Does anyone know of a way to set the bit depth of png renders from Karma?
They seem to always come out 8 bits per channel.
Karma PNG bit depth
418 4 1- pixelninja
- Member
- 31 posts
- Joined: March 2017
- Online
- Enivob
- Member
- 2539 posts
- Joined: June 2008
- Offline
- pixelninja
- Member
- 31 posts
- Joined: March 2017
- Online
I figured it might be unlikely but thought I'd ask just in case.
The comper is specifically requesting separate png sequences.
Bonus question for anyone out there:
Can you render AOVs as exrs but to the C(rgba) channel?
If I output an AOV as a separate sequence, depth for example, it will output an exr sequence with no colour data and depth in the depth channel, requiring compositors who use After Effects to make the extra step of extracting the AOV data out to be used. Effectively being forced to use the multichannel workflow on single channel sequences.
The comper is specifically requesting separate png sequences.
Bonus question for anyone out there:
Can you render AOVs as exrs but to the C(rgba) channel?
If I output an AOV as a separate sequence, depth for example, it will output an exr sequence with no colour data and depth in the depth channel, requiring compositors who use After Effects to make the extra step of extracting the AOV data out to be used. Effectively being forced to use the multichannel workflow on single channel sequences.
- Enivob
- Member
- 2539 posts
- Joined: June 2008
- Offline
The comper could use After Effects 3D tools to extract the depth pass, and write it out as a PNG sequence. Not sure what the point is, however. That would be downgrading your depth value from 16 bit to 8 bit, losing precision in the process.
Using Houdini Indie 20.0
Windows 11 64GB Ryzen 16 core.
nVidia 3050RTX 8BG RAM.
Windows 11 64GB Ryzen 16 core.
nVidia 3050RTX 8BG RAM.
- pixelninja
- Member
- 31 posts
- Joined: March 2017
- Online
Yeah just trying to avoid any double handling of renders.
I agree it's not ideal, I'm just trying to meet the preference of the comper.
The project is EXTREMELY large scale and so it's an odd case in that filesize is more important than precision. I'm already having to normalise the data to fit the 0-1 range for the pngs anyway.
I might be able to convince them to go with compressed exr sequences if I can get the data into the colour channel to avoid them having to extract it for every pass.
I agree it's not ideal, I'm just trying to meet the preference of the comper.
The project is EXTREMELY large scale and so it's an odd case in that filesize is more important than precision. I'm already having to normalise the data to fit the 0-1 range for the pngs anyway.
I might be able to convince them to go with compressed exr sequences if I can get the data into the colour channel to avoid them having to extract it for every pass.
-
- Quick Links