Does converting audio lose quality?

They are about to convert something they care about and want to know what it costs. Every ranking tool page dodges this because the honest answer discourages the conversion — which is exactly why answering it properly is the strongest trust signal available to a site with no ratings.

Usually yes, and the size of the loss depends entirely on what you started with. Converting an M4A to an MP3 means decoding audio that has already been compressed once and compressing it again with a different encoder, so a little more detail goes each time. At 192 kbps or above, on a single pass, almost nobody can hear the difference. Converting an M4A to an .m4a, an .aac or an .m4r ringtone costs nothing at all, because the audio is copied rather than re-encoded. And no conversion in either direction can bring back what the first encoder discarded.

The three cases, and which one you are in

Every audio conversion falls into one of three buckets. The bucket, not the file extension, decides what it costs.

1. A stream copy. The compressed audio is already legal inside the container you asked for, so it is lifted across untouched. No decoder, no encoder, nothing lost.
2. Lossless to lossless. FLAC, ALAC and WAV all store every sample. Moving between them changes the file size and nothing else.
3. A fresh lossy encode. Anything that ends up as MP3, AAC, M4A or OGG Vorbis, from a source that is not already that codec, is a full decode followed by a new compression pass. This is the one that costs something.

The converter on this site tells you which row you are in before it starts, because it reads the actual codec inside the file rather than trusting the extension. A file called song.m4a can hold AAC or Apple Lossless, and those two behave completely differently.

Why the second encoder guesses worse than the first

A lossy encoder does not shrink audio by being clever about compression. It shrinks it by deciding which parts you will not notice are missing, and throwing those away. A loud cymbal masks the quiet detail underneath it, so the encoder spends no bits on that detail.

When you re-encode, the second encoder is handed the first one's output as if it were the original recording. Some of the masking sounds it would have relied on are already thinned. Its own model of what you can hear no longer matches the material, so it makes a slightly worse set of decisions than the first encoder did.

It is worse still between different codecs. MP3 and AAC both work by splitting the signal into frequency bands and quantising them, but they do not use the same bands or the same block lengths: an MP3 granule is 576 samples, an AAC long block is 1,024. What one encoder rounded off does not sit tidily inside the other's grid, so the second pass has to spend bits describing the first pass's artefacts. That is why an M4A converted to MP3 at the same nominal bitrate is not quite as good as an MP3 encoded from the original recording would have been.

On one pass this is small. The problem is that it stacks, and each round is permanent.

Turning the bitrate up does not help

The instinct when converting something you care about is to pick the biggest number available. A bitrate higher than the source had cannot add detail that is no longer in the file — it only makes the output bigger, and it stores the first encoder's mistakes in more precision.

If you have a 128 kbps M4A and you export a 320 kbps MP3, you get a file roughly two and a half times the size holding the same missing treble. Matching the source bitrate, or dropping slightly below it, is the honest choice. Our converter refuses to pretend otherwise and warns you when the number you picked is above the source.

The ceiling is worth knowing too. 320 kbps is the highest bitrate the MP3 format defines, and it is not CD quality: a CD is 1,411 kbps of uncompressed PCM. Most listeners cannot pick 320 out from 192 in a blind test. What 320 buys is a slightly higher treble cutoff, in a band above much adult hearing — 320 kbps vs 192 kbps has the measured cutoffs and how to test yourself.

The conversions that genuinely cost nothing

These are the ones worth knowing, because they are the ones people needlessly re-encode:

• AAC in an .m4a → .m4a, .aac, or an .m4r iPhone ringtone. Same codec, different wrapper.
• The soundtrack inside an .mp4 or .mov → .m4a or .aac. Phone video is almost always AAC already, so the audio comes out with no encoder involved. See extracting audio from MP4.
• ALAC → ALAC, MP3 → MP3, Vorbis → OGG, PCM → WAV. Container work only.

All of these finish in about a second regardless of length, because a copy moves bytes rather than re-encoding them. Trimming survives the copy: the converter seeks to your start point and still carries the stream across untouched, so the cut lands on the nearest compressed frame, within about a fortieth of a second of where you put it. Changing the sample rate, downmixing to mono or applying any effect does force a re-encode, so decide on those before you pick a format.

One word needs care here: identical. We measured a 243,340-byte .m4a converted to .m4r: the compressed AAC stream extracted from both files matched exactly, all 244,087 bytes, and the decoded PCM matched exactly, all 2,117,710 bytes. The files themselves were the same length but differed from offset 240,994, inside the trailing moov metadata block, because FFmpeg writes its own container tags and timestamps. So the audio is identical — the file is not, and any tool claiming an identical file hash is describing something it did not check. Tags and artwork may not survive the trip either.

Making the loss inaudible when it is unavoidable

Sometimes you have to re-encode — a car stereo that only reads MP3, a transcription service that rejects M4A, an editor that will not import AAC. Four things keep the cost down to nothing you will hear:

1. Convert from the best copy you have, not from a file that has already been through a converter twice.
2. Stay at 192 kbps or above for music. Speech survives much lower; 96 kbps mono is fine for a voice memo, and much smaller.
3. Do not ask for more than the source had. If the M4A is 128 kbps, export 128.
4. Keep the original until you are certain the new file works where you needed it. Converting again from the original always beats converting the conversion.

If you want to convert an M4A now, the M4A to MP3 tool shows the source codec and bitrate it detected, and says which of the three cases above applies before you press anything.

The thing that actually destroys audio

One conversion at a sensible bitrate is close to invisible. What ruins a file is a chain of them, each at a lower bitrate than the last: the voice memo converted to MP3, emailed, re-compressed by a messaging app, converted again by whoever received it, then trimmed and exported once more. Every pass throws away a bit more, and the artefacts of the earlier passes get encoded as if they were music.

You can hear this happening once the chain gets long. Cymbals turn to a hiss with no shape, sustained vowels acquire a warbling, watery edge, and the top end vanishes entirely. It is called generation loss, and it is the audio equivalent of photographing a photograph.

The fix is structural rather than technical: always go back to the earliest file in the chain and convert once from there. If you are handing a file to someone who will edit it, give them the original or a lossless copy, even though it is bigger. Their export will be a first-generation encode instead of a fourth.