Audio files & formats

What is actually inside a WAV or AIFF file? This page explains the basics — sample rate, bit depth, channels, file formats — and the dual-mono problem.

Digital audio in one paragraph

A microphone signal is a continuously changing voltage. To record it, an audio interface measures that voltage many thousand times per second. Each measurement is a sample: a number saying how far the speaker cone should be pushed out or pulled in at that moment. A list of these numbers, stored as they are, is called PCM (pulse-code modulation) — the uncompressed audio that Pro Tools and Logic Pro record.

Sample rate

The sample rate is how many samples are taken per second, per channel, in hertz (Hz) or kilohertz (kHz). The highest frequency a recording can contain is half the sample rate (the Nyquist frequency) — 44.1 kHz captures up to 22.05 kHz, just above the limit of human hearing.

All files in a session should share the session's sample rate. A file at the wrong rate plays at the wrong speed and pitch, or has to be converted on import.

Bit depth

The bit depth is how precisely each sample is stored. More bits mean finer steps between the quietest and loudest possible values — more dynamic range before the noise floor.

16-bit

About 96 dB of dynamic range. The CD format; fine for finished masters, tight for recording.

24-bit

About 144 dB in theory — more than any microphone or converter delivers. The standard for recording.

32-bit float

Stores samples as floating-point numbers. Signals above 0 dBFS don't clip inside the file, so it gives enormous headroom. Common for bounces, intermediate files and some field recorders.

Channels: mono, stereo and more

Each channel is its own stream of samples. A mono file has one, stereo two (left and right), surround files six (5.1) or more. Most sources — a voice, a kick drum, a bass DI — are mono. Stereo makes sense when the left and right channels actually differ: a stereo microphone pair, a synth with stereo effects, a mix.

How big is an audio file?

Uncompressed audio size is simple arithmetic: sample rate × bytes per sample × channels × seconds.

A stereo file is exactly twice the size of the same audio in mono — which is why dual mono wastes space.

File formats (containers)

A file format is a container: it wraps the audio data together with a description of it (sample rate, bit depth, channels) and, optionally, metadata.

WAV (.wav)

The most widely used format for uncompressed audio, from Microsoft and IBM's RIFF family. Works everywhere.

Broadcast Wave / BWF (.wav, sometimes .bwf)

A WAV file with an extra chunk describing the recording: a description, who made it, when, and where it belongs on the timeline. Pro Tools records Broadcast Wave by default. Broadcast Wave explained →

RF64 and BW64

WAV variants for files larger than 4 GB. Regular WAV stores sizes in 32-bit numbers, which limits it to 4 GB (about 4 hours of 48 kHz / 24-bit stereo). RF64 (EBU) and BW64 (ITU) use 64-bit sizes. BW64 can also carry ADM metadata for object-based audio such as Dolby Atmos.

AIFF / AIFF-C (.aif, .aiff, .aifc)

Apple's long-standing uncompressed format, the Mac counterpart to WAV. AIFF-C can also hold compressed or floating-point audio.

CAF (.caf)

Apple's Core Audio Format. No 4 GB limit, flexible metadata, and can hold almost any kind of audio. Logic Pro can record to CAF.

MP3, AAC / M4A

Lossy formats: they discard sound the ear is unlikely to notice to make files much smaller. Great for listening and sharing, not for production — every decode and re-encode loses a bit more.

FLAC, ALAC

Lossless compression: smaller files that decode to exactly the original samples. Used for distribution and archiving rather than inside DAW sessions.

Inside a file: chunks

WAV, AIFF and CAF files are built from chunks: labelled blocks, each with a four-letter ID (a FourCC) and a length. A program reads the chunks it understands and skips the rest. A typical Broadcast Wave file from Pro Tools looks like this:

This structure is why metadata can get lost: a tool that converts a file by decoding the audio and writing a new file often keeps only fmt and data. The Media Manager instead copies every chunk it doesn't need to change. See Audio metadata.

Dual mono

A dual-mono file is a stereo file whose left and right channels are identical, sample for sample. It sounds exactly like a mono file played through both speakers — but takes twice the space.

Where it comes from

• A mono source (vocal, bass DI, guitar amp mic) recorded on a stereo track or through a stereo input.
• A mono track bounced or exported as stereo — some exports always write stereo.
• Sample libraries and loops delivered as stereo for convenience.
• Files passed between collaborators and tools that normalize everything to stereo.

Why it's worth fixing

• Disk space and bandwidth — twice the size for no extra information.
• Track and plug-in behavior — the file needs a stereo track; panning a stereo track balances two identical channels instead of placing a mono source, and stereo plug-in instances use more CPU.
• Clarity — you can't see from the file that it's really mono, so collaborators treat it as stereo.

Identical, or just similar?

A file is only dual mono when the channels are bit-identical. Files with a tiny difference — a dithered bounce, a slight stereo effect, a pan that isn't quite centered — are not, and converting them would throw that difference away. You can hear the difference with a null test: subtract one channel from the other, and a true dual-mono file leaves complete silence.

The Media Manager checks for bit-exact equality, so it never converts a file that has real stereo content.

Related: mono compatibility

Dual mono is not the same as mono compatibility. A real stereo mix is mono-compatible when it still sounds right summed to mono (on a phone speaker, a club PA, a radio). Listen with Mono sum (L+R) in the Media Manager preview: if parts become thin or disappear, the left and right channels are partly out of phase.