"sudo apt install libdrm-tests"Hi,
thanks for the update. I can get “kmsprint” result as “Connector 0 (32)”. I was able to switch to ycbcr444 without any problem.
I get “modetest” command not found error. I did not install it because the other link is for pi4.
What other link was for Pi4?
Selecting between RGB full, RGB limited, and YCbCr (limited) switches between different colour conversion matrices in the HDMI block. That's what is defined a https://github.com/6by9/linux/blob/rpi- ... 1234-L1261 for Pi0-3, and https://github.com/6by9/linux/blob/rpi- ... 1315-L1444 for Pi4.There is a problem in my work, when I compare it with other pattern generators, RGB Limited, RGB Full Range, YCbcr444, I get different measurements in RGB white level on the screen. the results are not far away but not close.
Is there also an algorithm that regulates the colors by xrandr or hdmi driver. I would appreciate your help.
YCbCr does have a number of options for exactly which matrix to use, with BT601 being the standard for SD resolutions, BT709 for HD, and BT2020 for UHD (4k). The "Colorspace" property on the connector allow selection, although there seems to be some debate in mainline Linux as to whether that actually change the matrix, or only the metadata sent to the display.
I've only implemented BT601 for Pi0-3, but can look at adding BT709. BT2020 probably isn't worth it.
YCbCr does inherently have some colours from the RGB space that can't be represented as they run out of range in the values (reverse is also true). White should be correct though. Can you give some numbers to it.
We haven't designed the Pi as a reference pattern generator, so whilst we do our best to make sure the configuration is correct, don't go calibrating anything from it.
Statistics: Posted by 6by9 — Sat May 04, 2024 6:12 am