Extract Audio from MPG to AMR — Free Online Tool
Extract audio from MPG video files and convert it to AMR format, re-encoding the MP2 or MPEG audio stream using the libopencore_amrnb codec optimized for speech. AMR's narrow-band compression makes it ideal for archiving voice content from broadcast or VCD/DVD-sourced MPEG video at very low file sizes.
to
FFmpeg Command
Copy this command to run the same conversion locally with FFmpeg on your desktop. Download FFmpeg
Drop your MPG 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
MPG files typically carry MP2 (MPEG Layer 2) audio, the standard for VCD and broadcast MPEG video. Since AMR (Adaptive Multi-Rate) uses a completely different codec — libopencore_amrnb — a full audio transcode is required. FFmpeg strips the MPEG video stream entirely using the -vn flag, then decodes the MP2 audio and re-encodes it into AMR narrow-band at 12,200 bps. AMR operates at a fixed 8 kHz sample rate optimized for speech frequencies, so any music or wideband audio in the source MPG will be downsampled and may sound noticeably thin. This is a lossy-to-lossy conversion, meaning quality from the original MPG source is not recovered — only the speech-range content will translate well.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg tool, which handles reading the MPG container, decoding the MPEG audio stream, and encoding the AMR output — all three steps are needed since MP2 and AMR are incompatible codecs. |
-i input.mpg
|
Specifies the input MPG file. FFmpeg will parse the MPEG-1 or MPEG-2 container and identify both the video and audio streams; subsequent flags determine what gets processed and what gets discarded. |
-vn
|
Disables video output entirely, stripping the MPEG-1/2 video stream from the MPG file so only the audio track is processed and written to the AMR output. |
-c:a libopencore_amrnb
|
Selects the libopencore_amrnb encoder to transcode the MPG's audio (typically MP2) into AMR Narrow-Band format, the speech-optimized codec used in mobile telephony and voice applications. |
-b:a 12200
|
Sets the AMR encoding bitrate to 12,200 bps, the highest available AMR narrow-band mode, which yields the best speech quality the format can produce from the downsampled MPG audio. |
output.amr
|
Defines the output filename with the .amr extension, which FFmpeg uses to confirm the AMR container format and ensure compatibility with mobile devices and telephony systems that recognize this file type. |
Common Use Cases
- Extracting spoken dialogue or narration from digitized VHS or broadcast recordings stored as MPG files for use in mobile voice applications
- Converting voice-over tracks from MPEG-1 VCD video into AMR for playback on older mobile phones or embedded telephony systems
- Archiving speech segments from legacy DVD or VCD content in AMR format to minimize storage footprint on constrained devices
- Preparing interview or speech audio from broadcast MPG recordings for integration into IVR (interactive voice response) systems that require AMR input
- Extracting narration audio from educational or corporate VCD-format videos to create compact AMR voice files for distribution over low-bandwidth networks
- Isolating spoken commentary from MPEG-2 broadcast captures for transcription workflows that accept AMR audio input
Frequently Asked Questions
No — AMR narrow-band (libopencore_amrnb) is specifically engineered for human speech frequencies, sampling audio at only 8 kHz. Music, ambient sound, or any wideband content in your MPG's MP2 audio track will be aggressively downsampled and will sound muffled and degraded. AMR is the right choice only when the MPG source contains primarily spoken voice content.
libopencore_amrnb encodes AMR Narrow-Band, operating at 8 kHz sample rate with bitrates from 4,750 to 12,200 bps — this is the default and most compatible option for mobile telephony. libopencore_amrwb encodes AMR Wide-Band at 16 kHz, offering better speech quality at the cost of slightly larger files and narrower device compatibility. If your MPG contains higher-quality speech and your target device supports AMR-WB, switching to amrwb will yield noticeably cleaner audio.
Replace the value after -b:a with any of the valid AMR bitrate modes: 4750, 5150, 5900, 6700, 7400, 7950, 10200, or 12200 (in bps). For example, use '-b:a 7950' for a balance between file size and speech intelligibility, or '-b:a 4750' for the absolute smallest file. The default of 12200 bps gives the best quality AMR narrow-band audio from your MPG source.
Two major factors account for this. First, the entire MPEG video stream is discarded — only the audio is kept. Second, AMR narrow-band encodes at a maximum of 12,200 bits per second, which is dramatically lower than the 128–320 kbps typical of MP2 audio in MPG files. Even a lengthy MPG will produce a very small AMR file, which is by design for mobile telephony use cases.
This FFmpeg command works regardless of which audio codec is in the MPG container — whether it's MP2 (the most common for VCD/broadcast), MP3, or AAC. FFmpeg automatically detects and decodes whatever audio stream is present, then re-encodes it to AMR narrow-band. The -vn flag ensures only the audio is processed, so you don't need to know your MPG's specific audio codec in advance.
Yes. On Linux or macOS, you can run a shell loop: 'for f in *.mpg; do ffmpeg -i "$f" -vn -c:a libopencore_amrnb -b:a 12200 "${f%.mpg}.amr"; done'. On Windows Command Prompt, use: 'for %f in (*.mpg) do ffmpeg -i "%f" -vn -c:a libopencore_amrnb -b:a 12200 "%~nf.amr"'. This is especially useful for processing large batches of legacy VCD or broadcast MPG archives.
Technical Notes
AMR narrow-band imposes a hard constraint of 8 kHz audio sampling, meaning FFmpeg will automatically downsample the MPG's MP2 audio — typically sampled at 44.1 kHz or 48 kHz — during this transcode. This is a significant lossy step beyond the initial MP2 lossy encoding, so the AMR output represents two generations of lossy compression. Metadata from the MPG container (if any exists) is not carried into AMR, as the AMR format has no standard container-level metadata fields for title, artist, or track information. AMR files are mono by default with libopencore_amrnb; if your MPG has stereo audio, FFmpeg will mix it down to mono, which is appropriate for speech but will collapse stereo field positioning. The .amr file extension is compatible with most mobile platforms, and files encoded with libopencore_amrnb at 12200 bps can be played back on virtually all devices that support AMR, including legacy feature phones. Note that FFmpeg must be compiled with libopencore-amrnb support, which is not always included in default system builds due to patent considerations — the browser-based tool handles this transparently.