M4A vs WAV for voice recording

What the two formats actually are

WAV is a container for uncompressed audio, usually linear PCM. Every sample is stored as-is. Nothing is discarded, nothing is modelled, and the file size is pure arithmetic: sample rate × bit depth × channels × seconds. One minute of 44.1 kHz 16-bit mono is about 5 MB.

M4A is a container — the MPEG-4 one — that on iOS almost always holds AAC. AAC is lossy: it throws away detail a psychoacoustic model predicts you will not hear, and stores what remains far more compactly. The same minute lands somewhere around 0.5 MB depending on the encoder settings.

Worth naming a common confusion: M4A is the container, AAC is the codec. An M4A can also hold Apple Lossless, which is not lossy at all. If a spec sheet says only "M4A" it has not told you the important half.

Does lossy compression hurt speech?

Far less than people expect. AAC was designed against the way human hearing works, and speech is the easiest case for it: a narrow frequency range, one source, high redundancy. At the bitrates a phone recorder uses, the difference on a voice recording is not something you will hear on any playback you actually use.

What genuinely degrades a voice recording, in order: distance from the speaker, room reverberation, background noise, and clipping. All four dwarf the codec. Choosing WAV to protect quality while sitting three metres away in a noisy room is optimising the one variable that is not the problem.

When WAV is the right call

• You will process the audio through several stages — noise reduction, then EQ, then compression. Each lossy re-encode compounds artefacts; working in WAV avoids that.
• You are feeding audio to a tool or pipeline that only accepts PCM. Some analysis and forensic tools do.
• You need sample-accurate editing or precise measurement rather than listening.
• Archival policy requires an unmodified master, common in research and legal contexts.

The pattern: WAV is a working and archival format. M4A is a capture and delivery format. If you are recording a meeting, a lecture, an interview or a note to yourself, you are capturing and delivering.

Sample rate and channels

Sample rate

The rate sets the highest frequency that can be represented — half the rate, in theory. Speech intelligibility lives largely below 8 kHz, so even 16 kHz audio is intelligible; 22.05 kHz is comfortably clear; 44.1 kHz captures everything a voice produces with margin. Going to 48 kHz for a spoken voice changes essentially nothing you can hear, and mainly buys convenience if you later drop the audio into a video timeline that runs at 48 kHz.

Mono vs stereo

A phone recording a voice should be mono. Stereo doubles the file for a spatial image nobody wants from a spoken-word recording, and a single-microphone capture has no real stereo information anyway. Every serious voice recorder defaults to mono for this reason.

What Voice Studio records

Voice Studio records AAC in an .m4a file, mono, with three quality levels: standard at 22.05 kHz, high at 44.1 kHz, and ultra at 48 kHz. High is the default and is the right choice for essentially every spoken-word recording. There is no WAV option — it would multiply file sizes by roughly ten for a difference you would not hear on speech, so we did not add one.

The .m4a is a normal file. You can share it out of the app or reach it through the Files app, and it opens in anything — Music, QuickTime, Audacity, a video editor, Windows. If you ever do need PCM for a processing chain, converting M4A to WAV is a one-step job in any audio tool; converting back is the direction that loses something.

Common questions

Is M4A the same as AAC?

Not quite. M4A is the container; AAC is the codec normally inside it on iOS. An M4A can also contain Apple Lossless, which is not lossy.

Will I hear the difference between M4A and WAV on a voice recording?

On speech at the bitrates a phone recorder uses, no — not on any playback you actually use. Distance, room and background noise all matter far more.

Does Voice Studio record in WAV?

No. It records mono AAC in .m4a at 22.05, 44.1 or 48 kHz depending on the quality setting. You can convert the exported file to WAV in any audio tool if a workflow needs PCM.

Which sample rate should I pick for speech?

High, at 44.1 kHz, is the sensible default. Standard at 22.05 kHz is perfectly intelligible and smaller; ultra at 48 kHz mainly helps if the audio is heading into a 48 kHz video timeline.