AMR to AIFF Converter

Convert AMR to AIFF: What This Tutorial Covers

This tutorial is for anyone who needs an .amr voice recording — an old Android voice memo, a saved voicemail, an MMS audio clip, or a call recording — inside a Mac audio app or a DAW that expects uncompressed AIFF. It covers the two settings that actually matter on this page, sets honest expectations about what the conversion can and cannot do, and walks through the errors people hit with phone-recorded speech.

How to Convert AMR to AIFF

1. Upload Your AMR File: Drag and drop your .amr file onto the page or click "Add Files" to browse. Voice-recorder memos, exported voicemails, MMS audio attachments, and call recordings all work, and you can queue several files to run with the same settings.
2. Set Audio Channel and Audio Sample Rate: Both dropdowns start on ORIGINAL, which mirrors the source — the right answer for nearly every AMR file, because AMR is single-channel speech at 8 kHz (16 kHz for the wideband variant). Change the sample rate only when you need it to match a project rate.
3. Trim the Recording (Optional): Open the Trim panel and set a start time and a duration to keep only the part you need — handy for cutting an automated voicemail greeting off the front of a message.
4. Convert and Download: Click Convert and save the .aiff. The output is written as PCM signed 16-bit big-endian, the macOS-native AIFF flavour. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours — no sign-up, no watermark.

AMR and AIFF at a Glance

Walk-through: Channel, Sample Rate, and Why ORIGINAL Usually Wins

The AIFF encoder writes PCM signed 16-bit big-endian, the standard AIFF layout that Mac audio software reads without a plug-in. You do not need to think about the codec; the two dropdowns worth understanding are Audio Channel and Audio Sample Rate.

• Leave Audio Channel on ORIGINAL. Real-world AMR recordings are single-channel. Forcing STEREO duplicates that one channel into both sides: the file doubles in size and gains nothing, because both sides carry identical samples. A mono AIFF already plays through both speakers.
• Leave Audio Sample Rate on ORIGINAL too. An 8 kHz narrowband source cannot hold anything above 4 kHz, and its speech content is filtered to roughly 200–3,400 Hz before encoding. Resampling to 44,100 Hz multiplies the byte count by about 5.5× and invents nothing — interpolation cannot reconstruct frequencies the phone never captured.
• Upsample only for session alignment. If your DAW project runs at 44,100 or 48,000 Hz and you would rather not let the host resample on import, pick that rate here so the file drops straight into the timeline. This is a workflow convenience, not a fidelity gain.
• Downmixing to MONO is worth it in one case: if a file was already padded to two identical channels somewhere upstream, choosing MONO halves the AIFF without touching what you hear.

How Big Will the AIFF Be?

Uncompressed PCM has no guesswork in it — the size is sample rate × 2 bytes per sample × channels. That makes the jump from a speech codec to AIFF completely predictable:

AMR figures are the codec payload; the real file adds small per-frame headers, so treat them as a floor rather than an exact size.

Common Errors and How to Fix Them

• "The AIFF still sounds like a phone call." Expected, not a fault. Narrowband AMR discarded everything outside the telephone band at record time. AIFF stores that audio exactly; it cannot add back what was never there.
• "My AIFF is enormous compared with the AMR." Also expected. AMR stores a few speech-model parameters per 20 ms frame; AIFF stores every raw sample. The size table above shows the ratio — roughly 10–25× at the native rate, far more if you upsample.
• "My DAW opens the AIFF but refused the original AMR." Most desktop editors ship without an AMR decoder, which is the whole reason this conversion exists. The AIFF is plain PCM that every Mac audio app reads.
• "The file I have is .awb, not .amr." That is AMR wideband — 16 kHz, a fuller 50–7,000 Hz band. It decodes the same way; keep Audio Sample Rate on ORIGINAL to preserve the 16 kHz rate.
• "The output is silent or cuts off early." The source is usually truncated or partially transferred off the handset. Play the original first to confirm it has audio, then re-export a clean copy from the device and upload again.

When This Doesn't Work

Some voicemail downloads from carrier portals are DRM-wrapped or served as a stream rather than a plain file; those will not decode, and the fix is to obtain an unprotected export. Files that carry the .amr extension but a different payload inside also fail — if that happens, try AMR to WAV, which writes plain little-endian PCM and is worth a second attempt. If you want a small file to send rather than an uncompressed editing master, AMR to MP3 is the better target, and an oversized AIFF can be shrunk afterwards with the Audio Compressor. To go back the other way for legacy phone software, use AIFF to AMR.

Frequently Asked Questions

Will converting AMR to AIFF make my voice recording sound better?

No, and any tool that claims otherwise is overselling. AMR is a lossy speech codec: it models the vocal tract rather than storing the waveform, and it filters the input to roughly the 200–3,400 Hz telephone band before encoding at 4.75–12.2 kbit/s. Decoding that into uncompressed AIFF gives you a far larger, edit-friendly file, but the missing detail is gone for good. What you gain is an editable, universally readable master that will not lose any more quality as you work on it.

Should I upsample the AIFF to 44.1 kHz, or leave it at 8 kHz?

For listening or archiving, leave Audio Sample Rate on ORIGINAL. A narrowband source has no content that 8 kHz cannot hold, so a 44.1 kHz version is about 5.5× bigger and sounds identical. Upsample only when your session runs at 44,100 or 48,000 Hz and you want the file to import without the host resampling it — that is an alignment decision, not a quality one.

Why is my AIFF so much larger than the AMR?

Because the two formats store completely different things. A minute of narrowband AMR is roughly 36–92 KB depending on the mode; the same minute as 8 kHz mono 16-bit AIFF is about 0.96 MB, since uncompressed PCM spends 16,000 bytes on every second regardless of content. That is the price of an uncompressed format, not evidence that something went wrong.

What codec and byte order does the output AIFF use?

Our AIFF output is PCM signed 16-bit big-endian — the standard AIFF layout, and the reason file sizes here are exactly predictable rather than estimated: sample rate × 2 bytes × channels. Big-endian ordering is AIFF's defining trait, inherited from the Motorola 68000 processors in the first Macs, and it is what separates AIFF from WAV, which stores the same PCM samples little-endian inside a RIFF container.

Should I keep the recording mono or convert it to stereo?

Keep it mono. The source has one channel of speech; selecting STEREO writes two identical channels, doubling the file for no audible difference. Mono AIFF plays correctly through stereo speakers and headphones on its own. The only reason to choose stereo is if you intend to build a stereo mix around the voice and want the extra channel ready in the session.

AIFF or WAV — which should I pick for this recording?

They hold the same uncompressed PCM, so neither is higher quality. Choose AIFF for Mac-centred work: Logic Pro, GarageBand, and Final Cut treat it as a native format, and it carries ID3-style metadata cleanly. Choose AMR to WAV if the destination is Windows software, a transcription service, or a speech-to-text API, since WAV is the format those tools name first in their documentation.

Does this handle AMR-WB (.awb) wideband files too?

Yes. Wideband AMR samples at 16 kHz with a roughly 50–7,000 Hz band and nine bit rates from 6.6 to 23.85 kbit/s, and it decodes to AIFF the same way narrowband does. It will sound noticeably fuller than a narrowband recording because the phone captured more of the voice — leave Audio Sample Rate on ORIGINAL so the 16 kHz rate carries through untouched.

How are my files handled, and how long do you keep them?

Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours — no sign-up, no watermark, never shared or made public. The practical constraint on a long recording is upload size and time rather than anything on the conversion side, and because AIFF is uncompressed you can predict the output size before you start with the table above.