MTS to AMR Converter

Convert MTS to AMR — and When You Should Not

.mts is AVCHD camcorder footage: H.264 video with a 48 kHz Dolby Digital soundtrack, from the format Sony and Panasonic introduced in 2006. AMR is the opposite kind of thing entirely — Adaptive Multi-Rate, the speech codec built for GSM mobile calls, standardised for file storage in RFC 4867. Narrowband AMR samples at 8 kHz in a single mono channel and tops out at 12.2 kbps. That combination is excellent for a spoken-word recording that has to be tiny, and genuinely bad for anything else: music, ambience, applause and room tone all fall apart at those rates. This page explains how to do the conversion properly, and is honest about when you should pick something else.

How to Convert MTS to AMR

1. Upload Your MTS File: Drag your .mts clips onto the page or click "Add Files". AVCHD cameras split long recordings across numbered files on the card, so convert each in turn.
2. Pick a Codec in "Audio Codec" — Nothing Is Preselected: Choose "AMR Narrow Band" or "AMR Wide Band" before anything else. Until you do, the "Constant Bitrate" dropdown under "File Compression" renders with no options in it, because the rate ladder depends on which of the two you are using.
3. Set the Rate, and Check the Trim (Optional): With a codec chosen, "Quality Preset" and "Constant Bitrate" both become usable. "Audio Channel" is Mono only and "Audio Sample Rate" is locked to the one rate your codec allows — that is the codec's design, not a limitation of this page. Use "Trim" with a start time and duration to take only the section you need.
4. Convert and Download: Click Convert and download the .amr. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours. No sign-up, no watermark.

Narrow Band or Wide Band

This is the only decision that substantially changes what your file sounds like, and it is worth more attention than any rate setting.

If you have a free choice, pick Wide Band. Doubling the sample rate moves the ceiling from 4 kHz to 8 kHz, which is the difference between muffled and merely compressed on consonants and sibilance, and it still costs under 11 MB per hour at its top mode. Narrow Band exists for compatibility: some older phones, voice recorders and telephony systems only read the #!AMR variant, and if you are producing a file for one of those, that is the reason to choose it.

Why the Options Panel Looks Half-Empty at First

Worth explaining, because it looks like a fault. When the page loads, the "Audio Codec" group shows AMR Narrow Band and AMR Wide Band with neither of them selected, and the "Constant Bitrate" dropdown under "File Compression" is empty. That is a configuration quirk on this target: the default that is supposed to preselect a codec carries a value that does not correspond to either of the two visible entries, so nothing gets highlighted, and the rate list — which is filtered by codec — has nothing to show.

The practical consequence is simple: choose Narrow Band or Wide Band explicitly, first. Do that and everything downstream populates correctly — the sample rate pins itself to 8000 or 16000 Hz, the channel list shows Mono, and the bitrate ladder fills in with that codec's modes. If you convert without touching it, the encode still runs, but it falls back to the container's own default rather than anything you chose, and the file you get back may not be the one you intended.

For a predictable result the "Quality Preset" control is the steadier of the two rate settings here, since it is scaled against whichever codec you selected. If a specific bitrate is contractually or technically required, check the output file rather than assuming the label on the rung is what was delivered.

What AMR Costs You, in Numbers

The middle rows are the entire argument for AMR: a two-hour interview lands under 25 MB in wideband. The right-hand column is the entire argument against it for anything but speech. There is no rate on this page at which music from a camcorder microphone survives intact — mono at 8 or 16 kHz simply cannot carry it. If small-and-listenable is what you actually want, MTS to Opus is the modern answer: it handles speech at comparable bitrates and, unlike AMR, does not fall apart on everything else. If universal compatibility is what you want, MTS to MP3 remains the safest file to hand to anyone.

When This Doesn't Work

If your camcorder recording contains anything you would describe as music — a school concert, a wedding band, a busker in the background of a holiday clip — AMR is the wrong destination and no setting on this page will fix it. The codec models the human vocal tract, and material outside that model is reconstructed badly by design. Use MP3, AAC or Opus instead.

AMR is also mono-only in this pipeline, so a stereo field is summed away permanently; if a recording depends on stereo separation to be intelligible — two people on either side of a camera, for instance — convert to something that keeps both channels. And if you are producing a file for a specific device, confirm which variant it accepts before you convert: the narrowband and wideband storage formats carry different magic headers (#!AMR and #!AMR-WB), and a player that only understands one will refuse the other rather than degrade gracefully.

Frequently Asked Questions

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

Because the default configured for this target carries a codec value that does not match either of the two visible entries, so the interface has nothing to highlight. It is a known quirk rather than something you are doing wrong. Pick "AMR Narrow Band" or "AMR Wide Band" yourself as the first thing you do — that is also what makes the "Constant Bitrate" dropdown populate, since the rate ladder is filtered by codec and has nothing to display until one is chosen.

Why is the Constant Bitrate dropdown empty?

Same cause. AMR has two entirely separate rate ladders — eight modes from 4.75 to 12.2 kbps for narrowband, nine from 6.6 to 23.85 kbps for wideband — and the dropdown cannot show either of them until it knows which codec you are encoding with. Select a codec in the "Audio Codec" group and the list fills in immediately with that codec's modes. If you would rather not think about individual modes, "Quality Preset" scales itself to whichever codec you picked.

Why can I only choose Mono, and why is the sample rate locked?

Because those are properties of the codec, not restrictions we added. AMR narrowband is defined at 8000 Hz and AMR wideband at 16000 Hz, each frame covering 20 milliseconds — 160 samples for narrowband, 320 for wideband — and this encoding path produces a single channel. Offering you 48 kHz stereo would be offering something the encoder cannot produce. Every AMR file you have ever received from a phone has these same properties.

Will my camcorder's music or background sound survive the conversion?

No, and this is the honest answer rather than a hedge. Narrowband AMR can only represent audio up to 4 kHz and wideband up to 8 kHz, both in mono, at rates between roughly 5 and 24 kbps. Speech is intelligible because the codec is built around a model of the human voice; music, applause, rain, traffic and room ambience have nothing to do with that model and come out sounding artificial. AMR is the right choice for a talking head and the wrong choice for everything else a camcorder records.

How small will the file actually be?

Very. At narrowband's top mode of 12.2 kbps an hour of audio is roughly 5.5 MB; at wideband's top mode of 23.85 kbps it is roughly 10.7 MB. For comparison, the same hour as 128 kbps AAC is about 58 MB, and as uncompressed 48 kHz 16-bit stereo about 691 MB. That ratio is why AMR persists in voice-note and telephony workflows long after better codecs arrived — for spoken word where storage or bandwidth is the binding constraint, nothing much beats it on size.

Should I use AMR or Opus for a speech recording?

Opus, unless something specific requires AMR. Opus handles speech efficiently at comparable low bitrates, supports stereo and full-bandwidth audio when the material needs it, and degrades far more gracefully when a recording turns out to contain more than talking. The case for AMR is compatibility: old handsets, voice recorders, some telephony and dispatch systems, and archives that already consist of .amr files all expect it. If the destination is a general-purpose computer or phone from the last decade, Opus is the better engineering choice.

Can I convert only part of a long recording?

Yes. Switch the "Trim" control away from "Unchanged" and set a start time and a duration, in seconds or as HH:MM:SS.sss, and only that span is decoded and encoded. On a long camcorder file this is usually what you want, since the useful spoken section is often a few minutes inside an hour of footage. A general-purpose audio cutter will not accept an MTS file as input, so trimming here — or on the video before you convert — is the practical route.

How are my files handled while they convert?

Your MTS is uploaded to our servers over an encrypted connection, the audio track is decoded and re-encoded to AMR there, and both the upload and the result are deleted automatically after a few hours. There is no sign-up, no watermark, and your recordings are never shared or made public. Expect the upload to dominate the time: AVCHD records at up to 24 Mbit/s, so you are sending gigabytes of video in order to receive a few megabytes of speech back.