Audio Trimmer & Normalizer

How it works

The dropped file is decoded with the Web Audio API's decodeAudioData into raw sample data, and the waveform is drawn by taking the min/max sample within each pixel-wide column. Trimming just slices the underlying Float32 channel arrays between the two handle positions, and the optional peak-normalize step scales every remaining sample by a single gain factor so the loudest one reaches 0.98. Export hand-writes a standard 16-bit PCM WAV header around those samples — no external audio library involved.

FAQ

What does "peak-normalize" actually do?

It finds the single loudest sample in the trimmed selection and multiplies every sample by whatever gain brings that peak up to 0.98 (just under full scale) — it's one scaling factor applied evenly, not per-section loudness leveling.

What audio format does it export?

Always 16-bit PCM WAV, regardless of the format you uploaded — it decodes your source file to raw samples via the Web Audio API and re-encodes those samples into a WAV header, so the output isn't a re-encoded copy of a lossy format like MP3.