Can Second Life Catch Up With VRChat's Creation Tools? A Graphics Pro Explains the Problems -- And Offers Some Partial Solutions

Right: The same avatar made by Krample selling to VRChat users for $40 and (below) for SL users for L$550 (under $3)

I recently told you about a 3D creator called “Krample”, who has been an SL merchant for many years. Listed in the SL Marketplace as Orange Nova Avatars, he’s now finding far more success selling his avatar content to the VRChat market. That’s partly because the VRChat userbase will pay much more for his avatars -- $40 USD, versus a handful of USD in Linden Dollars.  But more than that, the switch was motivated by the creation tools VRChat has, versus what Second Life offers -- especially as Krample is a professional graphics artist in real life. 

“I've been making content in Second Life for the past decade or so,” as Krample told me last month, “but it's increasingly brutal to be working in an engine that's just so... old. Especially ever since I got a day job working in Unreal.” (If you've watched the Super Bowl in recent years, you've probably seen visual effects he helped create.) “[I] get to have a daily reminder about how much better things could be!”

He might not even still be creating SL content, he tells me, but for the pandemic: 

“If it hadn't been for COVID keeping me at home,” as he puts it, “it'd be flat out impossible to work on ANY new SL content. My spare time just wouldn't allow it, it'd take me years to finish anything. Several of my SL avatars have taken years to put together, as a matter of fact, just ask any of my disgruntled customers.”

I asked him to compare Second Life and VRChat’s creation tools -- and what Linden Lab might do, to catch up:

“The model [for SL and VRChat] is the same. But the engines, and what they let you do, are quite different -- so the avatar features in the two engines are very divergent, as is how they look. Whereas SL locks you down to the single shader used for everything, VRChat lets you build your own shaders if you desire - so even just visually, I can easily create much more aesthetically pleasing characters in VRChat, even though the model and textures are the same!"

Creating avatars in SL, versus VRChat

“Take facial animations, for instance. In modern game engines - and VRChat - you have what's known as ‘Morph targets’ or ‘Blendshapes’. It's basically an alternate state of a 3D model, that you can seamlessly morph to.

“So, if you have, say, a neutral face, then you can copy and sculpt a frown on that face, and make that a new blendshape. Then, you may seamlessly blend between the neutral face, to any level of frownyness you'd like.

“Combining multiple blendshapes lets you build expressions - and with a small library of blendshapes for the basic ‘Visemes’, or mouth shapes, for different sound archetypes made when we talk -- you can set up a convincing talking system in a day or two.

“However, SL does not have blendshapes. So instead, for my newest, unreleased SL avatar, I've had to do all my facial animations with bones. And it has been grueling: 305 unique animations, just for basic facial customization and expressiveness - weeks of work, for something that could be done in a day or two in VRChat.

Working in Unity vs. the Second Life Engine

“There's countless other things that make creating SL content grueling as well (No custom shaders, no custom armatures, primitive workflows all around), and it's just depressing to even think about starting a new project.

“Whereas working in Unity, which VRChat is based on, has been an absolute delight. Starting from scratch, it took me three days to learn Unity, figure out the VRC SDK, skim through their documentation, and put together my first working avatar in-world. It was an absolute breeze, and wonderful fun to work on!”

What Linden Lab can do to bring Second Life tools up to VRChat/Unity quality

“I don't actually know what they CAN do. They cannot drastically change or alter the engine without breaking old content. They cannot break the old client, because their userbase -- which, again, has individuals who probably have spent thousands, if not tens of thousands of USD over many years, on their accounts… Those users would revolt.

“I know that Linden Lab tries - my heart bleeds for all the resources that were thrown into Sansar instead of modernizing SL (Which could have experienced a renaissance in these COVID times, if it was just more state of the art and ready for corporate use).

“But in my eyes, they are shackled by legacy. They're also not going to be able to compete with the visual fidelity and quality level achievable by Unreal Engine 4 or even Unity -- gone are the days where you can just develop an impressive render engine in-house. 

"But taking a few steps towards streamlining the content creation pipeline would go a long way for me: 

If not custom shaders, then at least MORE base shaders to choose from.

if not custom rigs, then at least blendshapes!

Upping vertex weight limits from 4 to 5 would help a lot as well, what with having to skin for both deformation and avatar shape sliders (Fitmesh).

“Then maybe more creators would be able to create products like I do, without having to suffer the hurdles that need to be worked around.”

The paradox here may be that Linden Lab surely knows all this -- but also knows that changing any of it would incur drastic risks.

Comments

[Lex4art]

Good review from actual developer - SL is old, it's content creation pipeline is old, and that kind of legacy cannot evolve much further (guess SL takes insane effort just to keep SL on float with every new tiny modern feature added like mesh avatars).

[Nodoka Hanamura]

Honestly, as much as it pains me to say this, I feel Linden Lab has to step on some toes to keep SL competitive. If it means sunsetting support for things like flexiprims by converting them in bulk to mesh and no longer supporting uploading (as an example), or updating the basic avatar to not look like it came straight out of 2003, then I'm in favor of it. If it means keeping my virtual home alive, as long as it doesn't rip out its' very heart and soul, I'm willing to make some sacrifices. Question is, how willing is the rest of our population to do the same?

[Joey1058]

Second Life is at a crossroads for me lately. I've been very loyal, but I'm seeing worlds that hold so much more potential. All of them have their attractions and drawbacks. Will I stop using SL? Doubtful. But I'm seeing the writing on the wall. And that writing says to embrace the new and retire the old.

[Nodoka Hanamura]

My apologies, I meant Sculpties, not flexiprims, my bad.

[Pulsar]

It was old over a decade ago already, it's really an ancient engine by now. It was originally meant to render simple cartoonish avatars and basic but tweakable primitives with which you could create anything.
But was that bad? A game doesn't need an awesome graphic to become popular and can be fun even if made just by cubes (you know, Minecraft).

However, then Linden Lab took a different path and targeted a different kind of people. They hammered new features into SL (sculpt, mesh...), they failed at turning SL into a conference tool, but they pleased the niche who uses SL as a doll-dressing game or as a porn simulator, who is happy to see every detailed bit up to skin pores. So SL has became prettier, but the added complexity distanced the crowd who enjoys construction sandbox games.
Coincidentally, it was May 2009 when Minecraft has been released and later took off; SL had started its decline around that time.
It almost feel like LL missed a train by following those marketing ideas, instead of focusing on the construction game... and they missed it again in 2017 with Sansar, that among many other issues, it started serious a bit like a VR museum experience; while the most successful VR social game so far, VRChat, is a weird, funny, anime crowd - and although someone found a few more serious uses for it later, VRChat remains a crazy, colorful anime crowd.

And now SL is this mess.
User-friendliness and a good first user-experience are still lacking, but it may look prettier than before. Still it can't compete with modern engines anyway and it's falling behind more and more. At the same time it is no longer a fun prim construction game, because to make something considered good now you have to learn Blender or Maya. Still for creators the engine is too limited. And yet the virtual fashion and the virtual adult-stuff folks want photo-realistic avatars in SL: you could still make something decent, but inconsiderate folks who don't know or don't care of optimization make it severely worse without improving the visuals; but Linden Lab doesn't help about it either.
Now you can't go back: by taking that path, SL ended up with these SLers, who rightfully are happy with their niche stuff.

But could Linden Lab hammer further stuff into SL, to keep it somewhat update? Maybe something, but the problem isn't just about the broken legacy content.
Linden Lab even did that a few times. E.g. physics updates, but also invisprim are broken since deferred/advanced lighting has been implemented. Of curse they added alpha layers to compensate, but SL lacks of volumetric water too: invisprims were also a workaround for boats, that now look flooded unless you make them to float unrealistically higher above the waterline.

The point is that even willing to break the legacy content, you can do only so much with SL.
According to Oz, for example, the Lab felt that implementing mirrors wasn't a good idea in SL: https://youtu.be/FpU3rxxpeqQ?t=28m15s
You would need a full rewrite, that somehow can import the old asset and can maintain your friend list. Linden Lab bet otherwise and lost that bet. Now I don't expect they are going to do anything so radical.

Unless the new owners decide that SL is no longer profitable, I don't think SL will poof tomorrow, but I don't expect many improvements either. I'd be glad if they work on that, though. For sure SL has some competition now.

[Rairosu_Ishida]

This might be old, But so from what they are saying is that this is basically hell to create an avatar and that making an avatar at this level is like bashing your head on a wall?

I don't see why the creator is EVEN on SL if he thinks its suffering. I don't see other creators complaining about this and they still make a lot of money from sales and we actually liked what they make. Jomo being one of them.

Like jeez if your gonna go above and beyond to a Game Dev level of making stuff than maybe your not long fitting to be in SL you out grown it.

We don't want game quality Avatars that gets close to a video game character or main character. We just want cool looking avatars to be as and add some cool stuff on it.

I know they want to keep pushing the limits further and further. But come on relax what they done already is GOOD ENOUGH stop punishing yourself on trying to make something that is impossible. Old or not for SL When mesh came around it was a major shift, I use mostly mesh avatars and I am quite happy with it.