AMR is a speech codec tuned for voice recordings and phone calls, which is why music sounds poor in it.
Common Uses
Voice recordings
Mobile audio
Telephony
AMR Optimization FAQ
What is an AMR file?
+
AMR (Adaptive Multi-Rate) is an audio format optimized for speech compression.
How much quality has already been lost in an AMR file?
+
Whatever the original encoder decided to discard, and it cannot be recovered. AMR compresses by modelling what the ear will not notice — quiet sounds sitting under loud ones, frequencies at the edges of hearing — and permanently removing it. At a healthy bitrate that is inaudible to almost everyone. What it is not is reversible, so converting to a lossless format preserves the current state rather than restoring the original. AMR is a narrowband speech codec running between 4.75 and 12.2 kbps, modelling the human vocal tract rather than the full audio spectrum.
Why does my AMR recording sound thin compared to music?
+
Because it was never meant to carry music. AMR is a narrowband speech codec built for mobile networks, tuned to keep words intelligible at a bitrate measured in single-digit kilobits. Everything outside the range of the human voice is discarded at capture time. Converting to a full-range format keeps the recording usable everywhere, but it cannot add back frequencies that were never sampled. AMR destroys music: its 8 kHz bandwidth removes everything above telephone frequencies, and no amount of re-encoding brings it back.
Where do AMR files usually come from?
+
Voicemail systems, push-to-talk radios, older handsets and telephony recording — anywhere audio had to travel over a narrow, expensive channel. That heritage explains both why the files are tiny and why they sound the way they do. Converting makes them play in ordinary software; it does not, and cannot, add fidelity the capture never had. AMR is Adaptive Multi-Rate, the speech codec 3GPP standardised for GSM voice calls and mobile voice recorders.
Why does my AMR file show up with no track name or artwork?
+
Because the format has essentially nowhere to store them. AMR was specified around the audio itself, with no standard place for a title, an album or a cover image, so players fall back to displaying the filename. Converting to a tagged format is what makes a library browsable — the tags are written during conversion rather than recovered, so expect to fill in anything the source never carried.
Should I still be saving new files as AMR?
+
Only if something downstream demands it. The format still opens, and there is no urgency about converting an archive that is not causing trouble, but it is no longer where the tooling or the efficiency is going. For anything new, a current format costs nothing extra and avoids the next round of this question a few years from now.
Why will an AMR file not play in my browser?
+
Browsers decode a short, fixed list of audio formats and AMR is outside it, so the file downloads instead of playing or the player simply reports an error. Desktop software with a wider codec set opens it without trouble. If the audio has to play on a web page, converting it is the fix — no player configuration will add a decoder the browser does not ship.
What bitrate should I convert my AMR files to?
+
Match the destination, not the source. For music, 256 to 320 kbps is transparent to nearly every listener; for speech and podcasts, 96 to 128 kbps is plenty and halves the file again. Encoding above the source bitrate is wasted space — it faithfully preserves audio that has already been through a lossy pass and recovers nothing the first encoder discarded.
Does using AMR require a patent licence?
+
Not for you. The codecs involved are covered by patent pools, and the licensing is settled by the people who ship decoders — the phone, the browser, the operating system — long before a file reaches you. It is why the format is everywhere in hardware and why royalty-free alternatives exist at all, and it changes nothing about converting a file you already have.
Does converting change the sample rate of my AMR audio?
+
Not unless you ask it to. A conversion keeps the source rate — 48 kHz stays 48 kHz, 44.1 kHz stays 44.1 kHz — because resampling is a small quality cost for no benefit when nothing requires it. The exceptions are real: audio CDs need 44.1 kHz, some older hardware players accept nothing else, and voice pipelines often want 16 kHz. Set the rate explicitly in those cases and leave it alone otherwise.
Can I just rename the extension instead of converting AMR?
+
Almost never, and it is worth understanding why: the extension is a label, while the bytes inside are structured for one specific format. Renaming produces a file whose name promises something its contents do not deliver, and whatever opens it either refuses the file or misreads it. The one narrow exception is two formats that are genuinely the same container under different names — and even then a proper rewrap is safer than a rename.
Can I convert a whole folder of AMR files in one go?
+
With an account, yes — drop them in together and they are queued as one job rather than handled one at a time, with the same settings applied across the batch and the results returned together. Free conversions run a single file at a time (merge tools take two); a paid plan is what unlocks batch processing and the larger per-file limit, which is what makes converting a whole archive practical.
