AI-Hear: read volumedetect from stderr, not stdout

OP OpenVibers committed 18 Aug 2026 · c673ff6d8a · 1 file, +5 −3

execFileSync returns stdout; ffmpeg writes volumedetect to stderr, so the level
regex never matched and logged mean=?dB. Use spawnSync and read .stderr.

Co-Authored-By: Claude Opus 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Xw6Sf5zzSALw8m93XNu5JE

Changes

ddiff --git a/server/ai/stream-audio.js b/server/ai/stream-audio.js
iindex e8e9899c..5eb82090 100644
---- a/server/ai/stream-audio.js
++++ b/server/ai/stream-audio.js
@@@ -147,9 +147,11 @@ async function captureWebrtc(stream, seconds) {
          // silent capture is an audio-path bug, on a loud capture it is a model problem,
          // and until now we could not tell the two apart.
          try {
--            const { execFileSync } = require('child_process');
--            const err = execFileSync('ffmpeg', ['-hide_banner', '-i', out, '-af', 'volumedetect', '-f', 'null', '-'],
--                { encoding: 'utf8', stdio: ['ignore', 'ignore', 'pipe'], timeout: 15000 });
++            // volumedetect reports on STDERR, so spawnSync (execFileSync returns stdout).
++            const { spawnSync } = require('child_process');
++            const r = spawnSync('ffmpeg', ['-hide_banner', '-i', out, '-af', 'volumedetect', '-f', 'null', '-'],
++                { encoding: 'utf8', timeout: 15000 });
++            const err = (r && r.stderr) || '';
              const mean = /mean_volume:\s*(-?[\d.]+) dB/.exec(err);
              const max = /max_volume:\s*(-?[\d.]+) dB/.exec(err);
              console.log(`[AI-Hear] stream ${stream.id}: capture level mean=${mean ? mean[1] : '?'}dB max=${max ? max[1] : '?'}dB (${size} bytes)`);