yes! i tried the v2 but ran into a lot of weird things atm, but i guess im waiting for the near finished release
Just a note, there seems to be lesser ray depth on the second picture : some internal reflections/refractions are absent. I suppose this can bias the results a bit. But the Blender one is definitely darker…
1 Like
The other engine must have some trick at work: there are way less inner reflections, also in some parts looks like light is added (see both sides of “C” pedestal). Somehow looks like thin/architectural glass.
Also the whole scene is more lit. Can you share the blend?
1 Like
It looks like the second image has a higher exposure or something.You can see that the whole image with the backround is brighter.
1 Like
Sure. Any idea where to upload? or should i send it to you via Discord?
Its hard to compare i know.
In Blender i used AGX workflow
And in the other one i used aces workflow, cause it was build in.
I tried to match exposures as good as possible.
To clear out i just did this test for my personal feeling to get a feeling for Cycles vs another common engine im using cause its giving me awesome results.
Despite the image being darker overall, I think the Cycles rendering looks more convincing, showing nice internal reflections.
3 Likes
Cycles glass may indeed be darker than other renderers, but maybe not as obvious as in your example. Here are some of my tests.
Cycles: 200s Multiscatter GGX:
Cycles: 200s Principled V2:
Luxcore: 20min path tracing+light tracing+caustic cache:
Luxcore:6min path tracing+light tracing:
Although I don’t know why the brightness of the Cycles rendering result is darker than Luxcore PT+caustic cach(SDS path???) and almost the same as Luxcore PT, but in this example, 20 minutes of sampling is still not enough for Luxcore PT+light tracing+caustic cache, resulting in smearing of the image.
Luxcore’s caustic cache may have some problems. Previously in one of my metal scene, turning on the caustic cache or indirect light cache would cause the metal to reflect ‘‘patterns’’ that did not exist in the world environment.
Uncompressed image: https://files.catbox.moe/tgi4wf.zip
Are you rendering the Blender version with Path Guiding in 3.4? Glass loses energy without it which can make it appear darker.
1 Like
Maybe a furnace test can give better results to compare energy conservation.
Ah thanks for bringing up our new sheen model 
I recently became aware of this push for a “Principled BSDF v2” and was also about to suggest it at some point. I’d also be happy to help with an implementation—though I likely won’t have any free time to work on this before the end of January.
24 Likes
I am really looking forward to Principled V2 Any news?
10 Likes
From what I remember there will be option to set nk values, but in separate “Metallic BSDF”, that what Lukas has stated before.
Has anybody else noticed that bump effect isn’t as good looking as in other renderers?
Particulary I’m talking about fstorm. Maybe it’s just me. I don’t have any evidence to prove this.
Just my eye It is like fstorm bumping really takes good care of those micro details.
2 Likes
you may be running into the issue where the bump texture is being interpreted as a regular sRGB color texture when it ought to be set to “NonColor”
I always set bump textures to non color if that’s what you mean.
I’m not saying cycles bump method is bad looking. It is looking good.
However I feel like it lacks that little extra oomph.
I’m plugging always into displacement slot as it gives me nicer result but I always need to raise scale value to high number to get those shadows out nicely.
I would have to render same scene in fstorm and cycles to really check out if this is true or not.
Maybe it’s just my feeling deceiving me.
2 Likes
yes, I’ve also noticed this. Something seems flat about bump in Cycles.
1 Like
Has development on PV2 been abandoned? It’s been a while since any word on it and 3.5 and now 3.6 are on the horizon and things seem to be dead silent on PV2. I’m concerned because to me this is the single most important development in Cycles in a while.
11 Likes
there is something wrong since quite a while, I think:
4 Likes