How do I tell whether the audio inside is MP3 or not?

The surest route is to look with a tool that shows the file's track information; most media players write the audio codec into their file properties or media information screen. A rough clue is the origin of the file: in film and series archives compressed with DivX or Xvid the audio is usually MP3, in DVD transfers it is likely to be AC3, and in old camera recordings PCM. If you are not sure, the conversion tool will detect it itself anyway.

If lossless copying is possible, do I need to pick a bitrate?

No, and in fact it is better not to pick one. On the copying route the audio data is carried across untouched and whatever the original bitrate was is preserved. Imposing a target bitrate can push the tool into an unnecessary re-encode, and the only thing you will gain is quality loss. If the source is already MP3, the best strategy is not to touch any setting at all.

Would it be better to take AC3 audio into some format other than MP3?

It depends on your purpose. If you only want to listen, MP3 is perfectly sufficient and at a sensible bitrate the loss is inaudible. But if you are archiving and want to lose as little information as possible, exporting the AC3 track in its own original form, or taking it into a lossless target, is the better move. You can produce the MP3 separately as a listening copy.

If I convert the same AVI file twice, does quality drop further?

If you start again from the original AVI file each time, no, the result will be the same quality every time. Loss is related to the length of the chain, not the number of conversions. But if you take the MP3 you produced, convert it to another format and then back to MP3, each step throws away a little more information and the degradation accumulates. Always convert in a single step from the most original source.

AVI to MP3: Is Re-encoding Unavoidable?

There is a genuine fork in the road

In most video containers, the answer to "does turning the audio into MP3 cost quality?" is one word: yes. In WEBM definitely yes, in MOV almost always yes, in MKV most likely yes.

AVI is different. Here there is a genuine fork, and which branch you take changes the result completely:

If the audio inside is already MP3: lossless copying is possible, loss is zero.
If the audio inside is some other codec: re-encoding is compulsory, there is a small loss.

This article puts those two branches side by side and makes clear how to work out which one you are on, and what to do in each.

Branch 1: Lossless copying

When it applies: If the audio track inside the AVI container is already compressed with the MP3 codec.

How often does this happen? The period when AVI was widespread was also the period when MP3 was widespread. Using MP3 as the audio track in film and series files compressed with DivX and Xvid was the standard practice of the era. So if you are dealing with an old film archive, the chances that you are on this branch really are high.

What happens: The audio data is taken out of the container and written directly into a .mp3 file. No decoding, no recompression, no psychoacoustic model coming into play. The data stays bit for bit the same.

Quality: Loss is zero. The result is an exact copy of the original audio track.

Speed: Noticeably fast, because the processor does not have to do any compression. Even on a long file this stage takes little time — the actual time goes on reading and transferring the file.

Setting advice: Do not touch any settings. Imposing a target bitrate can push the tool into an unnecessary re-encode and the only thing you will gain is quality loss. On this branch the best strategy is not to intervene.

Important warning: This does not mean you can change the extension to .mp3 by hand. The AVI container's own header and index structures sit inside the file; they do not match the structure an MP3 file expects and most players cannot open such a file. The audio track genuinely has to be separated from the container.

Branch 2: Re-encoding

When it applies: When there is AC3, PCM, ADPCM, WMA inside — that is, anything other than MP3.

What happens: The audio track is decoded first, opened back up into a digital waveform, then compressed again with the MP3 algorithm.

Quality: Some loss is unavoidable. How much depends on the source:

If the source is AC3 or WMA (lossy): There is a double-encoding situation here. While compressing, AC3 threw away some information according to its own psychoacoustic model. The MP3 encoder takes that already-reduced audio and applies its own reduction on top. Because the two formats discard different information, the losses do not cancel each other out, they stack. At a sensible bitrate the result is still satisfying, but technically there is loss.

If the source is PCM or ADPCM: The situation is better. Because PCM is uncompressed raw audio, the MP3 encoder starts from a clean source. Lossy compression is still applied — MP3 discards information by definition — but there is no double-encoding problem and the result comes out better than from an AC3 source at the same bitrate.

Speed: Slower than copying. On long archive files the process can take minutes.

An extra factor — channel downmixing: AC3 tracks in DVD transfers are frequently multichannel. Because MP3 is in practice stereo, those tracks are reduced to two channels. The dialogue in the centre channel now shares the same space as music and effects; that is why speech sounds more buried on headphones. This is not a fault, it is the natural consequence of reducing the channel count.

Setting advice: Pick a bitrate that suits the source and the content — 128 kbps for speech and old family videos, 192 kbps for films and mixed content, 256 kbps and above for recordings dominated by music.

Which branch am I on, and how do I tell?

The sure route: Most media players write the audio codec into their file properties or media information screen. Looking there takes seconds.

A rough clue — guessing from the source:

DivX/Xvid film and series archive → most likely MP3
File transferred from a DVD → most likely AC3
Old camera recording, VHS digitisation → likely to be PCM
Old game video → could be ADPCM

The most practical route: You do not have to make the determination yourself. Our AVI to MP3 converter reads the audio track inside the container and prefers the lossless route where that is possible.

The decision section

If there is MP3 inside: Do not think twice, extract it by copying and do not touch the settings. In this scenario the result you get is an exact copy of the original and better is technically not possible.

If there is PCM inside and it has archival value: Consider producing two copies — a high bitrate MP3 for listening, a lossless target for keeping. PCM is a clean source, and reducing it to a single lossy copy is a choice you might regret later.

If there is AC3 inside and you are only going to listen: 192 kbps MP3 is more than sufficient. If the dialogue comes across quiet because of channel downmixing, apply level normalisation in an audio editor after you have taken the MP3.

If there is ADPCM or a low quality track inside: Do not try to gain quality by choosing a high bitrate. The source is the limiting factor; 128 kbps will usually give the same result in a smaller file.

The rule that does not change

Whichever branch you are on, two things always apply:

Avoid intermediate conversions. Converting the AVI to another format first and then turning that into MP3 stacks lossy steps on top of each other and needlessly worsens the result. Always convert in a single step from the most original source.

Conversion cannot make audio better than its source. The audio in an archive file from the nineties was recorded according to the constraints of that period. No bitrate setting will enrich a track that is mono, low sample rate or narrow in frequency range. When working with AVI, the most realistic approach is to try to preserve the best version of what you have — not to improve it.