Extract Audio from CAVS to AMR — Free Online Tool
Extract audio from CAVS (Chinese Audio Video Standard) video files and convert it to AMR format using the libopencore_amrnb codec — ideal for repurposing Chinese broadcast or streaming content into a speech-optimized, mobile-friendly audio format. AMR's ultra-low bitrates make it perfect for voice content derived from CAVS sources.
to
FFmpeg Command
Copy this command to run the same conversion locally with FFmpeg on your desktop. Download FFmpeg
Drop your CAVS 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
This tool strips the AAC audio track from your CAVS file and re-encodes it as AMR Narrowband audio using the libopencore_amrnb codec at 12,200 bps. Because CAVS uses AAC audio and AMR is a completely different codec architecture optimized for speech rather than general audio, a full re-encode is required — there is no remuxing shortcut here. The video stream is discarded entirely (via the -vn flag), and the AAC audio is decoded to raw PCM, then re-encoded into AMR-NB frames. AMR-NB operates at an 8kHz sample rate, so the audio is downsampled from the higher sample rate typical of CAVS broadcast content. This process prioritizes intelligibility of speech over fidelity, which is the core design philosophy of the AMR codec.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg multimedia processing tool, which handles decoding the CAVS container, separating its video and audio streams, and orchestrating the re-encode to AMR. |
-i input.cavs
|
Specifies the input file in CAVS format — a Chinese national standard video container that holds an H.264 video stream and an AAC audio track, both of which FFmpeg will parse before the output flags determine what to do with each stream. |
-vn
|
Disables video output entirely, discarding the H.264 video stream from the CAVS file so that only the audio is processed. Without this flag, FFmpeg would attempt to include a video stream in the output, which AMR does not support. |
-c:a libopencore_amrnb
|
Selects the libopencore_amrnb encoder to re-encode the decoded AAC audio from the CAVS file into AMR Narrowband format — a speech-optimized codec that operates at 8kHz and is natively supported by mobile telephony systems. |
-b:a 12200
|
Sets the AMR-NB encoding bitrate to 12,200 bps, which corresponds to the highest quality mode available in the AMR-NB standard (Mode 7 / MR122). This maximizes speech intelligibility when downsampling from the higher-quality AAC audio in the CAVS source. |
output.amr
|
Specifies the output filename with the .amr extension, which tells FFmpeg to write the result in the AMR file format containing the newly encoded libopencore_amrnb audio stream with no video. |
Common Use Cases
- Extracting voice commentary or narration from Chinese broadcast CAVS recordings to store as compact AMR files for playback on older mobile phones or feature phones that natively support AMR
- Converting CAVS-encoded news reports or documentary dialogue into AMR for integration with mobile telephony systems or IVR (Interactive Voice Response) platforms
- Archiving speech segments from CAVS broadcast content at minimal file sizes when audio fidelity is not critical — for example, transcription workflows or voice annotation archives
- Preparing voice audio from CAVS source material for use in VoIP applications or messaging apps that accept AMR attachments, common in Asian mobile ecosystems
- Extracting interview or spoken-word audio from CAVS files to share via systems with strict bandwidth or storage constraints, where AMR's low-bitrate efficiency is essential
- Converting CAVS-format lecture or presentation recordings into AMR for distribution on platforms or devices common in markets where AMR is the standard voice format
Frequently Asked Questions
AMR-NB (Narrowband) operates at a fixed 8kHz sample rate, which means frequencies above 4kHz — including much of the natural brightness and warmth of music or broadband speech — are simply cut off during conversion. Your CAVS file's AAC audio track was almost certainly encoded at 44.1kHz or 48kHz, so you will hear a clear reduction in audio bandwidth. AMR was designed exclusively for voice intelligibility in telephony, not high-fidelity playback, so this tradeoff is by design rather than a flaw in the conversion process.
Not particularly well. AMR-NB's codec algorithms are tuned specifically for human speech patterns, and they handle music, sound effects, or ambient audio from CAVS broadcast content poorly. Music will sound muffled and distorted because the 8kHz sample rate ceiling and the codec's speech-focused compression artifacts are ill-suited to tonal or harmonic content. If your CAVS file contains significant non-speech audio, you should consider an audio-extraction format like MP3 or AAC instead of AMR.
This tool uses AMR-NB (Narrowband), encoded with libopencore_amrnb, which operates at 8kHz sample rate and bitrates from 4,750 to 12,200 bps. AMR-WB (Wideband), encoded with libopencore_amrwb, operates at 16kHz and delivers significantly better speech quality. AMR-NB is chosen as the default because it has the broadest device compatibility, particularly with older mobile handsets and telephony systems. If your target platform supports AMR-WB, you can modify the FFmpeg command to use -c:a libopencore_amrwb for noticeably clearer voice audio from your CAVS source.
Replace the 12200 value in -b:a 12200 with any of the valid AMR-NB mode bitrates: 4750, 5150, 5900, 6700, 7400, 7950, or 10200 (all in bps). For example, -b:a 7950 gives a reasonable balance between file size and speech clarity. Note that 12200 bps (the default used here) is the highest quality AMR-NB mode and is recommended for CAVS broadcast source material where the original AAC audio may contain multiple speakers or background noise that benefits from the extra bitrate headroom.
Dramatically smaller. A CAVS file contains a full video stream (encoded with H.264/libx264) plus an AAC audio track, typically resulting in files of hundreds of megabytes for even short clips. The resulting AMR file contains only audio at 12,200 bps — roughly 90KB per minute of audio. A 30-minute CAVS broadcast recording that might be 1.5GB will produce an AMR file of only about 2.7MB, though the tradeoff is that the audio is downsampled to 8kHz narrowband quality.
Yes. On Linux or macOS, you can use a shell loop: for f in *.cavs; do ffmpeg -i "$f" -vn -c:a libopencore_amrnb -b:a 12200 "${f%.cavs}.amr"; done. On Windows Command Prompt, use: for %f in (*.cavs) do ffmpeg -i "%f" -vn -c:a libopencore_amrnb -b:a 12200 "%~nf.amr". This is especially useful if you have a library of CAVS broadcast recordings you need to convert to AMR in bulk, since the browser-based tool processes files one at a time.
Technical Notes
CAVS files store audio as AAC (Advanced Audio Codec), typically at 44.1kHz or 48kHz with bitrates ranging from 96k to 256k depending on the broadcast source. Converting to AMR-NB via libopencore_amrnb requires FFmpeg to first fully decode the AAC stream to PCM, then downsample the audio to 8kHz mono (AMR-NB is always mono), and finally encode using the AMR-NB codec at one of its fixed bitrate modes. The mandatory downmix from stereo to mono may cause phase cancellation artifacts if the CAVS source uses stereo separation for distinct audio elements. Metadata from the CAVS container — including title, language tags, or broadcast metadata — is not preserved in the AMR output, as the AMR format has very limited metadata support. There is no concept of chapters or multiple audio tracks in AMR. The libopencore_amrnb library used by FFmpeg is a clean-room implementation of the AMR-NB codec and is fully compatible with AMR files expected by mobile devices, though some proprietary systems may expect AMR files with specific header variants. If you need wideband quality, switch to libopencore_amrwb, but verify your target device or system supports the .amr wideband variant before doing so.