MPG to AMR Converter

MPG to AMR — Is AMR Really the Format You Want?

AMR is a speech codec, not a music codec. It was standardised by ETSI for GSM and later adopted by 3GPP as the mandatory codec for 3G cellular systems, and everything about it is tuned for one thing: carrying a human voice down a narrow radio channel. Extracting the audio from a .mpg or .mpeg video into AMR is the right call when the content is talking and the destination is a phone, a voice-note workflow, or an embedded device that only reads .amr. If the content is music, a film soundtrack, or anything you plan to listen to properly, convert to MPG to MP3 instead — AMR will destroy it, by design.

AMR Compared With the Usual Audio Targets

The size column is the whole argument for AMR: an hour of speech is a handful of megabytes. The sample-rate column is the whole argument against it for anything else — an 8 kHz sample rate cannot represent anything above 4 kHz, which removes most of what makes music sound like music.

When AMR Is the Right Choice

• The audio is speech: an interview, a lecture, a dictation, a voicemail, a call recording pulled out of a video file.
• The destination is a device or system that specifically expects .amr — some handsets, voice recorders, and telephony platforms accept nothing else.
• Storage or bandwidth is genuinely tight and intelligibility is the only requirement.
• You are matching an existing archive of AMR files and want the new material to behave identically.

When to Pick Something Else

• Anything musical. Even a spoken-word recording with background music will come out muddy, because the music is mostly above AMR-NB's ceiling.
• Anything stereo. AMR-NB is mono; a stereo mix is downmixed on the way through and the separation is gone for good.
• Anything you might edit later — MPG to WAV gives you uncompressed audio with nothing thrown away.
• Anything headed for a normal media library, where MPG to MP3 plays everywhere and AMR does not.

How to Convert MPG to AMR

1. Upload Your MPG File: Drag and drop your .mpg or .mpeg video, or click "+ Add Files". The video track is discarded; only the audio is extracted and re-encoded.
2. Select an Audio Codec Explicitly: The Audio Codec group offers AMR Narrow Band and AMR Wide Band. Nothing is preselected on this page, so click one rather than leaving it blank — narrowband for speech at 8 kHz, wideband for noticeably clearer voice at 16 kHz.
3. Set Quality Preset, Audio Sample Rate or Trim: File Compression starts on Quality Preset at "Very High (Recommended)". Audio Sample Rate offers 8000 Hz and 16000 Hz, Audio Channel is mono, and Trim takes a start time and a duration.
4. Convert and Download: Click Convert. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours — no sign-up, no watermark, never shared.

AMR Narrow Band vs AMR Wide Band

Whichever you choose, the codec fixes the sample rate: picking AMR Narrow Band forces 8 kHz and picking AMR Wide Band forces 16 kHz, regardless of what the Audio Sample Rate dropdown is showing. That is why selecting the codec first matters.

Frequently Asked Questions

Why is nothing selected in the Audio Codec list when I open the page?

Because the configuration for this target names a codec value that does not correspond to either of the two entries shown, so neither AMR Narrow Band nor AMR Wide Band starts highlighted. It is a configuration gap, not something you have done wrong. Click a codec explicitly before converting — that also makes the sample-rate handling predictable, since each AMR variant pins its own rate.

Why does the Constant Bitrate dropdown have no options in it?

The Constant Bitrate mode renders on this page with an empty rate list, so there is nothing to choose there. Use Quality Preset instead, which is the default mode and works normally. AMR is a fixed-mode codec anyway — it does not accept an arbitrary bitrate the way MP3 does, only one of the eight defined narrowband modes or nine wideband modes.

My music came out sounding like a phone call. Can I fix it?

Not by re-converting to AMR. AMR-NB samples at 8 kHz, so by the sampling theorem nothing above 4 kHz can survive — cymbals, string harmonics, and most of the brightness of a mix are gone before the encoder even starts, and the encoder itself models a human vocal tract rather than general audio. Go back to the original .mpg and use MPG to MP3 or MPG to WAV instead.

How big will the AMR file be?

Small enough that size stops being a consideration. In our testing, six seconds of speech-rate audio produced a 9,638-byte AMR-NB file at 12.4 kbit/s, which works out to roughly 96 KB per minute, or under 6 MB for an hour. Wideband roughly doubles that and is still tiny by any modern standard.

Will the AMR be stereo if my video's audio is stereo?

No. Both AMR variants are mono-only, and the Audio Channel control on this page offers mono for exactly that reason. A stereo source is downmixed to a single channel during conversion. If channel separation matters — a two-microphone interview panned left and right, for instance — AMR is the wrong container and you should use MPG to WAV or MP3.

Can I extract just one section of a long video?

Yes, without touching the video first. The Trim group on this page takes a start time and a duration in either seconds or HH:MM:SS.sss, and only that region is decoded and encoded. If you would rather cut the video itself before extracting audio, Video Cutter accepts .mpg and .mpeg directly.

What will open a .amr file once I have it?

VLC and most libavformat-based players read both AMR variants directly, as do many mobile handsets and voice-recorder apps. Support in general-purpose desktop media players and audio editors is patchy, and you should assume a recipient cannot open one without checking first. That gap is the main practical cost of choosing AMR, and it is why AMR is worth choosing only when something downstream specifically requires it.

Does converting to AMR keep any of the video's metadata?

No. The output is a bare AMR stream with a short format signature at the front and nothing else — no title, no timestamps, no chapter information, no artwork. It is one of the leanest audio containers in general use, which is part of the point. If you need metadata to survive, extract to MP3, which carries ID3 tags.