Well, IMO the results of the test in https://www.diyaudio.com/community/thre ... st-7616047 need more analysis. Just checking for presence of the IEC958 marks may not be enough.
Since @dom detected the corruption of the IEC958 bitstream with the HDMI analyzer, it may be useful to check the stream generated by the script above for these bits too, in the xrun positions of the data. That would identify if the corruption happens in userspace or later in the kernel driver/hardware.
The check should extract the corresponding bit from each sample, compose the 192-bit sequences, and check that all 192-frame blocks produce identical bitsequences within the large dump.
Since @dom detected the corruption of the IEC958 bitstream with the HDMI analyzer, it may be useful to check the stream generated by the script above for these bits too, in the xrun positions of the data. That would identify if the corruption happens in userspace or later in the kernel driver/hardware.
The check should extract the corresponding bit from each sample, compose the 192-bit sequences, and check that all 192-frame blocks produce identical bitsequences within the large dump.
Statistics: Posted by phofman — Wed Mar 27, 2024 6:54 am