Tweet
The big part of success of Fstorm is that it defaults to a digital camera rather than neutral linear computer renderer. I've tested it briefly, and tone mapping settings default to average camera response curve, DoF settings default to camera objective, and bloom/glare settings default to secondary optical effects of average DSLR camera.
Everytime I tried to do same with V-Ray, I have unfortunately failed. The weakest point seems to be V-Ray's lens effects, that just do not look real no matter what i try. Real glare and bloom has not only intensity, but also size based on intensity of source image highlight. V-Ray lens effects has constant glare size regardless of intensity, and instead of adding glare on top of the image, it simply blends blurry image with base image, giving you very hazy and dreamy look even when using very slight glare or bloom effect. (I do realize that is by design, to prevent increasing HDR image intensity beyond what it really is, but I would much rather take realistic glare that I can use on my final image, instead of fake looking one, that I can't, even at the cost of messing up HDR spots on my image.)
Another issue is absence of photographic tone mapping. Physical camera exposure controls that have this partial functionality (something similar to filmic tone mapping) are not supported by V-Ray. V-Ray luckily supports LUTs, that can load arbitrary camera response LUT, but those are not saved out with image. Sure you can do all of this in post, but it really changes your workflow, if you build a scene looking at it through camera objective rather than neutral linear renderer.
Next one is that V-Ray still defaults to Blinn BRDF and no internal reflections for refractive materials. That alone means that if you are user, that is not aware of advanced BRDF choices in V-Ray (you are for example just artsy guy, not tech oriented type), your Fstorm shading will come out a lot nicer. GGX reflections and highlights will look a bit deeper and crisper, and bunch of wine glasses on table will look a lot more detailed and interesting with internal reflections enabled.
So it all comes down to people doing just a few quick test, and getting very pretty results. It's love on first sight effect. It's not just about if it is possible to do with your tool, it's about how easy it is to do, what people often value more. Most of them do not even know you could do most of it (not all though, due to the limitations in V-ray lens effects) in V-Ray. They just run a renderer that has good defaults, and they see something that looks better, more pleasing to the eye. It takes their usual work, their current skill level, and pushes the output quality little bit higher without any effort from their side. And anyone who is ambitious when it comes to output quality is willing to do anything to push quality even 5-10% further, even if it means completely switching to a different renderer.
I am sure some regular good old people will come in the defense, using the "Just learn to use your renderer" argument, but the thing is, that these days most people don't want to, and frankly, don't even need to. If you can achieve 8/10 quality out of the box with minimal effort spent learning and digging into deep technical stuff, why would you take a lot more time consuming and sometimes headache-inducing path to arrive to the same 8/10 quality mark?
Everytime I tried to do same with V-Ray, I have unfortunately failed. The weakest point seems to be V-Ray's lens effects, that just do not look real no matter what i try. Real glare and bloom has not only intensity, but also size based on intensity of source image highlight. V-Ray lens effects has constant glare size regardless of intensity, and instead of adding glare on top of the image, it simply blends blurry image with base image, giving you very hazy and dreamy look even when using very slight glare or bloom effect. (I do realize that is by design, to prevent increasing HDR image intensity beyond what it really is, but I would much rather take realistic glare that I can use on my final image, instead of fake looking one, that I can't, even at the cost of messing up HDR spots on my image.)
Another issue is absence of photographic tone mapping. Physical camera exposure controls that have this partial functionality (something similar to filmic tone mapping) are not supported by V-Ray. V-Ray luckily supports LUTs, that can load arbitrary camera response LUT, but those are not saved out with image. Sure you can do all of this in post, but it really changes your workflow, if you build a scene looking at it through camera objective rather than neutral linear renderer.
Next one is that V-Ray still defaults to Blinn BRDF and no internal reflections for refractive materials. That alone means that if you are user, that is not aware of advanced BRDF choices in V-Ray (you are for example just artsy guy, not tech oriented type), your Fstorm shading will come out a lot nicer. GGX reflections and highlights will look a bit deeper and crisper, and bunch of wine glasses on table will look a lot more detailed and interesting with internal reflections enabled.
So it all comes down to people doing just a few quick test, and getting very pretty results. It's love on first sight effect. It's not just about if it is possible to do with your tool, it's about how easy it is to do, what people often value more. Most of them do not even know you could do most of it (not all though, due to the limitations in V-ray lens effects) in V-Ray. They just run a renderer that has good defaults, and they see something that looks better, more pleasing to the eye. It takes their usual work, their current skill level, and pushes the output quality little bit higher without any effort from their side. And anyone who is ambitious when it comes to output quality is willing to do anything to push quality even 5-10% further, even if it means completely switching to a different renderer.
I am sure some regular good old people will come in the defense, using the "Just learn to use your renderer" argument, but the thing is, that these days most people don't want to, and frankly, don't even need to. If you can achieve 8/10 quality out of the box with minimal effort spent learning and digging into deep technical stuff, why would you take a lot more time consuming and sometimes headache-inducing path to arrive to the same 8/10 quality mark?
Comment