Research on how to best encode data within audio stream

[vmdocua]

Research on how to best encode data within audio stream (e.g. sequence of beeps), and tools/reprostim-timesync-stimuli to present that audio cue upon win.flip so we could match that timing to qrcode information in visual stream.

• Review https://psychopy.org/api/sound/playback.html and possibly use PTB's facilities in PsychoPy for precise audio placement in time.
• Add representation with error correction, e.g. https://github.com/sfp932705/ecc .

[yarikoptic]

just a comment on some additional utilities we identified while researching online, if e.g. approaching from aspect of watermarking:

• https://github.com/wavmark/wavmark -- "AI"-based

[yarikoptic]

here is how I was adding distortion at the level of encoded data before going to audio stream

❯ git diff
diff --git a/tools/audio-codes.py b/tools/audio-codes.py
index abf1273..8464ed3 100644
--- a/tools/audio-codes.py
+++ b/tools/audio-codes.py
@@ -113,6 +113,9 @@ class AudioFrame:
         if self.use_ecc:
             b = bytes(self.rsc.encode(b))
             logger.debug(f"  - ecc       : {len(b)} bytes, {b}")
+            for i in [0, 1]:
+                # distortion
+                b = b[:i] + bytes([(b[i]+1)]) + b[i+1:]
         return b