Convert AMR to WAV — Free Online Tool
Convert AMR audio files to WAV format by decoding the speech-optimized AMR bitstream into uncompressed PCM audio (pcm_s16le), making your voice recordings and mobile telephony audio universally compatible with any audio editor, media player, or broadcast workflow.
to
FFmpeg Command
Copy this command to run the same conversion locally with FFmpeg on your desktop. Download FFmpeg
Drop your AMR file here
or click to browse
Free — no uploads, no signups. Your files never leave your browser.
Settings
Note: Browser-based encoding uses approximate quality targets. For precise CRF compression, copy the FFmpeg command above and run it on your desktop.
Estimated output:
Conversion Complete!
DownloadHow It Works
AMR (Adaptive Multi-Rate) files store audio using highly compressed, speech-tuned codecs — either narrowband (AMR-NB, typically 8 kHz sample rate) or wideband (AMR-WB, typically 16 kHz) — that discard non-speech audio information to achieve very low bitrates as small as 4.75 kbps. During this conversion, FFmpeg decodes the AMR bitstream using the libopencore_amrnb or libopencore_amrwb decoder and re-encodes the audio as uncompressed 16-bit signed little-endian PCM (pcm_s16le) inside a WAV container. Because AMR was designed exclusively for speech at narrow bandwidths, the resulting WAV file will be uncompressed but will reflect the limited frequency range of the original AMR recording — typically capturing only up to 4 kHz for AMR-NB or 8 kHz for AMR-WB, regardless of the WAV format's theoretical capacity. The file size will increase dramatically, since WAV stores raw PCM samples rather than compressed data.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg command-line tool, which handles all decoding, processing, and encoding. In the browser-based version of this tool, FFmpeg runs locally via WebAssembly (FFmpeg.wasm) — no data leaves your device. |
-i input.amr
|
Specifies the input file — an AMR audio file, which may contain either AMR-NB (narrowband, 8 kHz) or AMR-WB (wideband, 16 kHz) speech-coded audio. FFmpeg reads the file header to automatically detect which AMR variant is present. |
-c:a pcm_s16le
|
Sets the output audio codec to 16-bit signed little-endian PCM, which is the standard uncompressed audio encoding inside WAV files. This decodes the lossy AMR speech bitstream into raw PCM samples, producing a file that any audio software can read without needing AMR codec support. |
output.wav
|
Specifies the output filename and container format. The .wav extension tells FFmpeg to write a RIFF WAV container, which wraps the uncompressed PCM audio data along with a standard WAV header describing sample rate, bit depth, and channel count inherited from the decoded AMR stream. |
Common Use Cases
- Transcribing voice memos or call recordings saved from Android or feature phones, which often export audio in AMR format, into WAV for use with transcription software like Whisper or Audacity
- Importing mobile voice recordings into digital audio workstations (DAWs) such as Adobe Audition, Logic Pro, or Reaper, which may not natively support AMR containers
- Preparing archived telephony recordings or voicemail audio for legal or compliance review, where WAV is required as an uncompressed archival format
- Converting AMR voice notes from messaging apps or older Nokia/Samsung devices into WAV so they can be played on desktop computers without specialized codecs installed
- Processing AMR speech recordings in Python or MATLAB audio pipelines that expect standard PCM WAV input for speech analysis or machine learning datasets
- Normalizing, trimming, or noise-reducing a mobile voice recording in a WAV-based audio editor before converting to a distribution format like MP3 or FLAC
Frequently Asked Questions
No — converting AMR to WAV does not restore any audio quality that was lost during the original AMR encoding. AMR is a lossy codec, and when the file was first saved as AMR, frequencies above roughly 4 kHz (AMR-NB) or 8 kHz (AMR-WB) were permanently discarded. The resulting WAV file is uncompressed and will be much larger, but it faithfully represents what the AMR file actually contained — it does not recover lost detail. Think of it as lossless preservation of a lossy original.
This is an inherent characteristic of the AMR format, not a conversion artifact. AMR-NB (the most common variant on mobile phones) was designed for telephony and only encodes frequencies up to approximately 4 kHz — the same bandwidth as a traditional phone call. Even though WAV can represent full 20 kHz audio, the source material simply does not contain those higher frequencies. If your original was AMR-WB, you will have somewhat better quality with up to ~8 kHz bandwidth, but it will still sound narrower than a typical studio or even consumer microphone recording.
You can usually tell by the file extension: .amr files are typically AMR-NB (narrowband), while .awb files are typically AMR-WB (wideband). FFmpeg auto-detects the codec from the file header and selects the appropriate decoder (libopencore_amrnb or libopencore_amrwb) automatically, so the same ffmpeg command works for both. The distinction matters for output quality — AMR-WB will produce a WAV with a 16 kHz sample rate and noticeably better speech clarity than the 8 kHz output from AMR-NB.
Expect the WAV file to be dramatically larger — often 10x to 40x the size of the AMR original. For example, a 1-minute AMR-NB file encoded at 12.2 kbps is roughly 90 KB, while the equivalent uncompressed pcm_s16le WAV at 8 kHz mono would be approximately 960 KB. This size increase reflects the switch from compressed speech coding to raw PCM samples with no compression whatsoever.
You can change the output codec by replacing -c:a pcm_s16le with another WAV-compatible codec such as pcm_s24le or pcm_f32le for higher bit depth, but this will not improve audio fidelity beyond what the AMR source contains. You can also add -ar 44100 to upsample to a standard 44.1 kHz sample rate (e.g., ffmpeg -i input.amr -c:a pcm_s16le -ar 44100 output.wav), which can improve compatibility with some software that expects standard sample rates, though it again adds no new audio information. For most speech transcription or editing purposes, the default command is sufficient.
The single-file command shown is a direct reference for one file at a time. To batch convert in a Unix/macOS shell, you can use: for f in *.amr; do ffmpeg -i "$f" -c:a pcm_s16le "${f%.amr}.wav"; done. On Windows Command Prompt, use: for %f in (*.amr) do ffmpeg -i "%f" -c:a pcm_s16le "%~nf.wav". This is particularly useful when exporting many voice memos or call recordings at once from a mobile backup.
Technical Notes
AMR audio presents a few specific considerations when converting to WAV. The two AMR variants — narrowband (AMR-NB, 8 kHz, codecs up to 12.2 kbps) and wideband (AMR-WB, 16 kHz, codecs up to 23.85 kbps) — produce WAV output at their native sample rates, so downstream tools should not assume 44.1 kHz or 48 kHz output without explicit resampling. The default pcm_s16le codec (16-bit signed little-endian PCM) is the universally compatible WAV encoding and is appropriate for virtually all use cases including transcription, editing, and archival. AMR files typically carry minimal metadata — usually just duration — so do not expect ID3-style tags like artist or title to transfer into the WAV output. Because AMR is a mono-only format by design (it was built for single-channel telephony), the output WAV will always be mono; stereo WAV output is not possible without artificially duplicating the channel. The libopencore_amrnb and libopencore_amrwb decoders used by FFmpeg are derived from the 3GPP reference implementation and are considered accurate and stable for all valid AMR files, but corrupted or truncated AMR files (common in incomplete mobile downloads) may produce silence or errors at the point of corruption.