These are not real people...

[Suspiria Finucane]

---

Jenni Darkwatch wrote:

One way to go would be to introduce a 2nd more modern avi mesh to SL while leaving the old one in place. That way people wouldn't be all pissed... oh... nevermind. Change anything at all and people will be pissed. Nevermind.

---

Excellent idea! So what if people get p.o. ed :matte-motes-agape:

[Suspiria Finucane]

---

Maryanne Solo wrote:

Second second life.

SL2

Your world your imaginationion.

You can only login if your comp is quad quad architecture

3 graphics cards, (no outdated appleish stuffs)

optic fibre internet..

No avatars under 1gb in resouce size... :smileysurprised:

Everytime you log in a 3rd world country goes pouffe' :smileysad:

Oh well....

Bring it on! \o/

---

[Penny Patton]

---

Jenni Darkwatch wrote:

---

Niamh Suisei wrote:

EG if every user has the ability to make things out of lego-block prims, SL will always look like 2004 poo. If you want SL to look like Battlefield 4 - and no reason it can't, more or less - people need to get serious about content creation. Whether textures, learning Blender/Maya, using the forthcoming normal maps etc. Battlefield 4 doesn't look like that because they asked a team of random people to make it, but because they had professionals using professional tools and tecniques making it. If you want SL to look like that, you need to get professional

And I don't mean charge L$10,000 for a model of an armchair, I just mean professional in your approaches to creation. Stop with the "ah thats good enough" if obviously its not. (But then, it's our fault for not demanding excellence at the checkout)

---

That's a bit simplistic.

SL has plenty of users who have truly old, ancient hardware and that won't change.

Content in SL cannot be as optimized as game content. In games, designers use every trick in the book to make sure content looks good:

• Don't use too many different textures
• Use all kinds of techniques to perfectly control how many polys are on screen at any given time
• Use low-res textures/low poly models for stuff players are unlikely to notice anyway

In SL, you buy maybe a skin from creator A - that's one big texture. Then a jacket, maybe even mesh. But you'll expect yours to be detailed, so it's a high poly mesh and large texture. Same for pants, shoes, hair, jewelry etc.pp. - for one avi that works fine. Now go to a venue. Everyone wants to look different. Different skin, different outfit, different accessories. Very soon even the best gaming rig will just drop to really low frame rates.

The very same thing that makes SL so attactive (user generated content) is also its bane. Take user generated content away and you'll get nice FPS and all. But you lose the very thing that makes SL what it is. Leave user generated content in and there's no way at all you'll ever see great performance except in sims designed specifically for performance.

---

 It's important to keep everything in perspective.

 Sure, nearly all of SL's content is created by hobbyists having fun, rather than professionals. Ergo, the work they do won't be the quality or efficiency of professionals.

 Second, There's a LOT more variety in avatars in SL compared to characters in videogames, which reuse lots of textures and models to reduce the strain on your hardware.

 

 These things are true. HOWEVER, that's not the whole of the matter.

 Fact is, there is PLENTY Linden Lab could have done to encourage more efficiency from the userbase, without constricting freedom of creativity.

 Count texture use towards Land Impact costs, charge a higher upload fee for larger textures. Provide basic tutorials people can access via the viewer itself which walk content creators through basic concepts while teaching them how to build best for SL.

 

 Better tools could even make it easier for anyone, regardless of skill level, to create better looking and more efficient content. I can think of a dozen ways LL could improve the existing appearance editor making it super simple for anyone to create a great looking avatar shape. The addition of "materials" will allow people to create more detailed looking content without piling on the polygons like we do now.

 Better default settings can help people run SL more effectively.  For example, if avatars weren't 8' tall with a camera floating 2 metres above their heads, people wouldn't need to build to the huge scale we see in SL which just encourages people to push up their draw distances, forcing their videocards to render more.

 

 Will everything in SL ever look as good as a professionally made game? No, almost certainly not. But Linden Lab should do what they can to make it easier for everyone to get the most out of what SL is capable of and so far they haven't.

[Jenni Darkwatch]

LL has a simple problem: If they'd raise LI even higher, everyone would scream bloody murder. They can't raise it for old content without causing mayhem on sims full of old content.

Regarding materials: I think those will raise GPU requirements, not lower them. It adds additional "texture" downloads to surfaces that previously didn't have them: Box prims and so on.

[Penny Patton]

---

Jenni Darkwatch wrote:

LL has a simple problem: If they'd raise LI even higher, everyone would scream bloody murder. They can't raise it for old content without causing mayhem on sims full of old content.

---

I agree. They needed to include textures in LI costs back when they first introduced mesh. That would have saved them the public outcry that will come if they ever do it now.

To be clear, reasonable use of textures should not increase LI of existing mesh model work too much, if at all. There are ways to weight these costs which keep efficient mesh content low LI.

That bar counter example I gave? With a single 1024x1024 texture, it should cost about as much LI as it does now. But loaded up with 8 of them like it is? It's LI cost should be way more than would be worth it to anyone to use. The size of the textures, the number of unique textures, and the size of the object should all factor into these cost calculations.

 

 By not including texture use in LI from the beginning, LL really shot themselves in the foot. They still absolutely need to do this, but now it's going to be more difficult.

---

Regarding materials: I think those will raise GPU requirements, not lower them. It adds additional "texture" downloads to surfaces that previously didn't have them: Box prims and so on.

---

You're probably right, but again it's because texture use carries no LI consequences. And I'm pretty confident LL will make the same mistake in not including material map use in LI costs, compounding the issue.

 

 

The problem is not materials themselves. They are a tool which allows us, when using them correctly, to do more with less. The problem is LL fumbling the ball, each and every time, when it comes to developing graphics releated tools and features.

 That's one reason why they need an art director so badly, to avoid these costly missteps when rolling out new content creation features.

[Jenni Darkwatch]

True. I'm just guessing, but I'm almost betting they didn't want to hobble mesh too much compared to the old content. There's still cases where using a (costly to render) sculpt beats mesh hands-down. 64x64 artificial landscaping for example... often done with mesh physics but sculpted shape. Bad for viewer performance, but rewarded by sculpts being always 1LI (unless linked to mesh or opted into mesh accounting)

[Penny Patton]

---

Jenni Darkwatch wrote:

True. I'm just guessing, but I'm almost betting they didn't want to hobble mesh too much compared to the old content. There's still cases where using a (costly to render) sculpt beats mesh hands-down. 64x64 artificial landscaping for example... often done with mesh physics but sculpted shape. Bad for viewer performance, but rewarded by sculpts being always 1LI (unless linked to mesh or opted into mesh accounting)

---

I agree and even I threw a tantrum when I saw some of the LI costs of mesh compared to sculpts, but eventually I realized the benefits outwayed the costs even in these cases, and most of the time you really are able to get better quality from mesh over sculpts and prims for the LI cost.

 However, I'm sure not everyone would think that way which is why, again, I'd say "reasonable" use of textures should amount to LI costs comparable to what you see now, with only excessive texture use impacting LI costs.

In addition, every new feature LL introduces which requires you to use LI over old prim costs sweetens the pot. Keyframing, pathfinding both require you to use mesh LI calculations (even if using prims), and there's plenty more features LL could also tie into that (such as materials).

 In light of that, there's no reason to think that reasonable LI costs attributed to texture use would halt mesh adoption.

[Jenni Darkwatch]

True, with those new additions they might have enough reason to use mesh even if it's higher LI.