Guide

M4A, MP3, WAV, MOV: getting text out of whatever you recorded

The format your recording happens to be in should not decide whether you can transcribe it. Mostly it does not — but there are a few real traps.

30 August 2026

Search for "m4a to text" and you will find a hundred sites that want you to upload the file. You almost never need them: the format is rarely the actual obstacle, and where it is, the fix is not a converter website.

Here is what each format you are likely to be holding actually means.

The short answer #

For transcription, a good tool converts whatever you give it into the same thing internally — 16 kHz mono audio — and works from that. The container is a detail. What matters is:

1. Is there an audio track in there at all?
2. Was it compressed so hard that the speech detail is gone?
3. Can the tool open that particular container?

Only the third question depends on the file extension, and only for a couple of awkward cases.

The formats #

M4A — What an iPhone Voice Memo is. AAC audio in an MP4 container. Compressed, but at the bitrates phones use, quality is good and transcription is unaffected. The single most common file people want transcribed, and it is a non-issue.

MP3 — The universal one. Same story: lossy, fine in practice. Very low-bitrate MP3s (64 kbps and below, common in old podcast archives and phone-system recordings) do measurably hurt accuracy, because the compression removes the high frequencies that distinguish similar consonants. Nothing can put that back.

WAV — Uncompressed. The best case for accuracy and the worst case for file size: two hours is several hundred megabytes, which is why upload-based services often choke on WAVs specifically. Local transcription does not care.

AAC — The codec inside most M4A files, sometimes appearing as a bare .aac. Same properties as M4A.

CAF — Apple's Core Audio Format. Turns up from macOS and iOS recording tools. Uncommon elsewhere, handled fine by anything built on Apple's audio frameworks.

OGG / Opus — The web and Android side of the world. WhatsApp voice notes are Opus. Excellent codec for speech at low bitrates, genuinely better than MP3 at the same size. Some Apple-native tools cannot open it without a decoding step, which is a tooling limitation rather than a quality one.

MOV and MP4 — Video. Transcription only needs the audio track, which is extracted first. Practically every screen recording, phone video and camera file is one of these.

The traps that are real #

A video file with no audio track #

Obvious once you know, invisible until then. Some screen recordings are captured without audio, and some cameras write a separate audio file. A transcription tool cannot do anything with a silent video, and a good one should say so rather than returning an empty transcript.

Legacy MOV files #

Not all .mov files are alike. Older QuickTime files can contain a single muxed MPEG stream with no separate audio track that modern frameworks can address. Apple's own tools will report zero audio tracks. If a specific old .mov fails everywhere while modern ones work, this is usually why. Re-exporting it from QuickTime Player is the practical fix.

Very low bitrate anything #

Below about 64 kbps for speech, you lose accuracy no matter what you transcribe with. If you have a choice at recording time, this is the setting worth caring about — far more than which app you use afterwards.

A file that is not what its extension says #

Renaming recording.aiff to recording.mp3 does not convert it, and tools that trust the extension will fail confusingly. If a file behaves strangely, check what it actually is (file recording.mp3 in Terminal will tell you).

Do you need to convert first? #

Usually no, and converting can cost you.

Every conversion between lossy formats — MP3 to M4A, say — decodes and re-encodes, losing a little more detail each time. If your goal is accurate transcription, feeding the original file to a tool that reads it directly is strictly better than converting it first.

The exception is a format your tool genuinely cannot open. Then convert once, from the original, to WAV rather than to another lossy format.

What we support #

Offline Transcription reads MP3, WAV, M4A, AAC, CAF and OGG for audio, and MOV and MP4 for video, on both Mac and iPhone. Everything is converted internally to the 16 kHz mono the model expects, so you drop in the original file and nothing else is required.

Unsupported or undecodable files are listed with a clear reason rather than failing silently — including the legacy .mov case above, which we keep as a test fixture precisely because it is the one that fails everywhere.

The short version #

The format is almost never the problem. Compression below about 64 kbps is a real problem, video files with no audio track are a real problem, and legacy MOV files are an occasional real problem. Everything else — M4A, MP3, WAV, AAC, CAF, OGG, MP4 — transcribes directly, and converting first usually makes things slightly worse rather than better.