Initializing... drag & drop files here
Supports: M2TS
An .m2ts file is a BDAV MPEG-2 transport stream — the container Blu-ray discs use and the one AVCHD camcorders from Sony, Panasonic, Canon and JVC write to their memory cards. A JFIF is a single still picture: JPEG File Interchange Format, which is the specification behind every file people call a JPEG. This page decodes one frame from the recording and saves it as that picture.
The conversion itself is simple. What trips people up with .m2ts sources is finding the right frame in the first place, because camcorder recordings are usually split across several numbered files, and because the interlaced footage most consumer camcorders produce does not freeze cleanly. Both are covered below.
.m2ts onto the page or click "Add Files". AVCHD clips are large, so upload is normally the slowest step. Several clips can be queued and they share one set of frame and quality settings.0, which on a camcorder clip is the moment the record button was pressed — often a swing, a lens still racking focus, or a hand leaving the frame. Type the second you want; decimals such as 18.75 work. "Multiple Screenshots" with a Capture Rate samples the whole clip instead..jfif. Files are uploaded over an encrypted connection, processed on our servers and deleted automatically after a few hours — no sign-up, no watermark, never shared or made public.An AVCHD card is a folder structure, not a single movie file, and camcorders that format their cards as FAT32 have to break a long recording into parts of roughly two gigabytes. That means a two-hour event is spread across a numbered series, and the moment you remember is very likely not in the first file.
| Where to look | What it holds | Why it matters for a frame grab |
|---|---|---|
BDMV/STREAM/00000.MTS (or .m2ts) |
The actual picture and sound | This is the file to upload. Numbering continues across the split, so check the later ones. |
BDMV/PLAYLIST/*.mpls |
Playlists that stitch the parts back into one timeline | Explains why the camcorder shows one long clip while the card holds several files; the playlist itself has no picture in it. |
BDMV/CLIPINF/*.cpi |
Per-clip index information | Not media; nothing to convert. |
AVF_INFO, PRIVATE, thumbnail folders |
Camcorder housekeeping and previews | Ignore. |
Practically: if a specific moment refuses to appear where you expect, work out which numbered part covers it by clip duration, or scout with "Multiple Screenshots" at a slow Capture Rate before committing to an exact timestamp. And if the folder holds .MTS files rather than .m2ts, that is the same transport stream with plain 188-byte packets instead of the timestamped 192-byte BDAV packets — use the MTS page for those rather than renaming.
| Source | Typical frame | Scan | What to expect in a freeze frame |
|---|---|---|---|
| Consumer AVCHD camcorder, 1080i | 1920 x 1080 | Interlaced 50i/60i | Comb artefacts on anything moving; nothing in this pipeline de-interlaces |
| AVCHD camcorder, 1080p mode | 1920 x 1080 | Progressive | Clean stills; AVCHD 2.0 allows up to about 28 Mbit/s at 1080p60 |
| AVCHD Lite (older cameras) | 1280 x 720 | Progressive | Cleaner than 1080i for stills, at lower detail |
| Blu-ray rip | 1920 x 1080 | Usually progressive | Film grain and heavy compression show up in a still that playback hides |
| AVCHD Progressive / 4K derivative | 3840 x 2160 | Progressive | Large stills; a top-preset JFIF from a UHD frame runs into the megabytes |
The single biggest quality question is scan type. 1080i is two fields captured a fraction of a second apart and woven into one frame; playback hides that, a still does not. Choosing a moment where the subject and the camera are both still is the only fix available on this page.
.jfif the same thing as a .jpg?Yes, as far as the image data is concerned. JFIF is the interchange specification that describes JPEG files on disk, and .jfif, .jpg, .jpeg, .jpe and .jif all name the same encoded picture. Renaming the download to .jpg changes nothing about the bytes and is worth doing if a photo app, printing service or upload form refuses to recognise .jfif. If you would rather not rename anything, M2TS to JPG performs the identical decode and hands you a .jpg directly.
Because the recording is interlaced, which most consumer camcorders shooting 50i or 60i are. Each stored frame contains two fields captured at different instants, so freezing it shows both moments at once. There is no de-interlace control on this page and no de-interlacing anywhere in the pipeline, so choose a frame from a static shot, or de-interlace the resulting picture in an image editor.
"Very High (Recommended)" for most purposes. If file size matters, "High" is the last rung that keeps colour at full resolution — below it the encoder halves the chroma planes in both directions, which roughly halves the file but visibly softens coloured text, on-screen graphics and saturated edges. "Highest" adds size without much visible gain on footage that has already been through AVCHD's compression.
Yes. Switch Frame Selection to "Multiple Screenshots" and pick a Capture Rate. The entries are intervals rather than frame rates — "1 second per frame" writes one still per second of runtime, which on a twenty-minute recording is 1,200 files. Start at five or ten seconds per frame to find the section you want, then re-run with "Specific Frame" and an exact time.
.m2ts at all?Because .m2ts uses BDAV framing: each 188-byte transport packet is prefixed with a four-byte arrival timestamp, making 192-byte packets, so the file does not begin with the usual sync byte at offset zero. Tools written for plain broadcast transport streams sometimes reject that outright. Our decode handles it, which is why a clip your editor will not import can still yield a frame here.
The recording's own frame size, because Image resolution opens on Keep original — 1920 x 1080 from most AVCHD footage, 1280 x 720 from AVCHD Lite, 3840 x 2160 from a UHD source. If you need something smaller, note that Preset Resolutions sets height only and lets width follow the source aspect, so a 16:9 frame at the 720p preset comes out 1280 x 720. Use Width x Height when the exact pixel dimensions matter.
No. AVCHD stores date, time and camera settings in the stream's metadata and in the card's index files, and a JPEG's own EXIF block is written by the encoder rather than copied from the video, so that information does not travel with the still. If provenance matters, note the timestamp from the original clip before converting.
It is discarded — a still image has no audio capability. AVCHD soundtracks are normally Dolby Digital or uncompressed LPCM at 48 kHz, and if you need them, run the same file separately through M2TS to WAV or keep the original clip. Nothing on this page touches the audio.
JFIF when the picture is for sharing, printing or embedding — it is far smaller and opens everywhere. M2TS to PNG when you plan to crop, annotate and re-save repeatedly, since each JPEG save adds a new generation of artefacts while PNG does not, and when the frame is dominated by graphics or text rather than photographic detail.
Trim first. AVCHD runs up to roughly 28 Mbit/s, so a long recording is several gigabytes, and pushing all of it up for one still is wasteful. Trim M2TS cuts the clip down to the section you care about, and Video Cutter accepts .m2ts directly as well. Whatever you upload is transferred over an encrypted connection, processed on our servers and deleted automatically after a few hours.