Make an audio file smaller by re-encoding it at a lower bitrate. The size estimate in the rail updates as you change the format, bitrate, sample rate, and channel count, so the tradeoff is visible before the export runs.
Where the size goes
Bitrate is the number of bits the encoder may spend per second of audio, and it is what determines file size. A 3 minute track at 128 kbps lands near 2.9 MB regardless of what is in it.
Lossy encoders reach that budget by discarding content that is hardest to hear: frequencies masked by louder ones nearby, detail above the range that survives most playback, and stereo information that duplicates between channels. Lower the bitrate far enough and the discarding starts to be audible, first as a smeared, watery quality on cymbals and sibilance.
Choosing a bitrate
| Material | MP3 | Opus |
|---|---|---|
| Speech, podcasts | 96 to 128 kbps | 48 to 64 kbps |
| Music, general listening | 192 kbps | 96 to 128 kbps |
| Music, close listening | 256 to 320 kbps | 192 kbps |
Opus reaches comparable quality at roughly half the bitrate of MP3, so it is the better choice when size matters most. MP3 is the better choice when the file has to open in older software or hardware.
Compress to a target size
Switch the size control to Target size and enter a number in megabytes. The budget is divided by the duration, and the highest bitrate that fits under it appears next to the field.
MP3 encodes a fixed ladder of rates: 32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192, 224, 256, and 320 kbps. A solved rate lands on the nearest rung below the budget rather than between rungs, so a 4.7 MB target on a 5 minute track resolves to 128 kbps and not 131 kbps. AAC and Opus accept any rate, which is why a tight target usually lands closer in those formats.
A target smaller than the format’s lowest rate can produce is not silently rounded up. The rail names the floor instead, and Opus is the way down from there.
Sample rate and channels
22.05 kHz discards everything above roughly 11 kHz, half the sample rate, which is where the highest audible content lives. Mono sums the channels into one.
Neither shrinks a lossy file by itself. The encoder spends the bitrate you set per second of audio whatever it is handed, so what changes is how much material those bits have to cover: 64 kbps of 22.05 kHz mono speech sounds better than 64 kbps of 48 kHz stereo. Pair the two controls with a lower bitrate, not instead of one.
For WAV and FLAC the effect is direct. Sample rate multiplied by channels multiplied by bit depth is the file, so halving either halves the output.
Re-encoding costs quality every time
Compression artifacts accumulate. An encoder working on already-compressed audio treats the previous encoder’s artifacts as signal and spends bits preserving them, then adds its own.
Compressing a 128 kbps MP3 again at 128 kbps produces a file of about the same size that sounds worse. Start from the highest-quality source available, ideally the original WAV or FLAC, and compress once.
Why the estimate is approximate
The estimate is bitrate multiplied by duration plus a small allowance for container overhead. Real encoders vary their rate, spending more on dense passages and less on quiet ones, so the finished file lands near the estimate rather than exactly on it.
WAV and FLAC appear in the format list but have no bitrate control. FLAC compresses without discarding anything and typically lands near 60% of the WAV size, which is the option to reach for when the file has to stay lossless.
Not a dynamic range compressor
Two unrelated processes share the word. This tool reduces file size. A dynamic range compressor reduces the gap between the loudest and quietest parts of a performance, changes how the audio sounds, and does not make the file smaller.