Audio format

AMR converter

An .amr file stores speech compressed with the Adaptive Multi-Rate codec, engineered to squeeze a voice call into a few kilobits per second. Feature phones saved voice notes and MMS clips as .amr, and the format still surfaces in old phone backups and some Android recorder apps.

Private for supported formats — processed in your browser

Convert supported files

Runs on your device

Drop AMR files here

Batch files can each use a different output. Nothing uploads for local conversions.

Working inputs include camera RAW, browser-local audio/video, PDF, CBZ/CBR comics, office documents, ebooks, markup, 3D models, structured text, images, and archives.

Where AMR comes from

3GPP adopted AMR as the mandatory speech codec for GSM and UMTS networks in 1999, tuning it to survive radio interference by shifting among eight bitrates mid-call. AMR-WB, the 16 kHz wideband version standardized in 2001, later became the codec marketed as HD Voice.

How AMR works

  • Narrowband AMR samples at 8 kHz, mono only, across bitrates from 4.75 to 12.2 kbps
  • ACELP speech modeling keeps voices intelligible but smears music badly
  • Audio is packed in 20-millisecond frames behind a '#!AMR' magic header, per RFC 4867
  • The wideband sibling (.awb) doubles the sampled bandwidth to 16 kHz

When to use AMR

  • Retrieving voicemail exports and call recordings from old handsets
  • Converting voice memos to MP3 or WAV so transcription tools accept them
  • Archiving MMS voice clips from feature-phone backups

Strengths and limitations

Strengths

  • Remarkably small files, with a minute of speech around 90 KB at the top bitrate
  • Speech stays intelligible at bitrates where music codecs collapse
  • ffmpeg decodes it freely, so conversion out is painless

Limitations

  • Telephone-grade fidelity: nothing above roughly 3.4 kHz survives
  • Entirely unsuitable for music
  • Few desktop players open it without conversion, and it is mono only

Compatibility

Android generally still records and plays .amr, while iOS, desktop media players, and all browsers require a conversion step first. VLC and ffmpeg handle it fine, so a one-time transcode to MP3 or WAV is the practical fix.

None; the storage format is a bare magic number followed by speech frames, with no tagging structure at all.