Guides · updated October 22, 2026

24-bit WAV to MP3

MP3 has no bit depth. Here is what happens to a 24-bit or 32-bit float WAV when it becomes MP3, and when the conversion is the wrong move.

Tested: 22 October 2026 · 24-bit/96 kHz and 32-bit float WAV fixtures · Chrome 156, Firefox 157, WebKit 27.2

Short answer

MP3 has no bit-depth field. A 24-bit WAV is decoded to samples, resampled if needed, and encoded to MP3 at the bitrate you choose. The extra bit depth is not carried; it may help the encoder resolve quiet detail, but the output is still lossy. Keep the 24-bit WAV as the master.

What “24-bit” means in a WAV

PCM WAV stores samples at a fixed bit depth. 16-bit gives 65,536 possible values per sample; 24-bit gives about 16.7 million. The extra resolution lowers the quantization noise floor and gives more headroom for processing. It does not add high frequencies or “detail” in the way marketing copy suggests.

A 24-bit/96 kHz stereo WAV is about 34.6 MB per minute. That is the cost of the extra resolution and sample rate.

What happens in the converter

Read the WAV header. The tool reads the actual format chunk: 16, 24 or 32-bit, integer or float, and the sample rate.
Decode to samples. The PCM values are turned into the browser’s audio sample representation.
Resample if needed. MP3 accepts 8–48 kHz. A 96 kHz source becomes 48 kHz; 88.2 kHz becomes 44.1 kHz.
Encode. LAME writes MP3 at the chosen CBR bitrate.
Write the file. The MP3 contains no bit-depth field and no 96 kHz sample rate.

The 24-bit source may help the encoder resolve very quiet passages, but the output is still a lossy MP3. The extra bits are not stored.

When to keep the WAV

Editing or mixing. A DAW wants the PCM master, not an MP3.
Archiving. The WAV is the full signal; the MP3 is a copy.
Further processing. Each lossy encode adds a generation. Start from the WAV every time.
A client deliverable that asks for WAV. Do not convert just because MP3 is smaller.

When MP3 is the right output

A phone, car stereo or hardware player that does not read WAV.
A podcast or upload form with an MP3 requirement.
A quick review copy for someone who does not need the master.
A file size that has to fit an email or a messaging app.

The dither question

Dither is a technique for reducing bit depth in PCM: it replaces quantization distortion with low-level noise. MP3 encoding is a different process; it uses a psychoacoustic model and a transform, not a simple truncation. This converter does not add dither.

If your mastering chain ends with a 16-bit WAV, apply dither there, then convert that 16-bit WAV to MP3. Adding dither after the MP3 encode is impossible; the MP3 is the final lossy step.

A note on “high-resolution MP3”

There is no such thing. MP3 is defined by MPEG-1/2 Audio Layer III, with a maximum of 320 kbps and 48 kHz for MPEG-1. A file labeled “24-bit MP3” or “96 kHz MP3” is either mislabeled or not an MP3. The converter will not pretend otherwise.

FAQ

Does 24-bit WAV sound better than 16-bit when converted to MP3?

The extra bit depth gives more headroom and a lower noise floor in the source. Whether that survives the MP3 encode depends on the material and the bitrate. MP3 does not store 24-bit samples, so the output is not a 24-bit file.

What happens to a 96 kHz WAV?

MP3 supports up to 48 kHz. A 96 kHz WAV is resampled to 48 kHz; an 88.2 kHz file goes to 44.1 kHz. This is a format limit, not a quality setting.

Should I dither before converting to MP3?

Dither is used when reducing bit depth in a PCM workflow. MP3 encoding is not a simple bit-depth truncation, and this converter does not add dither. If you are mastering to 16-bit WAV first, dither there; then convert the 16-bit file to MP3.

Is 32-bit float WAV supported?

Yes. The converter reads 32-bit float PCM and encodes MP3. The float headroom is useful while editing, but it does not make the MP3 lossless.

Can I convert a 24-bit WAV to a 24-bit MP3?

No. MP3 has no bit-depth field. A “24-bit MP3” label is marketing or a renamed file. The output bit depth is whatever the decoder produces, usually 16-bit or float, depending on the player.