What we tested
We generated three mono 16-bit PCM WAV files at 22,050 Hz. Each file contained the same isolated C4, E4, G4, and C5 sine tones. Every note lasted 0.75 seconds, each gap lasted 0.35 seconds, and the first note began after 0.25 seconds. Only the silence after the final source note changed.
The final source note ended at 4.300 seconds. We then added 0.25, 1.25, or 3.25 seconds of silence, giving total audio lengths of 4.550, 5.550, and 7.550 seconds. Each file went through the current local converter at its default controls.
| Trailing silence | Audio length | MIDI end | MIDI size | Events |
|---|---|---|---|---|
| 0.25 s | 4.550 s | 4.286 s | 105 bytes | 6 |
| 1.25 s | 5.550 s | 4.286 s | 105 bytes | 6 |
| 3.25 s | 7.550 s | 4.286 s | 105 bytes | 6 |
What the result means
Extra silence after the notes did not become a longer MIDI tail. All three exports ended with the final C5 note at 4.286 seconds, 0.014 seconds before the designed source-note end. The additional one or three seconds of silent audio did not create note events or extend the file's musical duration.
The three MIDI files were identical in the measurements that matter here: six events, one musical track, the same note names, starts, durations, velocities, and 105-byte file size. Every browser run returned HTTP 200 and produced no console errors.
Do you need to trim silence from the end?
- Not to shorten this converter's MIDI result. In this test, the exported file ended at the last detected note rather than at the end of the audio.
- Trim it when it helps your workflow. A shorter source file can be easier to preview and store even when the MIDI result is unchanged.
- Check sustained sound, not just digital silence. Reverb, room noise, or a fading instrument can still be detected as musical content.
- Inspect the last note in your DAW. Its detected end can differ slightly from the source boundary even without an empty tail.
What this test does not prove
It compares three exact PCM files containing four synthetic notes and true zero-valued silence. It does not test reverb tails, background noise, compressed audio, live YouTube-link input, complete mixes, or other converters. It also does not promise that every MIDI editor displays duration in the same way.
For silence before the first note, see the leading-silence timing test. For sound that continues after a note, compare the echo test and sustain test.