Jump to content
You are about to reply to a thread that has been inactive for 1568 days.

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

Recommended Posts

Posted

Some of you have just hopped in and others have been playing in sl for years!

What do you think can be improved and why isn't it happening?

Is this restricted financially or are they waiting for the right skilled individual or group to tackle the issues?

 

My Questions

- What if they used a new engine?
- What if optimization and tools were already accessible and made simple so that we are made aware for everyone even creators/scripters etc
- Why are some against the move of the severs in the cloud than what they already have?
- Why can't creators be warned that unoptimized goods will inccur a reminder too many of these can cause a poor reputation so others can see.
- Why can't stores inworld and in marketplace include vital information such as polygon count, complexity, etc
- Are LL focusing on important things or could they do more or something different??

 

 

  • Like 1
  • Haha 1
Posted
3 hours ago, NiranV Dean said:

Because LL is a company, needs money to survive and can't do anything that would even remotely anger users... unless they absolutely have to.

Think that is inevitable either way someone is going to be annoyed and angered.

- Not doing enough = feel the back lash
- Putting out nice visual stuff mesh, bom etc = feel less back lash just a nice short term pain killer doesn't focus on the actual issue.
- fixing the issues = better experience feel the back lash of LL losing profit

The new investors hope they will inject some well needed funding to sl, it is 2020 the idea of sl is still strong so make it great again please 🤑!
https://youtu.be/hSdQ4A0xsCM

Posted
53 minutes ago, Zeotetra said:

Think that is inevitable either way someone is going to be annoyed and angered.

- Not doing enough = feel the back lash
- Putting out nice visual stuff mesh, bom etc = feel less back lash just a nice short term pain killer doesn't focus on the actual issue.
- fixing the issues = better experience feel the back lash of LL losing profit

The new investors hope they will inject some well needed funding to sl, it is 2020 the idea of sl is still strong so make it great again please 🤑!
https://youtu.be/hSdQ4A0xsCM

Well that's the issue right there.

Fixing something means breaking something that people rely on = backlash.

Not doing anything gets people angry = backlash.

So whatever they do (or not) they will get backlash, what do they choose? Do whatever causes less backlash and best case less work.

  • Like 2
Posted
24 minutes ago, NiranV Dean said:

So whatever they do (or not) they will get backlash, what do they choose? Do whatever causes less backlash and best case less work.

Still wont fix the issue unless that is something down the line ??

Posted
8 hours ago, Zeotetra said:

Some of you have just hopped in and others have been playing in sl for years!

What do you think can be improved and why isn't it happening?

Is this restricted financially or are they waiting for the right skilled individual or group to tackle the issues?

 

My Questions

- What if they used a new engine?
- What if optimization and tools were already accessible and made simple so that we are made aware for everyone even creators/scripters etc
- Why are some against the move of the severs in the cloud than what they already have?
- Why can't creators be warned that unoptimized goods will inccur a reminder too many of these can cause a poor reputation so others can see.
- Why can't stores inworld and in marketplace include vital information such as polygon count, complexity, etc
- Are LL focusing on important things or could they do more or something different??

 

 

Like most things in life, I think SL has areas that could be improved,  however, with the large variety of why people are here and what their particular focus is, I don't think that SL could be made perfect for every community's perspective, so I try to not be too critical about what features or components LL chooses to work on (or not work on).  I'm still here, and still spending money here, so it hasn't been too bad of an experience.  Overall, I have seen a lot of changes in the last 6 years that have improved or enhanced my experiences in-world, and I do think that there will continue to be further changes and improvements into the future, and that some of them will likely benefit me.  

In terms of some of the questions about building, optimizations and creators, although there may not be specific direct information from LL to creators, there is a lot of information here in the forums about what makes items optimized or unoptimized and information creators can use to improve their products.  So I put some of the leaning about how to optimize things for SL onto the creators.    There are things about SL that are both good and bad, such as allowing anyone of us to decide to become a creator regardless of our level of understanding about, and interest in, creating well-optimized objects, or in providing helpful information about an object for consumers to use in deciding whether to buy the object or not.   There are measures that might have been able to be taken back when the world was new that might have put more limits on what could be created and/or uploaded into the world, but with the size of SL now, and the huge number of creators, general users and objects, it is probably too late to implement some of those measures without really disrupting SL as we currently know and use it.  Non-creators have some responsibility as well, to understand and use some of the tools that are available for choosing what they purchase, caring about their own avatar complexity and how it might impact others, not setting their LOD value to "8" because a creator's NC said to do so, etc.   There a lot of tips that can be picked up from the forums for the general user.  

If building tools are made too simple, I would worry that some of the functionality and parameters needed by the tools for complex creations would be limited, which in turn would limit what creators were able to build. 

I think that it is probably a balancing act for LL in deciding what they can do that doesn't completely disrupt existing things, yet provides a better experience for a large part of the user community, yet stays within the workload budget of what can be accomplished with their development staff, yet allows them to keep their overhead costs within a range that doesn't require them to raise end costs for the users but allows them to perhaps decrease them. 

Posted
24 minutes ago, Moira Timmerman said:

I don't think that SL could be made perfect for every community's perspective

Yes but it could be made much better!

 

26 minutes ago, Moira Timmerman said:

I do think that there will continue to be further changes and improvements into the future

Truly hope so, not just for aesthetics but optimization and performance too.

 

28 minutes ago, Moira Timmerman said:

about building, optimizations and creators, although there may not be specific direct information from LL to creators, there is a lot of information here in the forums about what makes items optimized or unoptimized and information creators can use to improve their products

There is yes i agree, and it shouldn't just fall on the community to do that, LL can provide specific information & even at least approve community information also.

 

34 minutes ago, Moira Timmerman said:

it is probably too late to implement some of those measures without really disrupting SL as we currently know and use it.

To cope with a little disruption or to endure an issue and let it drag on ahh
Even road works have little disruptions but at the end they create new paths for vehicles to move.

 

38 minutes ago, Moira Timmerman said:

Non-creators have some responsibility as well, to understand and use some of the tools that are available for choosing what they purchase, caring about their own avatar complexity and how it might impact others,

LL should at least have updated information regarding this made clear, to be more aware and be more informed would help much more.
Instead the focus seems to be more on looking great, creating, and so on. Notice that on the main sl website, why not add a section on sl tools and how to optimize.

 

41 minutes ago, Moira Timmerman said:

If building tools are made too simple, I would worry that some of the functionality and parameters needed by the tools for complex creations would be limited, which in turn would limit what creators were able to build. 

  Yes i agree, but there should be some guidelines and implications that going beyond a certain point would result in added weight etc etc

 

43 minutes ago, Moira Timmerman said:

I think that it is probably a balancing act for LL in deciding what they can do that doesn't completely disrupt existing things, yet provides a better experience for a large part of the user community, yet stays within the workload budget of what can be accomplished with their development staff, yet allows them to keep their overhead costs within a range that doesn't require them to raise end costs for the users but allows them to perhaps decrease them. 

Very good point, and do hope they will continue to make progress towards a  better sl

 

 

Posted (edited)
1 hour ago, Zeotetra said:

To cope with a little disruption or to endure an issue and let it drag on ahh
Even road works have little disruptions but at the end they create new paths for vehicles to move.

It depends on how much disruption and how complex the changes are, I suppose.  I think there probably is a limit to what can be realistically done behind the scenes and merged into the current without shutting everything down while major updates/conversions are performed.   I don't know if things such as changing the base engine to some other or changing the scripting language from LSL to some other language, for example, could be done easily and without a lot of trauma to existing stuff.  The fact that so much of the world is user-created and user-scripted is another one of those good and bad things, I think.  

In terms of general users and optimizations, there is still the issue of individual choice.  For example,  I remember when jelly dolls and the displaying of your avatars complexity were introduced. Immediately a lot of people wanted to know how to turn those displays and warnings off.  There are people that are going to wear what they want to wear regardless of the complexity.  There are people who want complexity and script limits enforced for large events - and then those who do not want their ability to wear what they want, wherever they want, infringed on.  There is even disagreement among those who are much, much more knowledgeable than me about whether the complexity calculation is measuring the right things, or if it's only a partial, ball-park number and if that is the case, should we put a lot of weight on paying attention to it. 

There are times when I think if I hadn't ever started reading in the forums, there would be things that I wouldn't be concerned about - I'd not be worrying about avatar complexity, I'd be running around the world seeing everything wonderful because my LOD setting was at 4.0, my parcel would be filled with wonderfully soft-looking alpha blended plants and grass, I'd be wearing multi-onion skinned head and body without even thinking about finding out what BOM was all about, and probably still be 2.0m tall...   But overall, I am glad that I do read here, and that there are so many knowledgeable creators and users here posting, so that through time and discussions and debates, a sense of general standards and guidelines starts to emerge (which is not always black & white, cut and dried, so while I do agree there are some things LL could be better at informing users of,  some of it comes out of the discussions and debates here by those "in the trenches", so to speak). 

ETA:  I 've probably gotten off of your main reason for your post, so I'll stop so others can comment more on your main questions. 

Edited by Moira Timmerman
  • Like 1
Posted
2 hours ago, Moira Timmerman said:

It depends on how much disruption and how complex the changes are, I suppose.  I think there probably is a limit to what can be realistically done behind the scenes and merged into the current without shutting everything down while major updates/conversions are performed. 

There's a lot that could be done entirely viewer side.

Consider a viewer targeted for the hardware a typical Steam user has. That's an NVidia 1060 GPU (about 4 years old), with 4 CPU cores, 16 GB RAM, and a 3GHz+ CPU.  Here are the hardware requirements for GTA V. Many SL users have hardware well below that level. That holds SL back. LL is reluctant to go there yet. This came up at the last Creator User Group meeting. I get the feeling that the younger graphics Lindens would really like to get SL into the current decade of graphics.

That's enough machine to support Vulkan and multi-threaded rendering. Vulkan is an API for getting the GPU to do things. In general, with Vulkan, the GPU does more of the work and the CPUs do less. Tiny custom programs running in the GPU ("shaders") do more of the work. Minimum GPU for Vulkan is an NVidia 600 series, which is a decade old.

Many people in the industry are very interested in finding out how Unreal Engine 5 works.  Unreal Engine 4 does large amounts of preprocessing and optimization on content while building level maps. UE5 apparently does more at run-time, so it's better suited to a dynamic environment like Second Life. Technical details are scant at this point. There is an impressive demo. They managed to get the GPU to do level of detail. SL needs that to deal with SL's elaborate clothing.

All that gets the frame rate up and the artifacts down. Same old SL, but looking better.

  • Like 2
  • Haha 1
Posted (edited)
5 minutes ago, Lyssa Greymoon said:

Those really aren’t very different from SL’s practical requirements, except for HD space.

True. If you try to run SL with the stated requirements, you'll have to turn everything down to "low". But there are people doing that. Usually on older laptops. There's at least one user still on Windows XP. LL collects data on what hardware people have; the viewers send it in.

Edited by animats
  • Like 2
  • Haha 1
Posted

/me hides Toshiba notebook with Core Duo, Centrino and Windows 2000 Media Center Edition under the bed.

  • Haha 2
Posted
13 hours ago, Moira Timmerman said:

It depends on how much disruption and how complex the changes are, I suppose.  I think there probably is a limit to what can be realistically done behind the scenes and merged into the current without shutting everything down while major updates/conversions are performed.

   

yes, the question then would be could the community tolerate a disruption, and how long could it take LL to complete?
Also taking into account blips and tweaks those are bound to happen somewhere down the line.

 

13 hours ago, Moira Timmerman said:

There are people who want complexity and script limits enforced for large events - and then those who do not want their ability to wear what they want, wherever they want, infringed on.  There is even disagreement among those who are much, much more knowledgeable than me about whether the complexity calculation is measuring the right things,

Its a tool and an interesting tool to have, may not be 100% however it can give an idea than not knowing at all.
Usually i would use more than 1 tool e.g.when i am in a rp sim the sim owner also makes an effort to show others how much they are using such as the

On a board:

- Complexity
- Script amount & time

I like these tools too

- Wireframe:  Ctrl shift R
Use this as another way to see how dense the object, items, from what i saw on a video the more dense the wireframe the more complex and its use of polygon

- Hilight transparent : Ctrl Alt T
Use this to see any extra items i may have hidden away sometimes they can be inside the body too

- Then i also like to see what texture size i am using in some objects this from the sl wiki helped me
http://wiki.secondlife.com/wiki/Texture_Usage


Just wish all these tools could be easily accessed in one area many do not press all these short cuts and to activate those by accessing them in advanced menu
ctrl alt shift D and the same again to toggle off.

14 hours ago, Moira Timmerman said:

There are times when I think if I hadn't ever started reading in the forums, there would be things that I wouldn't be concerned about - I'd not be worrying about avatar complexity, I'd be running around the world seeing everything wonderful because my LOD setting was at 4.0, my parcel would be filled with wonderfully soft-looking alpha blended plants and grass, I'd be wearing multi-onion skinned head and body without even thinking about finding out what BOM was all about, and probably still be 2.0m tall...   But overall, I am glad that I do read here, and that there are so many knowledgeable creators and users here posting, so that through time and discussions and debates, a sense of general standards and guidelines starts to emerge (which is not always black & white, cut and dried, so while I do agree there are some things LL could be better at informing users of,  some of it comes out of the discussions and debates here by those "in the trenches", so to speak). 

Agree, wouldn't of known all this either, and it really makes sense and how we can make a better experience for the community.
You can make things look pretty, and all that, but if everyone did their part... sadly that is not always the case. An example with covid19 not all wear mask or take procedures to protect it is not just an individual it can effect everyone not the greatest analogy but i think you know what i mean :D
 

 

 

 

13 hours ago, Moira Timmerman said:

There are people that are going to wear what they want to wear regardless of the complexity

If it was optimized in the first place wouldn't be an issue

 

10 hours ago, animats said:

Many people in the industry are very interested in finding out how Unreal Engine 5 works.  Unreal Engine 4 does large amounts of preprocessing and optimization on content while building level maps. UE5 apparently does more at run-time, so it's better suited to a dynamic environment like Second Life. Technical details are scant at this point. There is an impressive demo. They managed to get the GPU to do level of detail. SL needs that to deal with SL's elaborate clothing.

I wonder what the minimum/recommended hardware would be for UE5 hmm

This is for UE4.25
https://docs.unrealengine.com/en-US/GettingStarted/RecommendedSpecifications/index.html

 

UE5 does look very interesting wonder what other engines would be of interest and how does it compare to the latest havok engine?

  • Haha 1
Posted
21 hours ago, Zeotetra said:

  226302548_criticalthinking-1.jpg.9dbd185c98fddc917f34ab48c5451c20.jpg

Somewhere along the line some self-proclaimed genius changed the RenderVolumeLOD limit within their third-party viewer from the design limit of 2 to something higher and then claimed they improved rendering.  Now we have third-party viewer evangelism and content creators using this change as their baseline. Do not forget this hooliganism and the negative results when pining for real improvements.

  • Like 2
Posted
On 8/22/2020 at 7:14 AM, Zeotetra said:

UE5 does look very interesting wonder what other engines would be of interest and how does it compare to the latest havok engine?

Havok is just a physics engine. Viewer side, it doesn't do much. It's used by the mesh uploader to preprocess the physics models; that's it. Sim side, it's responsible for in-world physics.

Unreal Engine is a huge, comprehensive game engine. UE5 is interesting because it has a dynamic level of detail system, something Second Life badly needs. A reveal of UE5 was expected at this year's Game Developer Conference, but the conference went virtual, shrank, and Epic didn't participate.

  • Thanks 2
  • Haha 2
Posted

Before anyone starts saying SL should be in a (insert generic other engine here), i've ported one of my Avatars to Unity just to see how it would look like if an engine was used that is capable of far more than just diffuse, spec and normal maps.

The results are... disappointing.

Snapshot_156.png

I'd go as far as saying in Unity it looks even slightly worse though with a lot of more tweaking this could be brought up to the same as it looks in SL. Note that while Unity does a lot of the things naturally, like extra texture AO, detail shadows for the texture and some other extra features, all of which i enabled and did properly set up, it still doesn't come close to what SL does with a very crippled materials system that is limited to diffuse, spec and norm maps. Though thats where SL basically caps out, it looks decent from the get go without much setup but also doesn't go much higher, whereas Unity offers the ability to look all the way from bad to possibly better with utilizing some extra features though this requires a much more elaborate setup, and probably more than a diffuse, spec, norm, shadow and ao map, all of which came pre-made professionally with the avatar already and were simply baked into the diffuse map for SL.

I suspect without heavy modification Unreal Engine wouldn't make any of SL's content look better either, infact it would most likely look worse with how SL implemented all its stuff and before reworking everything to look good for Unity/Unreal we could as well add a few new fancy things to SL and make it look better and with some optimization it runs decently well too.

  • Like 1
Posted
5 hours ago, NiranV Dean said:

Before anyone starts saying SL should be in a (insert generic other engine here), i've ported one of my Avatars to Unity just to see how it would look like if an engine was used that is capable of far more than just diffuse, spec and normal maps.

The results are... disappointing.

Please try a test case where people know what it's supposed to look like in real life. No basis for comparison on the example.

 

principledbdsf.png.e359799b76571ab401e6064820e0b29d.png

The layers of Principled BDSF. Those 10 layers are enough for Hollywood. Few materials need more than 4 or 5 layers. Blender can edit and render those right now. Newer GPUs can render them in real time. This is becoming the new standard for materials.

It might not be that hard to convert SL content to Principled BDSF, because not that many objects really need manual conversion. Humanoid skins, mostly. They need some extra layers, such as subsurface reflection, that make skin look real. Few other materials need that. There are also layers used mostly for automotive paint jobs. Vehicle builders could get a supply of auto paint colors with "clearcoat" and "clearcoat gloss" values, and we would have GTA V grade shinyness. (If you want to see what a retrofit looks like in practice, look up "GTA V photorealistic mod")

"Physically based shading means we approximate what light actually does as opposed to approximating what we intuitively think it should do." - UE4 manual.

  • Like 1
  • Haha 1
Posted

Open GL is almost dead.

It may be that EEP is necessary for a switch to Vulkan. Apple and Microsoft use different alternatives to Open GL. Vulkan is being supported by current hardware drivers.

Project Sansar was a huge distraction.

The really hard problem is poorly made content. I have a low-complexity mesh body that looks good, and I try to meet that standard with the clothes I make and wear. My total complexity is usually under 20,000. Most people I see are over three times that complexity.That's not the only measure, it may be missing parts of the problem, but better optimisation could easily double the apparent speed.

I'm not hopeful. The quality of the documentation is pretty poor.

 

  • Like 1
Posted
23 hours ago, animats said:

Please try a test case where people know what it's supposed to look like in real life. No basis for comparison on the example.

 

principledbdsf.png.e359799b76571ab401e6064820e0b29d.png

The layers of Principled BDSF. Those 10 layers are enough for Hollywood. Few materials need more than 4 or 5 layers. Blender can edit and render those right now. Newer GPUs can render them in real time. This is becoming the new standard for materials.

It might not be that hard to convert SL content to Principled BDSF, because not that many objects really need manual conversion. Humanoid skins, mostly. They need some extra layers, such as subsurface reflection, that make skin look real. Few other materials need that. There are also layers used mostly for automotive paint jobs. Vehicle builders could get a supply of auto paint colors with "clearcoat" and "clearcoat gloss" values, and we would have GTA V grade shinyness. (If you want to see what a retrofit looks like in practice, look up "GTA V photorealistic mod")

"Physically based shading means we approximate what light actually does as opposed to approximating what we intuitively think it should do." - UE4 manual.

Wasn't the point of this test. People always say we should port SL to Engine XYZ because it would run better (no it wouldn't) and look better, which i've shown here isn't really the case, not without heavy modifications which could be done in SL and much easier than porting everything to another engine and starting over from scratch. Your comment about GTAV is just another proof for that, graphics mods (especially GTAV and Skyrim for instance) have always been there (Reshade is very popular for a quick-improvement in most games). We could easily make SL look so much better without porting to another engine and we could also make it run a lot better, although that it shall be noted that this only goes so far because the biggest culprit is still the content and no matter of optimization clientside will help with simply bad content that takes a dump on your optimizations.

As far as i know (that's what i was told) LL could have chosen to have full PBR and the original Materials system was just that but for the sake of ... well the users and the fear of "breaking" content they chose to do a dumbed down version that keeps compatibility which is why its so limited right now (but again could be easily extended to full proper PBR but would require rebuilding a lot of content to look right again)

  • Like 1
You are about to reply to a thread that has been inactive for 1568 days.

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

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...