AVCHD to JPG Converter

Grab a JPG Still from AVCHD Footage: What This Tutorial Covers

AVCHD is the high-definition format Sony and Panasonic introduced in 2006 for camcorders: MPEG-4 AVC/H.264 video multiplexed into an MPEG-2 transport stream and written to the card as .MTS clips. Any single frame of that stream is a full-HD photograph waiting to be pulled out. This walk-through covers picking an exact frame by timestamp, sampling a run of stills across a clip, choosing a quality preset that does not throw away the detail your camera recorded, and dodging the two artifacts that catch people out — interlace combing and squashed anamorphic frames.

How to Convert AVCHD to JPG

1. Upload Your AVCHD File: Drag and drop the .MTS or .M2TS clip onto the page, or click "+ Add Files" to browse. Several clips can be queued at once and each is processed with the same settings.
2. Set Frame Selection: The tool opens on "Specific Frame" with Time (seconds) at 0, which captures the first frame. Type a timestamp such as 2.100 — 2 seconds and 100 milliseconds — to capture a different moment, or switch to "Multiple Screenshots" and choose a Capture Rate to sample stills across the whole clip.
3. Choose the Quality Preset and Image resolution: Under Image Compression the Quality Preset starts on "Very High (Recommended)"; leave it there for the cleanest still or step it down for a smaller file. Image resolution starts on "Keep original", which keeps every pixel the camera recorded.
4. Convert and Download: Click Convert and save the image. Your file is uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours. No sign-up, no watermark.

Walk-through: Choosing the Frame and the Quality Preset

The capture runs in two stages. The requested frame is first decoded out of the H.264 stream and written as a lossless PNG, and only then is it encoded as a JPG at the preset you picked. That matters: the still is not re-compressed twice on our side, so the only lossy step added to the camera original is the single JPG encode.

The Time (seconds) field takes decimals down to the millisecond, so the practical workflow is to scrub the clip in any player, read off the moment you want, and type it in. A few settings decide how the result turns out:

• One hero shot: "Specific Frame", an exact timestamp, Quality Preset on "Very High (Recommended)", Image resolution on "Keep original".
• A contact sheet to choose from: "Multiple Screenshots" with a Capture Rate of 1 second per frame (or slower — the list runs out to 10 seconds per frame) so a long clip does not produce thousands of images.
• Smaller files for email or the web: drop the Quality Preset a notch or two, or use "Resolution Percentage" or "Preset Resolutions" to scale the frame down before it is encoded.
• Maximum fidelity for editing: use the "Highest" preset, or skip lossy compression altogether with AVCHD to PNG.

What Frame Size Each AVCHD Mode Gives You

AVCHD's published video specification defines a handful of recording modes, and the still you extract matches the mode the camera was in:

JPG or PNG for a Camcorder Still

Common Errors and How to Fix Them

• Comb-like horizontal stripes on moving subjects — the clip was shot in one of AVCHD's interlaced modes (1080/60i or 1080/50i). Each stored frame is two fields captured a moment apart, and no deinterlacing filter is applied when the frame is grabbed, so movement between the fields shows as stripes. Pick a low-motion frame, or use footage shot in a progressive mode.
• The still is black or empty — the timestamp fell before the first frame or past the end of the clip. Set Time (seconds) to a point you know has picture, a couple of seconds in.
• The image looks horizontally squashed — 1440×1080 AVCHD modes record non-square pixels that a player stretches to 16:9 on playback. A frame grabbed from one of those modes can come out at its stored 1440×1080 shape; stretch it back to 1920×1080 in any editor.
• The file is too large to attach — a full-resolution still at the top preset can run to a few megabytes. Lower the Quality Preset before converting, or squeeze the result afterwards with the Image Compressor.
• The wrong moment was captured — Time (seconds) is a time value, not a frame number: 90 means one minute thirty seconds into the clip, not the ninetieth frame.

When This Doesn't Work

A frame grab only succeeds if the clip decodes cleanly. Files truncated when a battery died or a card was pulled mid-write may not seek to an arbitrary timestamp, and broadcast-style copy protection cannot be decoded by any web tool. If every usable moment in the clip is fast motion shot interlaced, no single grabbed frame will freeze cleanly — nothing in this pipeline deinterlaces, so the combing stays baked in. The fix is to deinterlace the footage in a desktop editor first and pull the still from that progressive export; an editor also lets you choose the deinterlacing algorithm, which a one-click grab never can.

Frequently Asked Questions

Which frame do I get if I never touch the timestamp?

The first one. Frame Selection opens on "Specific Frame" with Time (seconds) set to 0, so an untouched conversion captures frame zero. Camcorders usually roll a moment before anything interesting happens, so it is worth reading a better timestamp off a player and typing it in — decimals are accepted, so 12.250 lands at 12.25 seconds.

Why does my frame have comb-like lines through moving subjects?

Because AVCHD's 1080/60i and 1080/50i modes are interlaced. Each stored frame is woven from two fields captured a fraction of a second apart, so a subject that moved between them appears as interleaved stripes when frozen. Nothing is deinterlaced during capture, so the cure is to choose a frame with little motion, or to shoot in one of the progressive modes — 720/60p, 720/50p, 720/24p, or 1080p on AVCHD Progressive cameras.

What resolution will the JPG be?

Whatever the source frame is, since Image resolution starts on "Keep original". A 1080-line clip yields a 1920×1080 still of roughly 2.1 megapixels; 720p footage yields 1280×720. AVCHD's HD modes top out at 1920×1080, so that is the ceiling — you can scale a still down with "Preset Resolutions" or "Resolution Percentage", but no setting adds detail the camera never recorded.

Is the still compressed twice, once as H.264 and again as JPG?

Only once on our side. The frame is decoded from the H.264 stream and written as a lossless PNG, then encoded a single time into JPG at your chosen preset. The H.264 compression the camera applied is of course already baked into the footage, but the conversion itself does not stack an extra generation of JPG loss on top.

Can I pull several stills from one clip in a single pass?

Yes. Switch Frame Selection to "Multiple Screenshots" and pick a Capture Rate — one frame per second is a sensible starting point, and the list goes down to 0.1 second per frame for near-frame-by-frame sampling and up to 10 seconds per frame for long recordings. Every sampled frame is converted with the same quality and resolution settings.

Why does my still look horizontally squashed?

Because some AVCHD modes record 1440×1080 with non-square pixels and rely on the player to stretch them to a 16:9 shape. A still taken from that kind of clip can arrive at its stored 1440×1080 size and therefore look narrow. Resizing it to 1920×1080 in any image editor restores the intended proportions; nothing has been lost.

Is the file I download named .jpg or .jpeg, and does the difference matter?

This tool writes .jpg, and a File extension control under the options lets you switch the label to .jpeg if something downstream expects that spelling. They are the same format — the three-letter form is a leftover from MS-DOS filename limits, and every viewer, browser, and editor reads both, so renaming one to the other changes nothing about the file.

What happens to my camcorder footage after I upload it?

It is 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. The practical constraint is upload time rather than conversion time, because AVCHD carries full HD video at up to 24 Mbps (28 Mbps for 1080p progressive recordings); if you only want a still from one moment of a long take, trim the clip first and upload the short version.