01:15 <triskit> FWIW, I am looking at the D1 for the same, but am doubtful of the quality of the mic preamps.

01:16 <triskit> Hoping to make a sort of a kitchen sink audio gateway thing. My idea was to use the internal ADCs to capture from an FM demodulator for wireless mics, and use external preamps and ADCs for wired mics.

01:16 <triskit> AFAIK spdif tx is supported but not rx. I assume rx is a bit trickier to implement since the driver needs to keep the FIFO at a consistent level.

01:16 <triskit> (In case you read the logs)

08:19 <blathijs> And the plot thickens - apparently the decoder problem (or another problem with the same symptom) I am chasing already exists in our original 6.1 build, it is just extremely rare (took over 700 playbacks to trigger it once last night)... So that complicates bisecting...

11:55 <blathijs> linkmauve: Ok, managed to bisect things back to this commit: https://github.com/torvalds/linux/commit/fec94f8c995 It does not *cause* the issue, but makes it significantly likely to occur. When I run 6.7.12 with armbian patches applied, I get the faulty decoding 20 out of 50 playbacks. With that commit reverted, only 5 out of 50 playbacks fail. So that does not tell me what the

11:56 <blathijs> problem itself is, but it might provide some clues about what is wrong. After lunch, I'll have a closer look at that commit and all of the cedrus code and maybe sprinkle around some debug output. If anyone has suggestions for things to check, I'm happy to hear them :-)

13:51 <timsu> triskit: Yeah I'm following along

13:51 <timsu> It is quite hard to find a SBC with SPDIF RX support.

13:53 <timsu> The datasheet says 80db THD+N for the audio adc so the quality should be okay

14:11 <triskit> I was also looking at the RK3308

14:25 <timsu> Also seems quite interesting, but I could find only two boards with it, none of them have the spdif rx pin connected.

14:26 <timsu> RK3588 also has spdif in with some more interesting features and quite active linux development, but no analog codec