MP3 to MIDI FLAC and MP3 consistency check: same source, same detector, same output.
In this controlled four-note run, the FLAC and 128 kbps MP3 encodes both produced four events with the same pitches and recorded boundaries; both silent controls produced zero events.
Input condition
The same source melody was encoded as lossless FLAC and 128 kbps MP3, then decoded to mono 11025 Hz and evaluated with the deployed app.js detector using confidence 0.8, minimum note length 100 ms, and transpose 0.
| Input | Observed output | What to inspect |
|---|---|---|
| source.flac | Four events: pitches 62, 65, 69, 74 | Lossless baseline boundaries |
| source.mp3 | Four events: pitches 62, 65, 69, 74 | Whether the encoded result matches |
| silent.flac / silent.mp3 | Zero events for both controls | Quiet-input behavior |
The complete recorded results.json contains outputs and SHA-256 records.
What this does not prove
This comparison does not prove that FLAC and MP3 always produce identical results, or that a short synthetic melody represents full-song accuracy. It documents only this source, these encodes, and this detector configuration.
After conversion
Keep the source format, decoder settings, detector parameters, and result file together when comparing a new encoding. If the input changes, rerun the same check instead of reusing this result.
Questions about FLAC and MP3 inputs
Did FLAC and 128 kbps MP3 differ in this run?
No. Both produced four matching pitch and boundary records.
What did the silent controls produce?
Both silent controls produced zero events.
Does this prove every song will match?
No. It is a controlled comparison of one generated melody and one detector setup.