Initializing... drag & drop files here
Supports: MTS
.mts is AVCHD camcorder footage — H.264 video at up to 24 Mbit/s with a Dolby Digital soundtrack, from the format Sony and Panasonic launched in 2006. AV1 is the codec from the Alliance for Open Media that replaced it as the efficiency benchmark: same picture, far fewer bits, no licensing to worry about. Re-encoding a shelf of old camcorder files to AV1 is one of the few conversions that is genuinely worth the CPU time, because AVCHD's bitrates were set by what a memory card could sustain in 2006 rather than by what the picture needs. Before you start, one thing to know: the file you get back is a Matroska container wearing an .av1 extension, and you will want to rename it.
.mts clips onto the page or click "Add Files". AVCHD cameras split long takes into numbered files on the card; convert each one and join them afterwards if you need a single clip.There is no such thing as a general-purpose .av1 container format, and we do not write a raw elementary stream either. What comes out of this conversion is a Matroska file — the same container as .mkv — holding an AV1 video track and your chosen audio track, saved under an .av1 name. That is deliberate: an audio track needs a container to live in, and Matroska is the one that carries AV1 properly.
The practical consequence is a rename. Most players sniff the file contents and will play it anyway, but plenty of media libraries, TVs and upload forms decide what to do from the extension alone and will refuse a file they have never heard of. Change clip.av1 to clip.mkv and it behaves like any other Matroska file — same bytes, no re-encoding, no quality change. If you would rather skip that step entirely, MTS to MKV produces a correctly named container from the start.
AV1 is a CRF codec, which means all six rate-control modes on this page work — unlike several legacy targets where most of them do nothing. The AV1 quality scale runs from 0 to 63, and lower numbers mean better quality and bigger files.
| Mode | What it does on AV1 | Use it when |
|---|---|---|
| Quality Preset | Maps each step to a fixed CRF: Highest 18, Very High 20, High 25, Medium 33, Low 45, Very Low 50, Lowest 55 | Almost always — this is the mode the page opens in |
| Constant Quality | Sets the CRF value directly on the 16–63 scale | You want a specific CRF rather than a named step |
| Constraint Quality | CRF plus Qmin, Qmax and a maximum bitrate ceiling | Streaming or a device with a hard bitrate limit |
| Constant Bitrate | A fixed bitrate, with an optional Auto Scale that resizes to help hit it | Rarely — CRF is a better fit for archival re-encodes |
| Variable Bitrate | A target bitrate with minimum and maximum bounds | You are matching a delivery specification |
| Specific file size | Searches for settings that land near a size you name | A hard size limit, for example an upload cap |
For converting a camcorder archive, leave it on Quality Preset. "Very High" at CRF 20 keeps the picture essentially indistinguishable from the AVCHD source while typically cutting the file size substantially, because AVCHD's H.264 encoding was tuned for real-time recording rather than for compression efficiency. If storage is the point of the exercise, "High" at CRF 25 is still very good and noticeably smaller. Two notes on the numeric controls: the Constant Quality box opens at a preset value, so set it deliberately rather than trusting whatever number is showing, and remember that with AV1 a change of six CRF points roughly doubles or halves the file size.
Three limitations worth knowing before you re-encode an archive, because none of them can be fixed after the fact.
| Property | AVCHD (.mts source) | AV1 (output) |
|---|---|---|
| Developed by | Sony and Panasonic, 2006 | Alliance for Open Media |
| Video codec | H.264/AVC, Main or High Profile, Level 4.1 (4.2 in AVCHD 2.0) | AV1, encoded here with SVT-AV1 |
| Typical bitrate | Up to 24 Mbit/s, or 28 Mbit/s for AVCHD Progressive | Whatever CRF 20 needs — usually a fraction of the source |
| Audio | Dolby Digital AC-3 at 64–640 kbit/s, or LPCM on pro models | Opus by default; AAC, AC3, EAC3, DTS, FLAC, MP2, MP3 and Vorbis also offered |
| Container | MPEG-2 transport stream (.mts on the card, .m2ts once imported) |
Matroska, saved under an .av1 name |
| Scanning | Frequently 1080i interlaced | Whatever the source was — no deinterlacing happens |
| Licensing | Patent-licensed | Royalty-free |
| Encoding speed | Real time, in-camera | Slow — the trade-off you make for the compression |
Because the extension is unfamiliar, not because the file is broken. The output is a Matroska container — the same format as .mkv — holding an AV1 video track and an audio track, and it is only the .av1 name that confuses software which decides how to handle a file from its extension. Rename it to .mkv and it opens normally in VLC, MPV, Plex, Jellyfin and anything else that reads Matroska. No re-encoding is involved in the rename; the bytes are already correct.
Substantially, though the exact figure depends on the material. AVCHD's H.264 encoding was constrained by what a camcorder could do in real time in the late 2000s, so it leaves a lot of efficiency on the table, and AV1 at CRF 20 typically reproduces that picture at a fraction of the bitrate. Detailed, noisy handheld footage compresses less well than a static tripod shot; grain in particular is expensive to encode. If the numbers matter to your storage plan, convert one representative clip and measure before committing to a whole archive.
Because AV1 does far more work per frame. It is a much newer and more sophisticated codec than H.264, and the gains in compression come directly from analysis that costs CPU time. We encode with SVT-AV1, which is among the faster AV1 encoders, but a long AVCHD clip will still take considerably longer than the same clip converted to H.264. That trade is worth it for an archive you will keep, and usually not worth it for a file you are about to email.
Yes, if it was interlaced to begin with. Nothing in this pipeline deinterlaces, so the field structure is re-encoded as-is and the comb artefacts on horizontal motion survive into the AV1 file. This matters because 1080i was the standard AVCHD recording mode and the only option on many camcorders. If your footage shows the tell-tale horizontal stripes on moving edges, deinterlace it in a desktop tool before uploading — once it is baked into a re-encode, removing it cleanly is much harder.
Opus is preselected because it is the best-performing audio codec available for this container: at any given bitrate it sounds better than AAC or MP3, it is royalty-free like AV1 itself, and Matroska carries it natively. Keep it unless something downstream cannot read it. AAC is the safest alternative for broad device compatibility, AC3 or EAC3 preserve a surround mix in a form that home theatre equipment understands, and FLAC is there if the source was LPCM and you want to keep the audio lossless.
Increasingly, yes, but check first. AV1 hardware decoding is present in recent smart TVs, phones and GPUs, and modern browsers and desktop players handle it in software. Older hardware may decode it slowly or not at all, which shows up as stuttering rather than a refusal to play. If a specific device is the destination, test one converted clip on it before you re-encode an archive — and remember to rename to .mkv first, because an unfamiliar extension will often fail before the codec is even considered.
AV1 if you are archiving and playing back on modern equipment; HEVC if you need broader hardware support today. AV1 compresses better and is royalty-free, which is why it is the right long-term choice for a personal archive. HEVC has been in hardware for longer and is more widely supported by older TVs and phones. Note that our HEVC output has the same container quirk as this one — it also writes Matroska bytes under a codec-named extension — so the rename applies either way. If neither efficiency nor licensing matters and you just want the footage to play everywhere, MTS to MP4 is the simple answer.
Your MTS is uploaded to our servers over an encrypted connection, decoded and re-encoded to AV1 there, and both the upload and the result are deleted automatically after a few hours. There is no sign-up, no watermark, and your footage is never shared or made public. The two things that take time are the upload — AVCHD runs at up to 24 Mbit/s, so an hour of footage is several gigabytes — and the AV1 encode itself. For a large archive, converting in batches and checking one result before committing to the rest is the sensible approach.