Originally Posted by sordna
NVIDIA folks, I narrowed down the problem to the HDMI audio!
If I modify the EDID and zero out one or two bytes from the 2nd 128byte block
to get the HDMI audio disabled (I discovered the relevant byte offsets it via trial and error)
then the monitor goes to sleep. Since we are observing the issue in 2 different monitors, it looks
like a driver issue.
Possibly, the solution is: upon initiating powersave, before shutting off the video signal,
the driver should should shut off the SPDIF audio signal. Upon ending the powersave,
the driver should enable the SPDIF audio signal after the video signal is restored.
I am available to try a development driver to test this for you.
I have attached a tarfile that includes the original and modified edid files for my monitor (includes hex editor (actually vim) text dumps to see the difference).
Interesting!! I'll mention a possibly related observation. The Viewsonic VX2835wm has a separate
audio input which I have tried to use and I'm using DVI->HDMI. When the edid is enabled, the
audio on the Viewsonic VX2835wm does not work. With UseEDID false, the audio works. I had
assumed that this was an intrinsic "feature" of using DVI->HDMI at the hardware level in the monitor.
Apparently the NVidia <==> Monitor interaction is a bit deeper. Is the audio being intentionally or