Repository navigation
Stem presence re-decodes every stem via ffmpeg #287
Description
Location: analyze.py:217
Severity: HIGH
compute_stem_presence spawns one ffmpeg per stem (6 subprocesses, 180 s decode each) immediately before compute_stem_peaks reads the same files again via soundfile. Two full passes over ~1.5 GB of WAV data per job.
Compute RMS in the same streamed pass as peaks — one read, zero extra subprocesses. Saves ~10–30 s per job on HDD/older machines.
From the health report. Part of #274 (Phase 3 — Performance).
Activity
perf(pipeline): single-pass streamed peaks + presence (stemdeckapp#306)