Initializing... drag & drop files here
Supports: MTS
You have .MTS clips off a Sony, Panasonic or Canon AVCHD camcorder, and something on the other end — an old Windows editor, a DVD-authoring tool, a media box — wants an .avi. This page walks through that conversion and is blunt about the two settings that trip people up here: the quality slider labelled "CRF" is not a CRF slider, and the preset that sounds safest makes your file bigger.
AVCHD is a 2006 Sony/Panasonic camcorder format. Per the official AVCHD specification it stores MPEG-4 AVC/H.264 video at 1920×1080 or 1440×1080 with Dolby Digital (AC-3) or linear PCM audio, at up to 24 Mbps. AVI is a Microsoft container from 1992. Nothing in that pairing is a copy operation — the H.264 is fully decoded and re-encoded into an older codec, so the settings below decide what you end up with.
.mts only; queue as many clips as you like and they all convert with the same settings.This is the part no other converter tells you, and it explains almost every "why is my AVI worse/bigger than the original" complaint.
The field labelled "CRF" is really qscale. With MPEG-4, DivX or Xvid selected, the encoder does not support CRF at all. The number you type is passed to the encoder as -q:v on a 1–31 scale where lower is better, not the 0–51 x264 scale. If you type 23 out of x264 muscle memory, you are asking for near-worst quality, not "good default". Two internal configuration contexts both claim this field, so the value it opens on is not dependable — always set it yourself rather than trusting whatever number is showing. In our own measurement on a 10-second 720p clip, -q:v 2 produced 23.5 MB and -q:v 23 produced 2.28 MB from the same source: a 10x swing driven entirely by that one field.
The Quality Preset does work here, but it scales the source bitrate. On a video source (unlike a still-image source) the preset multiplies the bitrate your camcorder actually recorded. The multiplier is 0.3 + level × 1.2, which means the default sits above 1.0 — the reason so many people end up with an AVI larger than the MTS they started with.
| Quality Preset | Multiplier | A 17 Mbps AVCHD clip becomes |
|---|---|---|
| Highest | 1.50x | ~25.5 Mbps (much bigger) |
| Very High (the default) | 1.32x | ~22.4 Mbps (bigger) |
| High | 1.14x | ~19.4 Mbps (bigger) |
| Medium | 0.90x | ~15.3 Mbps (first rung that shrinks) |
| Low | 0.72x | ~12.2 Mbps |
| Very Low | 0.54x | ~9.2 Mbps |
| Lowest | 0.30x | ~5.1 Mbps |
Medium is the first preset that produces a smaller file than the source. Spending a step on "High" to save space does the opposite.
| Mode | What you set | Behaviour with MPEG-4 / DivX / Xvid |
|---|---|---|
| Quality Preset | A named preset | Scales the source bitrate by the table above. Default is Very High. |
| Specific file size | A size in MB (opens at 24) | Targets that output size; Auto Scale may drop resolution to reach it. |
| Constant Bitrate | One bitrate | Fixed rate throughout, simple and predictable. |
| Variable Bitrate | Target, min and max | Rate moves with scene complexity within your bounds. |
| Constant Quality | A number labelled "CRF" | Emits -q:v on a 1–31 scale, lower is better. Works — just not a CRF scale. |
| Constraint Quality | That number plus Max bitrate | Same qscale, with a hard ceiling. The most predictable choice here. |
.mts. Files copied off a camcorder through software are often renamed .m2ts, which needs the sibling M2TS to AVI page even though the underlying stream is the same.AVI is rarely the right answer in 2026. It has no standardised aspect-ratio metadata, no subtitle or attachment support, and its original design predates the B-frame referencing that modern codecs rely on. If your goal is just a file that plays widely and stays small, convert MTS to MP4 instead. If you want to keep the AVCHD format but reduce the size, compress MTS directly. And if you only need a section of a long camcorder clip, cut it first with the video cutter so you are not re-encoding an hour of footage to keep 40 seconds.
Yes. The H.264 video inside an AVCHD file is decoded and re-encoded into an older codec, so this is generation loss by definition — there is no pass-through path. How much you lose is set almost entirely by the File Compression mode. A quality number of 2–4 under Constraint Quality is close to visually transparent; anything above about 15 on that 1–31 scale is visibly soft.
Because the Quality Preset defaults to Very High, which targets 1.32 times your camcorder's actual recorded bitrate. Presets on a video source scale the source rate rather than setting an absolute quality, and every preset above Medium multiplies upward. Choose Medium (0.90x) or lower, or switch to Constraint Quality and set the size yourself.
Something between 2 and 10, and treat lower as better. With MPEG-4, DivX or Xvid selected, that field is not an x264 CRF — it is qscale, capped to 1–31, where 1 is best quality. The familiar x264 advice of "23 is a good default" produces a badly degraded file here. If you switch Video Codec to H.264, the same box changes to a real 0–51 CRF scale and 23 becomes reasonable again.
No. There is no deinterlacing step anywhere in this conversion, so the combing your source has is copied into the output frames. Converting to MP4 or any other container does not change this either — that is a common misconception. Deinterlacing has to happen in an editor or player that offers the filter explicitly.
Pick MPEG-4 (the default) or DivX if the target is a genuinely old editor, DVD tool or set-top player, since those are the MPEG-4 Part 2 codecs such software expects. Pick H.264 if the destination just wants the .avi extension but runs on anything modern — it gives noticeably better quality per megabyte and unlocks a real CRF scale. Xvid is functionally interchangeable with DivX for playback purposes.
MP3 is preselected and is the most reliably decoded choice for old AVI players. Choose AC3 if your camcorder recorded a 5.1 surround mix and you want to keep the channel layout intact, since AVCHD's Dolby Digital track can carry surround. Uncompressed PCM avoids a second lossy audio pass at the cost of a much larger file.
Yes — Video resolution opens on "Keep original", so no scaling happens unless you ask for it. Note that the preset resolutions set height only, with width following the source aspect ratio, so "1080p" on a 1440×1080 anamorphic clip will not give you 1920×1080. Use Width x Height when you need exact dimensions.
Yes, and it is the normal way to work with AVCHD, since long recordings get split into multiple .mts segments on the card. Queue them together and they convert with identical settings. They stay separate files, though — this converter does not join segments back into one continuous clip.
Your .MTS file is uploaded over an encrypted connection and converted on our servers. Uploaded files and their results are deleted automatically after a few hours, are never shared or made public, and there is no sign-up and no watermark on the output.