huge cost to detail.

 

Although its certainly a step in the right direction another 25% off would imo go a long way towards making mesh a sucess.

Link to post
Share on other sites

Yes a much lower PE cost. ^_^

I have two rooves, which were 26 PE and 6 PE (1k4 and 0k2 tris repsectively) both approx 28x18x5 metres with a custom physics shape.  When rezzed in MSB#21 dropped to 8 PE and 2 PE respectively.   This is looking much better for building now.

Link to post
Share on other sites

Yes!  Much much better PE today!  Had 64 meter mesh that was 1070 prims now 979.  Two smaller meshes one which was a dolphin which was around 30 now 10 and an small space craft that was about 50 prims now a much better 20 prims.

 

Let's hope PE stays at this level or crosses her fingers in hopes it may come down a bit more.

Link to post
Share on other sites

OK, following items are now lower prim count:

5 meter palm (900 triangles) which used to be 9 prims is now 3,

Grand piano (7200 triangles) which used to be 25 prims is now 19,

32x32m terrain segment (625 triangles) which used to be 15 prims is now 12.  I can link a box prim to it for no added cost, while linking a torus gives an absurd PE cost.

Single mesh items still have a factor of 5 penalty for medium items relative to sculpts, when counting triangles/prim, and worse at larger sizes.  When linked to a box prim root and suitably sized you can get the penalty down to 2, which would make them barely competitive: a higher prim cost, but better UV map, physics, texture count, and vertex accuracy.  If I was doing something medium to large, I would look at using sculpts for the visible details, and a mesh purely for the physics, with a low triangle count for the LODs.  That might give an interesting result.

Link to post
Share on other sites

Here is my optimized sword (left), my first attempt from yesterday(middle) and the Sculpted Prim variant (right)

The sword length is ~93 centimeters.

excalibur_mesh4.png

excalibur_mesh5.png

 

The closeup is from the optimized mesh sword:

excalibur_mesh6.png

 

And actually i found that i made a mistake in the LOD0 of the mesh-Blade (easy to fix):

excalibur_mesh7.png

 

But actually not realy important:

excalibur_mesh8.png

And now i get back from Sword Escapade to Tutorial making ;-)

 

Link to post
Share on other sites

My meshes still having the same high cost than few weeks ago in other sims :/

 

This is before, when meshes were nice lol...

This mesh was 32 prims at 64x64x64 :D , 11 at the default size.

 

snapshot003[1].jpg

 

And this is now, when meshes arent so nice and dont change too much the PE whatever you do (optimize LODs or Phys shape)...

Now is 76 prims at the default size (about 160 at 64m o.O) LoLz!

 

Snapshot_025.png

 

Conclussion, the PE is now about x5. I should say too, that the second pic have a phys shape of just 1 triangle, while the first one have a nice basic but more complex shape and the LODs of the second pic are also lower than the first pic.

So maybe Im missing something, but I still getting the same PE no matter what region I rezz the mesh. Maybe something went wrong or is a bug in the viewer, I dont know, I just still seeing this really unfair :/.

Anyway, what worries me more actually are that dont matter what I try the PE never goes down. Why if I lower the polygon count on the LODs (not the higuest one of course) using the automatic LODs generator the PE never goes down? And why I get the same PE with a simple triangle as phys shape than using the whole highest LOD? It have no sense.. D:

Link to post
Share on other sites

The first pic is from an old viewer (cant remember exactly), the second pic is from 2.7.5.233867.

I updated the viewer to the latest version 2.8.1.235907 and that problem about lowering the PE using better LODs and phys shapes is fixed.

Anyway, the viewer still have issues with COLLADA v1.4.1, only files of v1.4 works perfectly. The UVW are messed with 1.4.1.

BTW, the displayed price is near to the real one? I heard that the minimum price will be 150 L$, but I uploaded somes meshes for just 11 L$ and even that dragon is just 23 L$. Just wondering if are random numbers there or they will be similar to the oficial ones.

Oh, didnt like too much paying in the beta grid :/, thats really bad for testing purposes. Even if the price is displayed, should be free, or at least someplace to get money -_-.

Link to post
Share on other sites

That price looks more reasonable. Isnt like before was too expensive either, but that can make a REALLY expensive products at time to sell them.

That about a Linden that gives you money sounds funny lol. I always wonder, if the beta grid is just for tests purposes, why there arent any "box" or "machine" where you click and get money. It could make all more easy for everyone.

Thanks for your comment ;)

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...