Audio bitrate explained: what 32 to 320 kbps really do to file size and the frequencies you keep

“128 kbps is fine, 320 is best” is repeated everywhere and rarely explained. We encoded test audio with the same MP3 encoder our MP3 Compressor uses, at every common bitrate, and measured two things you can actually verify: how big the file is, and how much of the frequency range the encoder keeps. The results explain why mono does not shrink a file the way people expect, and why raising the bitrate on a file that is already compressed only makes it bigger.

The one formula that predicts every size

An MP3 at a constant bitrate (CBR) uses the same number of bits every second, so the size is bitrate (kbps) × 1,000 × seconds ÷ 8 bytes. We checked it on a 66-second stereo clip:

For other lengths, this is what the formula gives:

The last row is uncompressed audio. A 128 kbps MP3 is 11 times smaller than the same sound as a WAV, and that is before any quality question. Our 66-second clip is 11.2 MB as a WAV. The Audio Bitrate Calculator does these sums for any length and bitrate.

What each bitrate actually keeps

An encoder has a fixed budget of bits, so at lower bitrates it discards the highest frequencies first. We fed a pink-noise test signal (energy at every frequency, falling gently towards the top like music) through the encoder and measured the highest frequency left in the result:

Two readings matter. First, the curve flattens: going from 192 to 320 kbps buys about 1.3 kHz at the very top, while the file grows by two thirds. Second, mono keeps much more bandwidth at the same bitrate(16.6 kHz at 64 kbps against 10.9 kHz for stereo), because all the bits go to one channel. That is why mono is the right choice for speech, where stereo adds nothing.

This measures how much of the spectrum survives, not how music sounds. How much the missing top end matters depends on the material and on your ears and speakers, so for anything you care about, listen to a short sample at the setting you are considering before encoding the whole file.

Mono, stereo and sample rate: what changes and what does not

At a fixed bitrate the size does not depend on the number of channels. In our test, 128 kbps stereo and 128 kbps mono produced files of the identical size (192,679 bytes for a 12-second probe), and so did 192 kbps. Switching to mono only saves space if you also pick a lower bitrate, as the presets do. Their settings are:

The sample rate sets a hard ceiling: a file can never contain frequencies above half its sample rate, so 22.05 kHz audio tops out at 11.025 kHz and 32 kHz audio at 16 kHz. That is fine for a voice memo and the wrong choice for music. It is also why the lower presets cut file size so much for speech: they discard bandwidth speech never used.

Why raising the bitrate of an MP3 never helps

The sample MP3 we used averages 32 kbps. We re-encoded it at higher bitrates, and every step produced a bigger file with no new detail, because the detail that was missing was thrown away when the 32 kbps file was first made: 260 KB at 32 kbps, 520 KB at 64, 1,039 KB at 128, four times the original. Quality cannot be added back by spending more bits on an already-compressed source.

The MP3 Compressor guards against this. When we ran our 32 kbps file through all four presets, it did not honour the preset. Each time it lowered the bitrate to 24 kbps mono so the result would be smaller (the page said “lowered from 128 kbps so the file gets smaller”), and saved 65 KB. If you see that message, your file was already as compressed as the preset you chose, or more so.

Choosing a bitrate

Step by step

1. Work out the size first with the Audio Bitrate Calculator: enter the length and a bitrate.
2. Open the MP3 Compressor and add your file. It shows the file's own bitrate; if you choose higher, the tool will cap it.
3. Pick the preset that matches your audio, or use the advanced options to set the bitrate, mono and sample rate.
4. Compress, then play the start, the middle and the end before you keep the result.
5. To change the format instead of the size, use the Audio Converter, and trim silence off the ends with the Audio Trimmer first, since silence costs bits too.

Everything runs in your browser and your audio is not uploaded. The MP3 encoder itself is downloaded the first time you compress, so you need an internet connection for that step.

What we did not test

We tested one encoder at constant bitrates. We did not test variable bitrate (VBR), AAC, Opus or FLAC, and we did not run listening tests, so the table of bandwidths is a measurement, not a verdict on how each setting sounds. The 32 kbps sample clip is short and simple, and a different recording would behave differently at the lowest bitrates.