Do you get lower complexity using baked on Mesh vs Onion Avatar?

[Rob Huntsman]

Catwa has finally released their baked-on-mesh update for male avatars. And I was wondering before I finally try baked on mesh. Will this also lower my complexity for my avatar drastically? Compared to the normal onion avatars we use now?

Edited December 25, 2019 by Haselden

[Maitimo]

You'll save some but not a drastic amount. The normal onion-layer heads from Catwa add about 5,000 to your complexity. Switching to a BOM version will remove two out of the three layers, so the BOM version will save you about 3,300. @Skell Dagger will know the exact figures but I will be very surprised if it's not around there somewhere.

It will save some scripts too, since a BoM head with only one layer won't need all the make-up/beard/hairbase scripts.

Edited December 26, 2019 by Matty Luminos

[Rob Huntsman]

29 minutes ago, Matty Luminos said:

You'll save some but not a drastic amount. The normal onion-layer heads from Catwa add about 5,000 to your complexity. Switching to a BOM version will remove two out of the three layers, so the BOM version will save you about 3,300. @Skell Dagger will know the exact figures but I will be very surprised if it's not around there somewhere.

It will save some scripts too, since a BoM head with only one layer won't need all the make-up/beard/hairbase scripts.

So you lose about 7-8k, because signature has three layers as well. Good to know, gonna see if stray dog can just update their skins for it.

[Chic Aeon]

So far --- I didn't install the BOM viewer partly because I didn't care and partly because friends had a lot of issues and went back to a previous version ---- I am NOT seeing a lot of odd colored blue and yellow folks. So as far as I can tell BOM isn't going over the way LL hoped it might.    You can get your complexity down really pretty low just by choosing some well-made clothes and hair.   I have a friend with Lara and a altamura head and she stays under 30,000 with not much effort at all. I am normally around 40,000 or so.  

Edited December 26, 2019 by Chic Aeon

[Fritigern Gothly]

2 hours ago, Chic Aeon said:

I have a friend with Lara and a altamura head and she stays under 30,000 with not much effort at all. I am normally around 40,000 or so.  

I have an Akeruka head and a Belleza body. If I go full-on BoM and be careful what hair and shoes I wear, I will clock in between 12,000 and 20,000 depending on the hair and shoes.

[Chic Aeon]

40 minutes ago, Fritigern Gothly said:

I have an Akeruka head and a Belleza body. If I go full-on BoM and be careful what hair and shoes I wear, I will clock in between 12,000 and 20,000 depending on the hair and shoes.

If I am careful I can do that now without BOM :D.  The bigger problem is that the majority of folks are simply not careful.  But glad you are! 

 

[Conifer Dada]

I still use a classic 'system' avatar although I've tried mesh demos.  I've also tried bakes on mesh on a good freebie mesh body and was surprised how well it worked.  So I'll talk about my observations regarding other peoples' mesh avatars.  If I go to a busy place, they take a long time to rez properly, even if I reduce my jellydoll threshold.  I have a good computer and fast internet connection these days.  But the biggest problem with other peoples' mesh avatars is the time it takes applied textures like tattoos to rez - sometimes they never rez properly.  So I'd be interested to know if bakes on mesh rez any faster than 'appliers' textures. 

Edited December 26, 2019 by Conifer Dada

[Cinos Field]

It's not about the complexity, which does not represent actual load accurately. That's why they're reworking it! For now, you can right click and inspect the heads and look at vertices and triangles: the BoM one has a significantly smaller impact on everyone around you. Therefore it's very worth updating to if you care about that kind of thing!

And yes, since BoM only loads one composite texture it means that if someone can see your skin they can also see your tattoos, freckles, scars and whatnot.

[Lillith Hapmouche]

1 hour ago, Cinos Field said:

And yes, since BoM only loads one composite texture it means that if someone can see your skin they can also see your tattoos, freckles, scars and whatnot.

Plus, start to say goodbye to (most) alpha blending issues, appliers overlapping and other such nonsense. That's my favourite part of the system.

[Qie Niangao]

There are different ways to do BoM, and unfortunately some mesh avatars are still using alpha-cuts even though they're cutting up a BoM surface that was designed to be selectively hidden by alpha masks. The BoM aspect of such avatars means they only need a depth of one surface, but if they've chopped it up into all those little alpha-cut-eligible sides, then there's still a whole bunch of extra geometry and rendering complexity (not to mention the extra server-side complexity of the multiple links needed to hold all those independent cut sides, and the scripting to change their visibility). So if a mesh avatar still enables alpha cuts, it's... well, it's just lame, and it definitely won't reduce complexity as much as would a proper BoM avatar.

Mesh heads, though, lag for so many reasons I'm not sure geometric complexity is even in the top ten. On the other hand, done right, BoM could definitely reduce heads' user interface complexity by removing confusing options better handled by simple baked texture layering. (And "better" not only because it's simpler, but also because it fixes a whole range of messiness as @Lillith Hapmouche mentioned.)

[ChinRey]

19 minutes ago, Cinos Field said:

It's not about the complexity, which does not represent actual load accurately.

The render complexity in itself isn't too bad, the problem is that it doesn't work with fitted mesh.

When LL developed fitted mesh, they made a huge mistake. They forgot that fitted mesh is resized when worn and since isze is such an important factor for render complexity, it means the calculation is way off. CatWa discovered this mistake before LL and rather than report it, they decided to take advantage of it. Other fitmesh makers followed suit and by the time LL caught on to the problem, it was too late.

I don't know exactly what the actual render complexity of a CatWa head is but it has to be well over 100,000 is pribably several hundred thousand and may well be close to a million. With that in mind, cutting it to a third becomes very significant.

Btw, don't blame CatWa for it, they're here to make money not to fix bugs.

[Cinos Field]

5 minutes ago, ChinRey said:

The render complexity in itself isn't too bad, the problem is that it doesn't work with fitted mesh.

When LL developed fitted mesh, they made a huge mistake. They forgot that fitted mesh is resized when worn and since isze is such an important factor for render complexity, it means the calculation is way off. CatWa discovered this mistake before LL and rather than report it, they decided to take advantage of it. Other fitmesh makers followed suit and by the time LL caught on to the problem, it was too late.

I don't know exactly what the actual render complexity of a CatWa head is but it has to be well over 100,000 is pribably several hundred thousand and may well be close to a million. With that in mind, cutting it to a third becomes very significant.

Btw, don't blame CatWa for it, they're here to make money not to fix bugs.

That's the sad truth of it, many big creators are here to make money. So they'll use any trick they can regardless of how it affects those of us who actually play the game.

Edited December 26, 2019 by Cinos Field
Perhaps not most, but many

[ChinRey]

1 hour ago, Cinos Field said:

That's the sad truth of it, many big creators are here to make money. So they'll use any trick they can regardless of how it affects those of us who actually play the game.

That's how they become "big" creators. Still, it's not their faults. This is exactly what a free market means and they just play by the rules.

[Lucia Nightfire]

1 hour ago, Qie Niangao said:

There are different ways to do BoM, and unfortunately some mesh avatars are still using alpha-cuts even though they're cutting up a BoM surface that was designed to be selectively hidden by alpha masks. The BoM aspect of such avatars means they only need a depth of one surface, but if they've chopped it up into all those little alpha-cut-eligible sides, then there's still a whole bunch of extra geometry and rendering complexity (not to mention the extra server-side complexity of the multiple links needed to hold all those independent cut sides, and the scripting to change their visibility). So if a mesh avatar still enables alpha cuts, it's... well, it's just lame, and it definitely won't reduce complexity as much as would a proper BoM avatar.

Mesh heads, though, lag for so many reasons I'm not sure geometric complexity is even in the top ten. On the other hand, done right, BoM could definitely reduce heads' user interface complexity by removing confusing options better handled by simple baked texture layering. (And "better" not only because it's simpler, but also because it fixes a whole range of messiness as @Lillith Hapmouche mentioned.)

Sadly, IMHO, I believe alpha cuts are here to stay as I do not see clothing makers providing custom alpha skin layers to hide parts of the body, nor do I see people building outfits with every combination of alpha texturing necessary to hide their body under whatever clothing  is in the outfit, nor do I see that many textures working on any one particular body layer stack.

Time will tell, but I have my doubts.

[Cinos Field]

Slink switched entirely away from alpha cuts and provides a pretty big library of standard layers that from my experience work with almost all clothes. Very rarely do I have to make my own unless I want to.

That set of alpha layers is free in her store, by the way, even if you don't have the body, IIRC

[Lucia Nightfire]

Project Arctan is going to drastically change how complexity is calculated.

I look forward to onion avatars that once reported 20K getting bloated to over 200K.

I can almost hear the crying now...or is that the sound of people raising their maximum complexity sliders.

[Maryanne Solo]

Catwho? lol Go with a serious creator who got it right the first time. *ducks laughing.

[Anna Nova]

2 hours ago, ChinRey said:

...

Btw, don't blame CatWa for it, they're here to make money not to fix bugs.

They'd make a lot of money if LL fixed the problem and all the Catwa users had to buy a new head...

Maybe the best way to fix it for LL would be to make a new FREE Mesh Head that doesn't exploit the bug, then fix the bug.  Everyone happy, except Catwa, Lelutka, Genius....

Personally, I'd make it a condition of sale in SL to report such bugs, or out.

Edited December 26, 2019 by anna2358

[Cinos Field]

Are there any (human) mesh head makers who don't use that bug, by the way? It's a little off topic but I'm always looking to optimize my avatars as much as possibile.

[Desiree Moonwinder]

25 minutes ago, Cinos Field said:

Are there any (human) mesh head makers who don't use that bug, by the way? It's a little off topic but I'm always looking to optimize my avatars as much as possibile.

No bug was named. 

[Cinos Field]

12 minutes ago, Desiree Moonwinder said:

No bug was named. 

Bug, glitch, exploit, they're all fairly synonymous to me. But I also don't want to sidetrack this topic even more with semantics, as much as I love discussing them! Sorry!

Edited December 26, 2019 by Cinos Field

[Desiree Moonwinder]

9 minutes ago, Cinos Field said:

Bug, glitch, exploit, they're all fairly synonymous to me. But I also don't want to sidetrack this topic even more with semantics, as much as I love discussing them! Sorry!

No bug, glitch, exploit was named or described 

[Fauve Aeon]

6 hours ago, Cinos Field said:

Therefore it's very worth updating to if you care about that kind of thing!

This could be said on a sliding scale for everything in the world as any changes are made. But there’s a complex equation that involves human typists, awareness, skills, time, money, vanity, embodiment, computing power, internet speed, a learning curve, tolerance, the willingness to make changes and probably lots of other factors I’m skimming. Everyone sits at a different place within it for just as many different reasons. Which is why discussion and teaching will likely get more and better results than the scolding, shaming, nagging or haranguing about complexity that some people do. That last part probably belongs in the pet peeves thread, I admit! 😉 I like to believe that more SL people just care more for their avatar looks and pretty things than they do about taking the time to learn about optimization but also...what people think is ‘good enough’ or ‘caring enough’ about optimization is also on a sliding scale so getting a majority of people to care more about lowering complexity without alienating them is a tricky business that has not yet been accomplished. 

Edited December 26, 2019 by Fauve Aeon

[Qie Niangao]

24 minutes ago, Desiree Moonwinder said:

No bug, glitch, exploit was named or described 

Not sure whether there's a jira (or jiras) about it, but it's apparently a big motivator for the whole ARCTAN project, and what ChinRey describes thus:

4 hours ago, ChinRey said:

[...] When LL developed fitted mesh, they made a huge mistake. They forgot that fitted mesh is resized when worn and since isze is such an important factor for render complexity, it means the calculation is way off. CatWa discovered this mistake before LL and rather than report it, they decided to take advantage of it. Other fitmesh makers followed suit and by the time LL caught on to the problem, it was too late.

I take ChinRey's word on this because I have zero expertise about rendering complexity of fitted mesh. Is there a different assessment of the situation??

[Desiree Moonwinder]

Agreed  ARCTAN is about updating metrics across the board.