One audio finishing workflow
Choose a file, set an optional time selection, remove edge silence if needed, change channels, reverse, remove DC offset, add fades and set the final peak level. Process, listen to the preview, then download a genuine PCM WAV.
Know what changes
Speed changes the WAV sample rate and therefore changes pitch as well as duration; no time-stretching or resampling occurs for PCM WAV input. Fades use the output duration after speed changes. Mono averages channels, so opposite-polarity signals can cancel. Peak normalization overrides manual gain and is not LUFS loudness matching. Excess gain is rejected unless hard clipping is explicitly enabled.
Format support and limits
PCM WAV and float32 WAV are decoded by the local WAV engine. Other audio formats use your browser decoder and may be resampled to its AudioContext rate; codec support varies. Compressed or extensible WAV containers are not supported by the WAV parser. Input is limited to 40 MB and 12 million decoded interleaved samples. Audio decoding may consume more memory than the compressed file. No noise removal, speech transcription, multitrack mixing or metadata preservation is claimed.
Privacy and reusable settings
Processing stays on your device. Recipes store editing settings, never audio. Undo/redo changes settings, not source files. Exports support 16, 24 and 32-bit integer PCM without dithering.
Finish the processed selection with leading/trailing silence, up to ten repeats and gaps. Stereo balance attenuates one channel and requires two output channels. Polarity inversion changes signal sign. Repeats include the applied fades. Output is capped at 12 million samples.