Hi,
I was trying bump (checkerboard) but I don’t think its correct, it should be hard edges, not smooth.
I tried Displacement also… but its also smooth.
Any other setting in bump channel…?
render view and bump material.
Thanks.
…
I’ll admit to having no idea what those two images show 
You mention using checkerboard in bump and displacement. Based on that, I added the C4D native Checkerboard surface to the bump channel of one CoronaMtl, and to the Displacement of another, and these are the quick and dirty results, bump on left, displace on right (used displacement World Size of 0.02 to give it fine detail and no artifacts).
Let me know more specifically what you were doing, if this was not what you meant.
Thanks!
Tom
Hi TomG,
Thanks for reply.
That was render screen shot and used bum.
I am from VrayforC4D, A simple scene I did and then I tried with Corona to do it, I am not getting same bump results with Corona. Vray bump is sharp even a small tiny bumps but in Corona I am loosing the details.
May be some other method for good result.
See c4d file attached and render screen shots also.
Thanks.
…
p.s.
I am sorry if I am comparing this vray.
bump test.zip (2.09 MB)
I tried it myself but I don’t know where your problem is.
The only thing that comes to my mind that would result in that noisy bump is that you have only let it run for one pass or so.
I used the c4d tiles shader(lines v-dir.) for my experiment but also tried a similar bitmap that also looked fine.
The only thing where artifacts occurred was with the displacement but that could easily be resolved by lowering the screen size setting and/or maybe unwrapping the uvs at the top&bottom.(though it was a pretty small object and not too close to the camera)
here’s my take on it:
Ring on a cylinder on a plane with corona sun.
Left: Bump / Right: Displ.
Anyone tired my test scene file…? given above.
Thanks..
…
I tried your scene.
The problem is with the displacement screen size. The default value of 2 is too high for the fine contours you need.
It’s also an adaptive approach and if the camera were closer, it would look fine.
The artifacts can be resolved by setting the displ. screen value to 0.4 or lower or if you want it to be static 0.001 cm will also do the job.
I’d also recommend to let it run for more than ten passes(I used 30 but less would already do it) and you’ll see a crisp line.
The denoiser only makes sense when the image has gathered enough samples and finer details are already visible. The default value for such a scene would also be too high and you’ll have to play around with the values until you’re satisfied.
One more thing for the normal map that was in your folder.
If you want to use those, you first need a corona normal and inside of that the corona bitmap shader and set the color space to linear.
It also looks like the y value should be inverted for that map.
Hope I could help.
I’ve tried it again. Sure there are small artifacts but the area in the original view is very tiny, pixel wide and wouldn’t be noticeable in a final render.
I used 16 passes. I had s.size 0.4 as max and a finer version with 0.2. In the close ups, I exaggerated the displacement to 0.15cm
1.original size 2
2. size 0.4
- close up 0.4 300% zoom
- close up 0.2 300% zoom
It should be cleaner but with final HQ settings, it wouldn’t be noticeable. Maybe a little tweak on the AA settings could do the rest.
I don’t know how the c4d shaders are handled. Some other renderers like Octane bake the shaders and you can choose a size.
edit: small mistake writing the numbers onto the pics 0.4, not 0.04.
Glad I could help.
There is no fixed HQ setting. This all depends on what you are going to create and tweak the values accordingly.
If you have a forest with lots of leaves for example, then a higher raydepth could be useful.
A scene that has a strong or many caustics could benefit from a higher value in the light sample multiplier and AA settings.
This is all very individual. Imagine it like tweaking the values in Vray (global subdivision multiplier, GI depth, dmc sampler etc. etc.)
Just learn what each value does, there aren’t that many. That’s the good thing of unbiased engines.
Then you can easily set up the rendering values for what you need. The documentation and the official YT tutorials are quite helpful.
thanks for update.
…
In this case I think the best solution would be to model the displacement, or use a modifier, like in this (3ds maxa) example here: https://forum.corona-renderer.com/ Problem with clean displacement
Bottom line is corona bump is terrible and has always been in C4D, I dont think its a problem in max otherwise we would have heard about it. I’m not a coder but there seems to be a stumbling block on this one. For me its coronas Achilles heal. Toms checkerboard example is a good one to show the poor bump performance. Perhaps the new bump converter shader coming in the next beta might help though? Quite often it helps to blur the map slightly in the shader - you get a more pronounced effect at the expense of sharpness.
I love corona but I have to admit the vray bump is undeniably better and much more reliable across its strength range and where using shaders or bitmaps.
Hi Rob,
There is still probably misunderstanding from our side what is your expectation about how the bump mapping should work. Could I ask you to create two simple scenes with one cube using the same setup (using the same bump map) first using VRay bump mapping and second using Corona bump mapping so we can compare the renders? We would like to improve bump mapping because it looks that current results are somehow counterintuitive but to be able to do it we need to know what is the desired result. Thank you
Nikola
Thanks for looking in to this again Nikola. I have put a link in to a drop box zip with the file and oak texture so hopefully that works.
Please see the attached images. I have just used the respective material preview ball scenes with a semi gloss grey diffuse base so we can just see the bump effect but there will be some slight calibration variation in hue. Also I have use the vray default of 1cm strength and picked 58% for the corona one as a sort of half way strength. Making it more or less doesnt seem to help / make much difference.
As a side note I personally prefer the cm scale range rather than a percentage - at least you can visualise roughly the amount you are dialing in in CMs, where as percentage is quite arbitrary especially if we can go over 100%. I wonder if there could be some calibration for this between black and white in corona?
So regarding the Oak bump images; I have just made a straight desaturated with enhanced curves jpeg of an oak texture and loaded that in to the bump slot in both mat editors. Straight out of the box you can see vray picks up every strand of wood grain and you get a real sense of a rough wood surface. There is a strong shading artifact in the preview ball but I think that is the way it is set up/lit. As for the corona one we just get a high frequency noise that doesnt represent the wood grain. If I ramp up the strength to 500% we still get this strange high frequency noise, but if i blur it by 0.5% magically we then get to see the wood grain but we have sacrificed a bit of texture detail in the process and I still prefer the detail and feel of the vray one as it is more fiberous where the corona one smooths out the subtle threads due to the bluring. Also we shouldnt need to blur this texture to see it!
Moving on to the checker. I have used the same strengths as the previous test (1cm and 58%) This is a hard test case as we are going from pure black to white in a pixel width. But vray has a reasonable stab at it and there is a bit of a surface there. Corona fails to create a surface. If I then blur the checker for vray giving a ramp between colours you get a good relief and you can feel that checker surface over the ball. Do the same with corona and its an even more noticable painted on strange cross pattern.
So hopefully this gives a clear guide on why i think vray is definately better at bump and works well with hardly any tinkering. Corona as it stands is always a battle to get the bump to work and work well with the right strength. I hope its a simple fix or a scan down the chaos group code now Im guessing you have access to it - although sadly not the vrayforc4d code. Thanks again and will always chose corona over vray regardless but it would be nice to get this fixed once and for all. Cheers.
Agreed. The Vray stuff does such a good job with bump maps. Hoping maybe they can bring some of that into Corona.
I’m also getting this result (Beta 1), its like the bump is expaning to the sides instead of upwards.
Annoyingly vrayforc4d is the only plugin not developed by chaos group so all that specific c4d knowledge is lost.