Flea Yatsenko

Refresh rate and frame rate are basically how much something happens in a second. It's confusing because monitors use refresh rates and hz and games use frame rates and FPS. But they are all pretty similar in the end. An LCD screen can only display a certain frame rate or refresh rate. CRTs had variable refresh rates. LCDs generally do not. So if you have a 60hz panel, it's going to update 60 times a second whether your frame rate is 2, 60, or 1000. Now imagine you have a panel that only updates 60 times a second, but you are updating the contents of the screen 48 times per second because you are getting 48fps. 48 does not go into 60 very well, so there's a bunch of janky stuff that has to go on to make the numbers work between your monitor only showing 60hz (fps if you really want to be consistent) and your graphics card only outputting stuff at 48fps. It also goes backwards, If you have a 60hz screen and your game is running at 65fps, it doesn't line up either. So it has to do some jank stuff to make it work. I don't know the details. But it's the same deal Now, if you have a factor that goes into 60hz, like 5, 10, 15, 20, 30 for your fps it feels smoother because the screen update time is not fighting the screen refresh time. Which is why 30fps probably feels smoother to you than 31fps. vsync is one option to fix it. It adds like a frame of lag (1/60th of a second) and it stops your screen from tearing. Like when it splits and you have a line across your screen for a little bit. You can also limit frame rates. Monitors with VRR/Freesync/G-Sync are LCDs where the refresh rate can change with the content on your screen. So if you have a VRR/Freesync/G-sync (just call it VRR, variable refresh rate), the graphics card can tell the monitor it's only getting 48fps and the monitor will only update when it needs to, instead of trying to cram 48 frames into 60 slots of updates per second. I have a gaming monitor that refreshes 175 times a second (175hz) and it has freesync down to like 40fps. So anything between 40fps and 175fps doesn't have the jank and feels way smoother.

You can have stuff that sells even if the listing isn't that great. I have lots of those and I'm afraid of changing them, lol. I think the story is very different if you have an established store with a good amount of listings and if you don't. I think we are both in similar situations even though we sell very different stuff. One of the big rules of SEO is don't change everything at once and make sure you have some A/B testing going on. You have a good store. There are two types of people unhappy with SLMP. Those who can't sell anything and can't gain any traction and those who have been successful in the past and have noticed trends that seem very out of their control. But I think there are a lot of people who are very upset and have made some quality products and search is showing something that looks like it crawled out of the very early mesh days.

There are probably around 20 to 25 million listings on marketplace. Because of that they have to do a lot of caching and slow updating. So it can take weeks for your changes to show up in search. I sell a lot of similar products and when I search the pattern of closest match with the least irrelevant words show up first, then it keeps going to longer and longer names. From my experience, nothing bad can come from leaving out the cool names for your products and your store tag. Looking at your store, you have to drop the "by Punk JUSTUS" in the title, it's killing you. Any irrelevant words should be removed. Paris for REBORN MOUNDS & KUPRA MOUNDS by Punk JUSTUS should be more like "pink black bikini (reborn, mounds, kupra)" That said I don't think search is the only problem. I think there are issues with the whole web ecosystem of SL. I check dashboard every day multiple times, sometimes I'll get forcibly logged out or get a redirect error. I know correlation does not imply causation but when that happens in a few days my sales are great and then die down. It's happened too much for me to not be suspicious. Another odd coincidence is how my SLMP sales are basically always in the same range no matter the time of year, and my sales are only good when I'm moving stuff in world. I used to only sell stuff in world and would see normal seasonal fluctuations. Now it's like SLMP is only good for so much L$ a month and if you want more you have to sell outside of it. I've been struggling with sales compared to year over year for a while now no matter what I do. The only thing that's given me normal (or even good) months has been in world sales. Feels like there's some sort of ceiling in place on marketplace that wasn't there before. I am going to rip out of a lot of the unique names in my products and put them in the description and keywords, and remove my store tag from a listing. There is no need to have a store name, people can find your store via the "merchants/stores" search. Also don't get worried about January/December sales. I always cash out around the 15th of every month and since 2017 January has been one of the worst months for me. Oddly enough before that for around 6 or 7 years January was one of the better ones. It's really easy for seasonal fluctuations to make you think the sky is falling.

Because they want to charge for it. Not being open source, pushing a single feature. It all screams to me "we are going to charge you a monthly fee to use our viewer." It's not open source because they don't want other viewers giving away their "features" for free, they want it monopolized. They are probably hoping people support it and don't question if it's phishing or whatever because they are thinking with their loins.

After working with PBR and a big builder's kit. I have some complaints The legacy method is terrible. It's a total waste to make new content and include the legacy texture because it'll be useless in a year or two yet you have to do it now. I wish you could just use the color map from the material. Yes I know it's wrong, there's problems with doing that. But who cares? If someone is at minimum specs in the PBR viewer there's lots of stuff that looks terrible and not like it's supposed to. Anyone running low settings isn't expecting eye candy. IF you still want to make baked builds (with AO and SOFT shadows when it's allowed like a scene with no external lights and environment coming in) the UV maps are usually totally different between PBR optimized and baked. So you either have something with good PBR maps and instancing that doesn't bake or bake dependent so the PBR maps are way too sparse on the texture and low resolution because you have 5 different objects that could be instanced or Two separate versions which destroys the purpose of making backwards compat since the PBR version is just gonna show up no textures Once you put a PBR material on something you are doomed. Did you apply your legacy texture and PBR then realize you need to go back and change something? Might as well start from scratch with a new object. Actually kinda sucks right now to make content for SL. Maybe small stuff doesn't matter but if you're making builder's kits and stuff it's a total nightmare. Building with PBR stuff in world is a lot more fun. I don't think people are realizing what you can do with a PBR builder's kit, lights, and reflection probes. I think the slow rezzing they introduced is over shadowing a lot of stuff about PBR.

A4-5000 is not Bulldozer like the FX, it's a Jaguar CPU and it's actually quite good, completely demolished Intel in the price bracket at the time, though it was competing with celerons. I did have both FX and this APU. Perspective for what I tested on, it was a $500 netbook released in 2011. The entire package (CPU and GPU) is only 15w. The most important thing I got from my little test was that the default settings are way too high and make me think it runs horribly, but when I turned them down I was impressed. I was fully expecting a completely un-usable experience based on what people have been saying about PBR. Not only does my experience represent 2.5% of users based on computer age, at the time this laptop was a slow netbook that got 8+ hours of battery life when the average was only like 2 hours. I think LL needs to update their min requirements and turn down default recommended settings on older hardware. Maybe this should be in another thread because there's no reason why TPVs should be doing this. But lowering draw distance should be the biggest factor for reducing VRAM usage, right?

I just did a quick and dirty test on my A4-5000 netbook with latest FS Alpha with PBR. The default settings were way, way too high for this thing, like 0 fps high. I had to turn everything off and set view distance to 32 and I was actually getting 15-25fps on my build platform. I went to a club and was still getting FPS in the 10s but I didn't have time to wait for the new texture fetch algo to load a whole club. The CPU is massively struggling with FS on that APU, more so than the GPU I think. It took like a minute for FS to actually load and show the log in screen. Maybe I will look a bit closer later on, but the HD 8330 on this APU is about equal to Intel Integrated on Haswell according to notebookcheck. https://www.notebookcheck.net/AMD-Radeon-HD-8330.92874.0.html Haswell came out in 2012. Surely the people who are complaining so much aren't on 11 to 12 year old laptops? I was running Linux with KDE. AMD's support for that GPU isn't very good in Windows anymore. Perhaps the default graphics settings are too high for the PBR viewer and it's making people think PBR runs far worse than it actually is? I figure LL is responsible for the recommended specs though.

I can guarantee you there will be people actively seeking out fully baked homes and stuff over PBR two years from now. That's just how some of the SL residents are. Surely if you make something like a fully baked, non-PBR skybox and you run PBR on the absolutely lowest settings it's not that resource intensive? I've got a truly anemic netbook, 128 GCN 2.0 cores, an A4-5000 AMD APU. It barely ran with forward rendering on low. R7 260 was the same generation and had 768 cores, with 290x at 2816 cores. Surely if that laptop can get over 10fps on low the problems with PBR are over-hyped.

If someone doesn't have the computer to run PBR, they aren't going to have the computer power to run ALM with materials and shadows enabled. They are going to have all that stuff disabled. There is no reason to use spec and normal for a PBR fallback. You just need diffuse. I think the people who are most unhappy with PBR are the ones that are using the render system before ALM because ALM doesn't work well for them in the first place. The real problem is if you want to make a baked fallback. With PBR you can instance parts of builds like crazy, with baking you can't really do that, maybe instance the mesh but not the texture if you are very lucky. And honestly the only kind of PBR fallback textures you should be doing are baked or just diffuse since baking fallback when building for PBR might not be possible, i.e. you have one piece with one mesh and UV map you use in several places. With baking you cram them into one material so you don't have so many textures. With PBR you want one mesh with a full UV you can re-use over and over again. Kind of off topic for this thread but yeah. The biggest problem I see with PBR fallback and a viewer right now is not being able to go between viewing both when making a legacy version and a PBR version. If you forget to add the diffuse to something before adding the PBR, non-PBR viewers are going to see white and you'll have no way of knowing you missed something in a PBR viewer. So you'll have to relog into a non-PBR viewer to check your legacy build, and if you forgot one you have to go through all the hoops. I think legacy blinn-phong should just remove spec/norm from the editor since it's a total waste of time (unless editing something that already has it), L$, and bandwidth and treat the legacy diffuse texture as a separate layer that can be edited instead of a fallback buried underneath PBR. There really needs to be a fallback toggle for viewing legacy textures in a viewer, specially since we're going to have to basically support non-PBR for at least a year and a half, probably longer because there will be viewers that won't force people into PBR after that.

Well, I will say this. With ALM and Blinn-Phong there was a very big difference between baked and in world material builds. I'm working on a build right now, fully baked legacy version and a PBR version with the new features. And honestly the PBR version looks pretty great when the lighting is all done, something I couldn't get with ALM. PBR lets you make much better interior spaces (skyboxes, for example) with much better lighting. I think building in world is going to shift. Materials aren't going to be that important anymore and lighting and reflection probes are going to be the new way to get builds to look the best. When you build, once you figure out the new lighting system you are going to have a lot more options for changing the appearance of things with just lighting. Hopefully LL sees this through because the foundation they are laying can make something really amazing.

Yeah, PBR isn't a huge upgrade. Specially considering the best parts of PBR are the new lighting and reflection probes, things we probably could have gotten without changing to GLTF. GLTF is mostly about getting content creators to make things the right way. When more people use PBR and they start getting things that look like Vaseline when they shouldn't be shiny at all people are going to realize how many times content creators messed up Blinn-Phong. Which is gonna make a lot of people mad because a new feature is breaking existing content. But LL can't do much to work around content creators doing things wrong. Sculpties had a ton of drawbacks and the change was huge. PBR, GLTF, and the new lighting isn't nearly as groundbreaking.

If it's an alpha texture, the alpha settings are saved in the material instead of object so you don't have to change alpha settings every time you use the texture. Material moves a lot of stuff you'd apply to the object to the material. I.E. if you needed a glowing texture you would make your material have glow (emission) with a white texture then change the tint color and you don't have to change glow on every object you use that texture/material on. It can make some things easier, if you are using just a texture with no changes to alpha and no glow it is a waste of time. But my point is even if for some reason LL got rid of just applying textures and forced everyone to use materials, you could still just create a material with a diffuse and us the material like a texture. Sounded like someone was under the impression that if you weren't making gltf materials with Blender or Substance Painter or whatever that you weren't going to be able to use textures anymore.

Once it's a material, yes. 1. Go to inventory 2. Create new material 3. Drag your texture to "base color" 4. Click save 5. Use the material like you always used textures A little more setup than just uploading a texture but once it's done it's the same as always.

You can still use regular textures with PBR, just create a material and ignore all the maps other than the albedo. PBR just strongly encourages you to use all the PBR maps while Blinn-Phong/specnorm made it more optional by making more steps. PBR is not stopping anyone from using just diffuse/albedo (just a regular texture) or whatever you want to call it. I don't have time to explain but it's something like create a material from inventory, just add your texture to albedo, then use the material like a regular texture. As far as building goes, the old way was to treat diffuse, normal, and spec as entirely different textures and PBR creates one item out of all the textures. It is actually a bit easier to build with, because if you're re-using the same material you just have to manage the PBR material instead of applying the diffuse, spec, and norm one by one for each texture face. But there is nothing saying you have to use every single map in a PBR material.

That post should be stickied. Very good. But I have a 3950x with a 7800XT and I have my SL cache on tmpfs in Gentoo with a speed test of 560/23 and rezzing is still super slow. I think their new algo just isn't very good and releasing it when PBR is released means the logical conclusion for a lot of people is that PBR is slow rezzing. Maybe it has to do with my frames feeling smoother in PBR though.

I don't think the slow rezzing has to do with PBR. I think it's an unrelated problem that SL broke something that happened to be included in the PBR release. This data is super important, thank you so much. Will be interesting to see how many people stick with PBR and how many go back to a non-PBR viewer. Seems like people are excited about PBR but some of them try it and are disappointed. Especially because the lighting can be very different from what they are used to with the old browser.

I don't know how they did such a good job making uploading PBR materials intuitive and easy and the rest of the system just kind of falls apart as far as figuring out what to do. Wiki page being confusing for a builder who doesn't know anything about PBR is a sure fire way to make sure lots of people are using PBR incorrectly and we end up with another spec/norm (Blinn Phong) debacle where people are using the wrong types of maps and all sorts of stuff. You can shriek about RTFM all day long but if people RTFM and it doesn't help them understand what they need to do then it's not really the user's fault. And honestly someone who just joined SL and wants to build and has zero experience with 3d content creation is going to be completely confused by the wiki pages for spec/norm and PBR. And now we are going between calling spec/norm blinn phong, material, spec/norm and now going between PBR material and Material and whatever. Which is gonna make trying to learn about this stuff by searching a nightmare for new users.

From what I've been told, it's far more difficult to convert spec/norm into PBR than it is to go from PBR to spec/norm. PBR has more data and more map info. So to go from spec/norm to PBR you basically have to make stuff up. Much like trying to generate spec/norm from a diffuse (such tools exist), at best you can get it kind of close but it'll never actually look right.

I remember watching a video from LL and they said they were prioritizing the difficult parts first, like rendering AVs. I'd be a lot more concerned if this was in private alpha and inventory worked but all AVs were just placeholder blocks. Why attach stuff from inventory if you can't see your AV? Outfits are a good start at least. If you can change outfits the mobile viewer can at least attach and detach stuff from your AV. We are easily at least a year out from mobile viewer being in the hands of the average person on App Store and Play.

Given the usage of spec norm you'd see around the grid after all these years, I think PBR is more of a "spec/norm failed so we're going to try something more modern." I'd argue spec/norm failed, it's not used very often and a lot of times people don't read the wiki page and they use it wrongly. I don't think PBR is gonna be popular for a long time anyways. Getting PBR into the hands of users depends almost entirely on SLMP and search. So good luck with that.

Yeah, I think people are gonna learn how much better PBR is. If you did something like used ALM and spec/norm in a builder's kit with tunnels, you'd have to add some sort of light source to get materials to do anything, but a long hallway with more than 8 pieces you could see at a time and the lights would start flickering. So you'd be stuck trying to add as few light sources as possible to something because the more you add the worse the flicker would get.

Looks really good. People are gonna have to get used to not working around the 8 light limit of ALM. I think people are used to just throwing a single light in a dark area and having that be good enough, because ALM would just turn off lights if there were more than 8, creating a weird effect like people randomly turning off lights.

That's a really good theory, I remember all the "gig/side hustle" stuff going on when COVID hit. It can be tested too if you have had a store before COVID. Just look at product IDs and dates in your listings and see how much the ID went up each day before, during, and after COVID. That would explain why SLMP search has gotten bad. A bunch of new users who understand nothing about search optimization and best practices flooding the marketplace. That said I don't think that's the only issue, because search was showing very old and outdated results before new ones. Unless that was LL's way of trying to bury the wave of new content from new creators?

I wanted to check in and see how everyone was doing. I revamped my in world store and have been working on some tools and stuff, my sales have gone closer to normal with them all, but my marketplace sales are still trending horribly like they have been the last year, and my in world sales have been clobbering marketplace the last 30 days (like a third marketplace, two thirds in world sales as far as income). I'll also have days when I sell like two L$99 products on SLMP then sell a couple high end products in the same day. I was wondering if people were still having problems with SLMP, or if what I did made some meaningful change for me. Or if sales just suddenly got better and I'm just cannibalizing my marketplace sales with in world sales. I am planning on releasing what I'm doing but I at least want to know if it's making a big difference or if it's just trends going on. Decembers for me are usually one of the worst months of the year.

R Removing forward renderer is a big mistake. I didn't think mobile viewer needed PBR but judging by the requirements for the mobile viewer it probably does. My laptop is old and terrible but no way it's gonna run SL anymore, it barely ran it at lowest settings before. I do like the change to PBR. it's a lot easier to work with and for content creators who don't read wiki pages it's a lot more difficult to screw up. Making a shiny dress, it's easy with spec norm to start going crazy with spec values. PBR basically says "how much metal is in this material?" and if you turn it to full it's enough to at least make alarm bells go off in your head that your silk dress contains no metal. Unlike specularity which is a term most people don't even know what it means other than higher = shinier. But it's a lot clearer if something is metal or not. I see why people are upset, SL moving towards a direction that is excluding people, like the minimum requirements for a game going up after release. I'm lucky enough to have a nice desktop but if I only had my phone and a laptop I would be in serious trouble getting into SL. I imagine a lot of people don't have nice desktops and are worried what's going to happen to their second lives. If i had to use my laptop or phone for SL I would rather be able to have something usable and have missing textures than have nothing at all. Specially since I'm wondering how I can do customer service when I'm away from home with a laptop that can't log into anything. Then again my laptop is a netbook from 2011 so lol. Not gonna fault LL for that. Still won't have any means of giving inventory or anything. I also have a crappy, super durable phone because I do volunteer first responder stuff and I need a phone I can drop in the snow or bleach if I have to call 911 and get body fluids on it. There are free and paid PBR texture libraries all over the internet. Since it's PBR it's a standard that works in several different game engines. If you can upload your own textures for things and apply them it's the same with PBR, just more maps. Content creators who don't read manuals could easily confuse spec strength and spec color maps, etc when downloading textures. You have to go out of your way to mess up PBR textures. If you're just doing basic texturing it's not like PBR is going to make this all go away. Just adds some more steps and your stuff is going to look way better. Biggest problem with PBR is going to be setting up reflection probes and lighting and that's where things are going to be a total mess. Interior scenes being blue from the sky or totally shadowed because no lights and no reflection probes is going to cause a lot of problems for people. If you were painting your diffuse in something like photoshop you are gonna have problems though. I would argue that PBR can actually make it easier to make really cool looking stuff without professional tools. I think people are just scared of change. But if you never made any content for a game (please don't kill me over that word), reading the wiki page for spec norm materials is a lot to digest. Downloading a PBR texture, selecting the GLTF file, selecting the drop down for the material you want, then SL automatically loading the maps, emission, and alpha settings for you and just pressing "save" is a heck of a lot easier. LL is never going to have viable competition. They have too much content. Even with SL's stability problems, lag, etc. it's extremely difficult to dethrone from it's position. Maybe things like Roblox but they went after a different market entirely. SL has the content to let you do pretty much anything you want. Like, you can think of the most obscure weird stuff and type it into market place and you'll find it. If someone wants to make a new open platform for content creators, they have to make it more appealing (profitable!) for content creators to leave SL, or at least add a new virtual world into their workflow. Or they have to pay people to make content. Both not very easy and very expensive. Even if you are a AAA studio and you pay for content you can easily end up with a GTA Trilogy: Definitive Edition situation where the content is all garbage anyways. Most of us probably remember that cringey stuff Meta did with their Metaverse stuff. That was a company several times larger than LL and that's what they gave everyone, lol