ProRender vs Modo vs Cycles – PBR 3D Render Comparison
Thomas Cheng writes:
ProRender vs Modo vs Cycles – PBR 3D Render Comparison
This is a comparison of three 3D engines that supports PBR textures.
The software used is:
- AMD Radeon ProRender 1.6
- The Foundry Modo 12
- Blender Cycles 2.79b
1-ProRender
2-Modo
3- Cycles
Unfortunately I have to agree with you…
I srongly disagree
1 cycles better vivid color nuances (darker areas not as if it was mixed with black)
2 ProRender vivid problem (black mixed) but faster than modo
3 Modo vivid problem and slow as well.
As if cycles got way more samples, or knows something about colors that the others dont know
As for light differences thats the sun size setup, cycles seams to have had a large sun here.
Of the three cycles did seem a little over bright and lacking detail in the scratches. But it’s that or spend another 2 extra hours with Modo, but I’m liking the look of prorenders results. Always nice to see the differences a new render can bring.
I hate these nonsense videos just to make money… Anyway, you give the scene to Cycles user and it will make that render look much better than you achieved. I’m not sure why you made it look so flat. Have you even consulted in a Blender forum?
1 – he doesn’t have advertising enabled on his channel
2 – if you don’t like it, don’t watch it
Sorry about the advertising … Just appeared a banner with suggestions that I closed quickly, thinking it was advertising. I’m really sorry.
The rest is a real criticism. In Blender you could make it look better even in Eevee right now. So I’m not sure why user did not manage good metalic look on Cycles.
I would have made the same criticism if any other engine had been the one that looked bad. I am sure that this is a fairly simple scene that should look quite similar in most modern engines, with good configurations in each of them made with users with experience in each of the engines.
Actually the point they are trying to make is a bit different. The video is purely intended for illustrating the raw capabilities of the render engine. As stated, the workflow was utilizing principled bsdf node in cycles, which is a relatively new solution. All that aside, render engine isn’t substitute for the massive amount of overhead post processing that can make any render look good, effects like scratches are usually added in post-production as they allow a more flexible workflow when faced in compositing node based environment.
Cycles is a very capable render engine but the point of whole video isn’t about “which is better” but rather to only put them side by side and show the individual which they’d prefer for a project. If somebody was to compare render engines on this forum and base their preference and get biased towards one (which isn’t exactly more than comparing mathematical simulations of light hitting an object and then being approximated by a camera and then favoring the raw result of either just because it looks good as it is) then they are wasting their time.
“2 – if you don’t like it, don’t watch it” – that’s pretty stupid argument. If you don’t like his comment, don’t read it. C’mon Bart, be a better person and not some offended snowflake when you read criticism.
I updated the description section to be more clear. Hope that helps.
“I tried to make this render as close to each other as possible. All three setups use the same HDRI, FBX, and PBR texture set. All 3 uses the same settings for the Principled Shaders. RPR calls their version of Principled as Uber, but it is essentially the same thing. The setup process was to import the FBX, apply the Principled/Uber shader and connect the textures. I didn’t tweak the shader settings as I wanted to see how the default settings look. Due to the different parameters of each render engine, I adjusted the settings to derive at similar noise levels. However, all three engine had at least 3 GI bounces.”
I’m in agreement with Raul. A little more time spent in the setup and it would have been well on par with the other two, and of course most likely the render time faster.
I found it rather interesting how different the textures looked with each render. I would not of expected such a deviation.
For further tests, I would like to see a target look and how much time it took to get that look realized in each material/lighting setup. Render time is important, but knowing how much time it took to get the setup is even more so.
It was all really simple and basic. Drop in FBX model, apply principled shader, connect appropriate PBR textures. No real tweaks to the shader to see raw output. I lower the samples in Cycles to 128, and increase the samples with RPR to 256. Both have denoiser on. In Modo, I up the GI bounce to 3 and increased the AA to get less noise, doesn’t have a denoiser. This model and PBR texture have both dielectric and conductor surface, so the base Principled shader fits well.
Probaply different sun setups (cycles larger shadow/light effect).
As for the material cycles seams to look better (color is more vivid, the others color-ramps towards black).