Help me understand max samples. (IRAY)
Flortale
Posts: 611
Scene A
- 4xGenesis 8 base figures
- max samples 1000
Scene B
- 1xGenesis 8 base figure
- max samples 1000
Will the Genesis 8 in Scene B look the same in quality as the four Genesis 8 figures in Scene A? I assume Scene A would just take longer to render?
Post edited by Flortale on
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Assuming that they both stopped for samples then they should be the same, if one or both stopped for time then they might differ in how converged they were.
Thank you for this confirmation and yes, they would both be stopped for samples. Good to know.
Do you happen to know of any reference charts that compares sample qualities? Does such a thing exist?
I don't think that would be possible - how many samples a sceen needs would depend on the scene (in geenral the more indirect light needed the more samples needed).
I'm assuming convergence ratio is the best factor for determining quality yes?
If you're render achieves 99% convergence, it will probably look absoultely stunning?
LOL, mine don't but that's often when things like when the last bit of glossy subtleness of skin gets rendered on and such I think.
99.8% then? Is there a huge difference between 99% and 99.8% ?
As I understand it, it means all rays are converged, it means the highest accuracy is reached in determining the pixel value. So the higher converge, the more accurate the picture. I dont know if its on a certain scale or curve, its prolly not logarithmic.
Also more rays (lights and bounces) = more time needed for all rays to reach convergence, hence the long times for indoor scenes.
I have my set at 95% and since people typically pay the most attention to humans in an image I notice the extra 3% - 4% results in a more naturally looking glossy skin, or wet skin, or sweaty skin, depending on what you are trying to do but not much improvment in other areas of the render than the viewer would likely notice.
it does depend which bits are last to converge - if there are a few "difficult" areas (eyes, for example, can be a problem) then they may still look very noisy. One option is to use the Spot render tool, set to render to a new window, to redo just the tricky bits and then add those to the base render (save as Tiff or PNG to get a mask); another option is to increase the render quality setting, which narrows the threshold for a pixel to be considered converged and so gives the tricky areas more chance to get caught up; the nw denoiser option in the beta can also help.
Another trick that I use fairly often is to render larger than my intended final image size. So if I want a 1600 x 1200 image in the end, I might render 50% larger (2400 x 1800) and then reduce the image size post render. The downsampling of the image will sometimes remove some of the noise from areas of the render.