Wolfy_527

142 posts

Wolfy_527 banner
Wolfy_527

Wolfy_527

@wolfy527

🧡 VRChat Avatar, World, and Content Creator. 🐾 Hobkin Lover! 🧡Never let a circumstance define you. 🐾

Katılım Ekim 2020
68 Takip Edilen186 Takipçiler
Sabitlenmiş Tweet
Wolfy_527
Wolfy_527@wolfy527·
Check out my VRChat, Unity, and Hobkin related discord server! Feel free to ask me questions and interact with fellow creators! discord.gg/sZydgrnPkV
English
0
0
5
1.1K
Wolfy_527
Wolfy_527@wolfy527·
Some super early progress pics of a set of street/techwear clothing I'm making for the OG Hobkin. Just the pants right now, but more to come! Planning on making a Furality Ultra setup for the clothing as well!
Wolfy_527 tweet mediaWolfy_527 tweet media
English
0
0
4
59
Wolfy_527
Wolfy_527@wolfy527·
@ImNetPet Scrolling emissions are awesome! You should check out pathing at some point if you ever want to dive deeper into scrolling emissions with audio link!
English
0
0
1
327
NetPet
NetPet@ImNetPet·
I found out how to do a scrolling emission... I don't think I'll have plain emissions on my avatars ever again
English
18
22
546
22.8K
Wolfy_527
Wolfy_527@wolfy527·
@LiindyVR Ran into the same issue when I started working on a dual saber like this a couple years ago and ended up just shelfing the project to do other stuff.
English
0
0
0
124
Liindy
Liindy@LiindyVR·
got most of the sounds working now the huge challenge with sounds for this asset is that each saber when active HAS to have humm, which is 2 sounds. And vrchats limit is 3 sounds MAX at the same time from an avatar. So everything else that emits sound has to come from 1 emitter
English
7
41
434
13.2K
Basis
Basis@BasisVR·
The big load test is this week! The video below was recorded from our regular small demo server
English
17
40
462
35.4K
Wolfy_527
Wolfy_527@wolfy527·
@BasisVR Absolutely! I just think it's awesome to see such a large lobby of people working so well in a VR setting! Stuff like this rivals what AAA MMOs strive for on a flat screen, let alone in VR. I'd love to check it out sometime if you guys do run open playtests at some point!
English
1
0
1
126
Basis
Basis@BasisVR·
@wolfy527 I can't speak to what VRChat does, but I'm sure they have done their best. A lot of smart people work at VRChat. Basis runs as well as it does from a lot of trial and error of what works performance-wise, we are down to break compatibility for more frames
English
1
0
1
94
Wolfy_527
Wolfy_527@wolfy527·
@BasisVR Are you guys already just splitting up a lot more out of the main thread than vrchat does then to get such good performance at such a large amount of people? I know the mainthread is definitely a big bottleneck for vrchat.
English
1
0
1
128
Basis
Basis@BasisVR·
@wolfy527 yep custom shaders, custom avatars, limits are very high by default and can be turned off for content that exceeds said limits. The main thread is always our bottleneck, so moving work off the main thread, as well as writing a lot of Unity jobs code, basically makes this possible
English
1
0
0
86
Wolfy_527
Wolfy_527@wolfy527·
@BasisVR what kinds of things are you reducing for efficiency? are people able to use custom shaders and such? i'd be interested to know how drawcalls and and polygon count are handled as things like that as well as IK, voices, and such can be a bottleneck for performance and networking.
English
2
0
0
113
Basis
Basis@BasisVR·
@wolfy527 Yes some are optimized and some are very much not, it's a lot of systems of reduction and efficiency
English
1
0
0
340
Wolfy_527
Wolfy_527@wolfy527·
We hit 1000 followers on Jinxxy! It genuinely means the world to me to be able to make stuff people love and enjoy! To celebrate, I have a store-wide 50% discount running on Jinxxy and Gumroad until midnight April 19th using the code: BWAH1000 Thank y'all for the support🧡🧡
Wolfy_527 tweet media
English
0
1
4
353
Wolfy_527
Wolfy_527@wolfy527·
Feeling a little dapper. New outfit soon!
Wolfy_527 tweet media
English
0
0
8
92
Wolfy_527 retweetledi
逻辑XR
逻辑XR@logicproducer·
1.VRChat到底是谁想出来的70,000面一刀切的性能标准,但凡给个140,000面的限制我都乐意优化模型,超过70,000面直接给你VeryPoor这谁受得了,我再优化骨骼优化碰撞优化mesh优化贴图都没用,超过70,000面就没用,在手机上跑的二游自机角色面数都能超过100,000。 2.一言以蔽之,这个阈值制造了一个断崖,而不是梯度。 3.这个断崖的问题在于,它把"70,001面"和"300,000面"放在同一个评级里。一个精心优化过骨骼、碰撞、贴图,只是主模型面数稍高的角色,和一个完全没有优化意识随手堆料的怪物,在系统眼里是同一类东西。这对有优化意识的创作者是一种惩罚。创作者做了90%的工作,因为最后那10%(面数)越线,所有努力在评级上归零。更坏的结果是它改变了创作者的行为——既然超过70,000就是Very Poor,很多创作者干脆不优化了,反正都是Very Poor,不如堆料。这个阈值在客观上鼓励了它本来想惩罚的行为。 4.一个更合理的评级逻辑应该是什么样的 面数只是性能开销的一个维度,而且不一定是最重要的那个。真正影响帧率的,权重大致是: 材质球数量 → DrawCall数量 → CPU瓶颈(影响最大) PhysBone数量和复杂度 → CPU物理线程(第二大) Shader复杂度 → GPU overdraw(第三) 面数 → GPU几何处理(其实在现代GPU上已经不是最大瓶颈) 面数排最后。VRChat的评级系统却把它放在最显眼的位置,权重也最高。这个优先级本身就值得质疑。 5.最后,怎么改? 升阈值固然是最简单的修复,但治标。真正治本的是把评级逻辑改成加权多维度综合评分,让面数回到它应有的权重位置,让材质和PhysBone承担更多评级压力。
逻辑XR tweet media
中文
26
147
1.4K
104.7K
Wolfy_527
Wolfy_527@wolfy527·
@dtupper @lil_xyzw Something I mentioned in the developer QA on Friday was a raise on the polygon limit. I'd happily sacrifice 8 material slots and skinned meshes on the limit to becoming very poor to get the ability to have like 30000 more tris on an avatar before becoming very poor.
English
0
0
0
99
tupper
tupper@dtupper·
@lil_xyzw 70,000 was chosen based on the median occupant value, which was around 10-20 at the time of evaluation. Using the 80 player example is a convenient way to show "this is how bad it could possibly get" but was not how the system was formed.
English
8
2
29
5.9K
lil
lil@lil_xyzw·
7万ポリゴンを80人インスタンス基準で計算してるならそれはnot for meな数値だから無視していいなという結論に至った🫠 ask.vrchat.com/t/developer-up…
日本語
2
60
362
59.2K
Atelier Elysara
Atelier Elysara@AtelierElysara·
@wolfy527 @VRUnderground nah taking a look at your jinxxy you could definitely reduce your polycounts a lot without losing any of the look of the model easily
English
1
0
0
16
VRUnderground
VRUnderground@VRUnderground·
I see there is a role for allowing very poor avatars. It might be tempting to pay wall that, but I don't recommend it. It might seem like a clever idea but doing so have consequences. Thoughts? #VRChat
English
8
0
19
1.3K
Wolfy_527
Wolfy_527@wolfy527·
@dtupper I don't just think it's the algorithm, but the English-speaking culture.
English
0
0
0
40
tupper
tupper@dtupper·
The more I think about it, the more I observe, and the more I participate, the more I convinced that this platform is an absolutely terrible place for actual, productive discussion with the goal of mutual understanding. There's no such thing as a discussion here, or someone spending time to understand the point of view of somebody else -- just hot takes, bad dunks, disinformation-as-fact, and anger. Even simply listening on this platform requires a collection of caveats that essentially invalidates your conclusions. You have to filter out so much because of what the algorithm encourages users to do that you end up losing most of your data alongside the trash. Twitter has become one of the least reliable data sources for social listening that we have on hand. Strangely, I feel a little guilty blaming just the platform. The reasoning for this isn't some kind of attachment or emotion, it's just that when I observe communities that are separated by a language barrier -- like the Japanese VRChat community here on Twitter -- the same issues simply are not present. It only seems to occur in the English-speaking sphere. Maybe that's the product of the algorithm? I'm not sure. I just know it didn't used to be this way, and now it is, and even if I want to try my hardest, this place just isn't healthy for your mind.
English
108
41
602
35.1K
Wolfy_527
Wolfy_527@wolfy527·
@IkigaiKitsune @VRUnderground I think that avatar "ranks" are just a slice of the pie on this limitation. IK, voices, and animators also become very, very taxing the more of them you have in one lobby, especially animators.
English
0
0
1
34
skel
skel@IkigaiKitsune·
@VRUnderground I'm curious if they can now raise the user cap of the instance with avatar limitations? I'd love to see this as the benefit of doing it.
English
2
0
2
106
Wolfy_527
Wolfy_527@wolfy527·
@VRUnderground @AtelierElysara I hear what you're saying when you say this, but I don't consider myself at "industry standard" with my capabilities. I would consider myself above average in my ability to create assets and avatars, but I feel like "industry standard" falls even further above my capabilities.
English
1
0
1
12
VRUnderground
VRUnderground@VRUnderground·
@AtelierElysara I agree, though, the difficulty is on par with someone who is industry capable and doing it for a living.
English
2
0
0
59
Wolfy_527
Wolfy_527@wolfy527·
@VRUnderground One of the main limiting factors here can be texture memory. But even that isn't very difficult if you only have 1 or 2 variations on each set of clothes.
English
0
0
1
9
Wolfy_527
Wolfy_527@wolfy527·
@VRUnderground I think the best part about what you said is that more can still be better, but it doesn't have to come at the cost it currently does. I can put 4 or 5 of my outfits (depending on the fit) on the Hobkin before it hits very poor. Under 70k tris, 16 materials, and 8 skinned meshes.
English
1
0
1
25
VRUnderground
VRUnderground@VRUnderground·
I cannot think of a single other game studio, dev team or individual that I have ever worked with where avatar limitations would not have been a day one design decision. Both in terms of stats and uploads. With that said its been so long that I am pretty used to very poors dominating every lobby at this point... We just never set limits in the VRU in general. I remember there was at least a few times I got blamed directly for allowing it lol... Now here is the real question, will people actively avoid lobbies that don't have poor or better limitations set? I suspect that we may very much get forced to use the setting or lose out on traffic. Why would any club goer actively choose to go to a place with no restrictions? Poor or better may become a standard. Thoughts?
English
6
0
8
935
Wolfy_527
Wolfy_527@wolfy527·
@VRUnderground Absolutely, I think limitation brings about innovation. Look at all the different kinds of assets people have created over the years to overcome different limitations. VRCLens, GogoLoco, VRCFury, and so many more out there. We just need more positive trendsetters in the industry.
English
0
0
1
25
VRUnderground
VRUnderground@VRUnderground·
@wolfy527 Its completely possible to be expressive within limitations. In fact, sometimes limitations are when the rubber meets the road on creativity and ingenuity. You can get a really accurate signal on residual self image within the current limitations.
English
1
0
1
16