Convert AIF to AMR — Free Online Tool
Convert AIF audio files to AMR format using the libopencore_amrnb codec, optimized for speech encoding and mobile telephony. This tool transforms Apple's lossless uncompressed audio into a highly compressed, bandwidth-efficient format designed for voice communications — ideal for preparing recordings for mobile devices or VoIP systems.
to
FFmpeg Command
Copy this command to run the same conversion locally with FFmpeg on your desktop. Download FFmpeg
Drop your AIF 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
AIF stores audio as uncompressed PCM data (typically pcm_s16be — 16-bit big-endian signed integers), which means every sample is stored at full fidelity with no compression. During conversion, FFmpeg decodes this raw PCM audio and re-encodes it using the libopencore_amrnb codec, which implements the Adaptive Multi-Rate Narrowband algorithm. AMR-NB is highly specialized — it operates at a fixed sample rate of 8000 Hz and a single mono channel, so if your AIF file contains stereo audio or was recorded at a higher sample rate (44.1 kHz, 48 kHz, etc.), FFmpeg will automatically downmix to mono and resample to 8 kHz. The bitrate is locked to one of AMR's defined mode rates, defaulting here to 12.2 kbps — the highest AMR-NB mode, which offers the best speech intelligibility within the format's constraints. This is a lossy conversion; the uncompressed fidelity of the original AIF cannot be recovered from the resulting AMR file.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg multimedia processing engine. This is the core tool running in your browser via WebAssembly (FFmpeg.wasm) — the same binary you would run on your desktop command line. |
-i input.aif
|
Specifies the input file — in this case an AIF file containing uncompressed PCM audio (typically pcm_s16be big-endian 16-bit). FFmpeg reads the AIFF container header to determine sample rate, bit depth, and channel count before decoding. |
-c:a libopencore_amrnb
|
Selects the libopencore_amrnb encoder for the audio stream. This encodes the decoded PCM audio using the Adaptive Multi-Rate Narrowband algorithm, a codec purpose-built for compressed speech transmission over mobile networks. |
-b:a 12200
|
Sets the AMR-NB bitrate mode to 12.2 kbps, which corresponds to the MR122 mode — the highest quality setting available in AMR-NB. This value must be one of AMR's eight fixed mode rates; 12200 is recommended when converting from a high-quality AIF source to maximize speech intelligibility in the output. |
output.amr
|
Defines the output filename with the .amr extension, which tells FFmpeg to write the encoded audio into the AMR file format as specified by RFC 4867. This file is directly playable on Android devices, many feature phones, and telephony software. |
Common Use Cases
- Preparing voice memos or speech recordings originally captured on a Mac in AIF format for playback or transmission over mobile telephony networks that require AMR encoding
- Compressing large lossless AIF interview recordings into small AMR files for storage on embedded systems or legacy mobile handsets with very limited storage capacity
- Converting AIF-format voiceover or spoken-word content into AMR for integration into IVR (Interactive Voice Response) systems or telephony platforms that only accept AMR input
- Archiving spoken AIF recordings from older Mac audio applications into AMR to share via messaging platforms or email where file size is a constraint
- Preparing audio captured in professional AIF format for use in mobile app prototypes or telecom testing environments that simulate GSM or 3GPP voice streams
Frequently Asked Questions
Yes — this is a significant quality reduction, but it is intentional and format-appropriate. AIF stores audio as uncompressed PCM at full bit depth (typically 16-bit or higher) and sample rates up to 192 kHz or beyond, while AMR-NB encodes at a fixed 8000 Hz sample rate in mono at just 12.2 kbps maximum. All audio content above 4000 Hz is discarded, stereo information is collapsed to mono, and the codec applies perceptual speech compression tuned specifically for human voice. For speech recordings, intelligibility remains high; for music or wide-bandwidth content, the quality loss is severe and the format is simply not designed for that use case.
AMR-NB (Adaptive Multi-Rate Narrowband) is a codec defined by the 3GPP telecommunications standard specifically for voice telephony over GSM and mobile networks. The standard mandates 8000 Hz mono as the only supported configuration — this mirrors the bandwidth of a standard phone call. FFmpeg handles the downmix and resampling automatically during conversion, but it is a hard constraint of the codec itself, not a limitation of this tool.
AMR-NB supports eight fixed bitrate modes: 4.75, 5.15, 5.9, 6.7, 7.4, 7.95, 10.2, and 12.2 kbps. These correspond to the codec's internal 'modes' (MR475 through MR122). Lower modes produce smaller files but less intelligible speech; higher modes preserve more speech detail. The default here is 12.2 kbps (MR122), which is the highest quality AMR-NB mode and is recommended unless file size is severely constrained. For relatively clean speech recordings from AIF source files, 7.4 kbps or above typically produces acceptable results.
Replace the value after -b:a in the command with any of the eight supported AMR-NB modes: 4750, 5150, 5900, 6700, 7400, 7950, 10200, or 12200 (values in bits per second). For example, to use the 7.4 kbps mode, run: ffmpeg -i input.aif -c:a libopencore_amrnb -b:a 7400 output.amr. Note that AMR ignores arbitrary bitrate values — FFmpeg will snap your input to the nearest valid AMR mode, so only the eight listed values should be used.
No. AMR is a bare audio format with extremely limited metadata support — it has no standardized container for embedding tags like artist name, title, or creation date. Any metadata stored in the AIF file (such as Apple-specific chunks or ID3-style tags) will be discarded during conversion. If metadata preservation matters, consider an intermediate conversion to a format like M4A or MP3 before distributing, or document the metadata separately.
Yes — on Linux or macOS you can use a shell loop: for f in *.aif; do ffmpeg -i "$f" -c:a libopencore_amrnb -b:a 12200 "${f%.aif}.amr"; done. On Windows Command Prompt: for %f in (*.aif) do ffmpeg -i "%f" -c:a libopencore_amrnb -b:a 12200 "%~nf.amr". This is particularly useful when you have many AIF recordings to compress for mobile delivery, since processing them one at a time in the browser tool would be impractical.
Technical Notes
The libopencore_amrnb encoder used in this conversion is an open-source implementation of the 3GPP AMR-NB standard, and must be compiled into FFmpeg separately (it is not included in all FFmpeg builds due to patent licensing history). If you run the command locally and receive an 'encoder not found' error, you will need an FFmpeg build that includes libopencore-amrnb — on Ubuntu this is available via the libavcodec-extra package, and on macOS via Homebrew's ffmpeg formula with the --with-opencore-amr option or equivalent. The output .amr file uses the AMR storage format as defined in RFC 4867, which is a simple file wrapper with a magic number header (#!AMR\n). This file is natively supported by many Android devices and mobile messaging apps. One important limitation: because AMR-NB hard-codes its output to 8 kHz mono, even the highest-quality AIF source (e.g., a 32-bit float stereo file at 96 kHz recorded via pcm_f32be) will be reduced to the same telephony-grade output as a basic 16-bit source. The source format quality simply cannot pass through the AMR-NB codec ceiling. If your use case involves wideband speech at 16 kHz, consider the AMR-WB variant (libopencore_amrwb), which supports higher fidelity speech while remaining a compact, mobile-friendly format.