Merlin bird app Sound ID does not work under /e/OS

i did reboot the device, uninstalled and reinstalled the app. re-granted mic permissions. still no spectrogtam. i was really interested about why the Merlin app might have worked up until 10 days ago. Just thought it was strange.

i am a human. i am experiencing a similar problem to the OP. I included all the information i uncovered in my investigations in case it might help someone else. your comment adds nothing to resolving the original issue.

gpt chats can be useful and good for you as an individual, but the signal to noise ratio can be bad when posted in full to a group, esp. when it crawls its own answer again and doesn’t filter it out, that ratio will further detoriate.

the thread was already ahead: isolated to the function name where things bail in an exception, established that the mic does record (ruling out permissions as cause), just the spectrogram not render. A lot of users of FP5 devices seem to run into this. The sampling rate is a likely candidate to be at issue.

1 Like

thanks for the info.

@szakib do you still have the raw logcat around before you filtered for ‘merlin’? would give you the whole stacktrace: grep '09-30 10:23:32.455 10197 6830' logcat.txt

the adjacent fairphone thread had the same ST with a given error of: java.lang.IllegalArgumentException: capacity < 0: (-257 < 0)

posted this to the merlin folk, they’ll know how the buffer capacity can end up with a -257 size

.. Merlin app [..] worked up until 10 days ago.

Is it possible that an update for /e/OS has rendered the Merlin app deaf over the last 2 weeks?

@jules22 so you did upgrade? if so, what version tags? would make sense to check for changes in the audio plumbing between the 2 versions then

The problem went away when the phone got updated to /e/OS 3.2-15-20251022539003-official-FP5

really wanted to satisfy my curiosity and see the exact change that did it, but I can’t source the 2 old images to compare (according to the changelog, FP5.TT4C.A.154 and FP5.TT46.A.144)

In LineageOS and e, the audio stack was previously imported from FP4 until the import from FP5.TT4C (lagoon_fp4 → yupikidp) that (also) fixed the mic issue - but the commit change is too big to easily point to something.. too bad