OiDN Denoiser - Standalone
- Thread starter erikals
- Start date
Planeguy
Active member
Sure, here you go. This render was rewritten, so it's got a bit more noise.
Attachments
Planeguy
Active member
Here's another one.
Min samples 1.
Max samples 3.
AS: 0.018
render times around 20 sec per frame on an old i7 3740qm CPU.
Color graded the final sequence. Added a bit more of noise removal (in shotcut), rgb shift, and color graded. There's some degradation of detail. If you look at the lamp it dissapears, but for quick previz demos should work fine. For better results I have to increase either the min samples (with AS disabled) or have a combo of min samples, and higher AS threshold. Something like:
Min sample: 1
Max samples 25
Threshold: 0.015 (or so)
Filter radius at default values (0.5)
These settings gave neatvideo something more to work on, and produced really nice renders in my opinion. Render times for these settings at 800x400 something, were around 5-7 mins (not for the video below). I'm sure a balance can be found that gives improved speed without much loss of detail.
But as you can see, the original sequence has a tremendous amount of artifacts, that neatvideo was able to clean up pretty well.
Min samples 1.
Max samples 3.
AS: 0.018
render times around 20 sec per frame on an old i7 3740qm CPU.
Color graded the final sequence. Added a bit more of noise removal (in shotcut), rgb shift, and color graded. There's some degradation of detail. If you look at the lamp it dissapears, but for quick previz demos should work fine. For better results I have to increase either the min samples (with AS disabled) or have a combo of min samples, and higher AS threshold. Something like:
Min sample: 1
Max samples 25
Threshold: 0.015 (or so)
Filter radius at default values (0.5)
These settings gave neatvideo something more to work on, and produced really nice renders in my opinion. Render times for these settings at 800x400 something, were around 5-7 mins (not for the video below). I'm sure a balance can be found that gives improved speed without much loss of detail.
But as you can see, the original sequence has a tremendous amount of artifacts, that neatvideo was able to clean up pretty well.
Planeguy
Active member
Sorry, I sent you the wrong one lol, give me a sec. I'll send you a higher res too.but there is almost no noise in that one ?
Planeguy
Active member
Here you go. I can send you the sequence later if you want.but there is almost no noise in that one ?
Attachments
Planeguy
Active member
yikes! That looks noisy, but that's an interesting texture looking noise there hmmmm. Do you have AE? If you haven't tested neat video, give it a shot. It cleans the image quite well.
Planeguy
Active member
I'm gonna try that next, and upping the light samples. Problem is, on my old processor, any small changes make a big change in render times. My comfort zone is frames rendered in the 3 min cap.
lardbros
Not so newbie member
With the normal pass and object pass the results would surely be better wouldn't they Erikals?
lardbros
Not so newbie member
My bet is Blender is using the best buffers to assist OIDN with its denoising. It'll do this automatically.
In LightWave, you have to choose which buffer will give OIDN the best info to stop it destroying your shaders. Best to experiment with which normal buffer, or ID pass is the best for OIDN to retain your textures and bumps.
No, I compared rendered output from Lightwave that I denoised in blender and it preserves way more detail. Did the same with a Blender output in Lightwave and compared the result.
The blender denoised end result is much better than the Lightwave output.
I'm not talking about rendering. I simply used the final saved buffers and denoised them in both programs and compared. The blender ones kept much more fine detail (denoising the same source material!)