Initializing... drag & drop files here
Supports: AVCHD
AVCHD is the format Sony and Panasonic camcorders record to — H.264/AVC video wrapped in an MPEG transport stream, the .mts (or .m2ts) files on the camera's memory card. This tool pulls a single frame out of that HD video at a timestamp you choose and saves it as a still WebP image — not an animated one. This walk-through covers picking a sharp frame, avoiding the combing you get on interlaced footage, and choosing lossless vs lossy so the still stays small.
.mts/.m2ts file, or click "+ Add Files" to select it. Files are uploaded over an encrypted connection and deleted automatically a few hours after conversion — no sign-up, no watermark.2.100 means 2 seconds and 100 milliseconds into the clip.The two settings that decide whether your still looks good are the timestamp and the lossless toggle.
12.5 is twelve and a half seconds in. AVCHD typically records at 25–30 frames per second, so neighbouring frames are only ~33–40 ms apart — nudge the value by a few hundredths if the frame you land on is mid-blink or mid-blur..m2ts files are large; on a slow connection the upload, not the conversion, is the bottleneck. Trim the clip first, or extract from a shorter segment.If the camcorder footage is copy-protected (some AVCHD discs and managed imports carry protection), the frame can't be extracted until the file is exported as a plain .mts/.m2ts. Frames from heavily compressed or low-bitrate AVCHD clips will show H.264 blocking that no still-image setting can remove — that detail simply isn't in the source. And if you actually wanted a short looping clip rather than a frozen photo, you want an animated format instead: convert the AVCHD to GIF and keep it as a moving image.
A single still image. The Specific Frame mode captures exactly one frame at the time you enter and writes it as a static WebP. WebP can hold animation, but this tool's still mode does not — if you want motion, use an animated format like GIF instead. Choose Multiple Screenshots if you want several separate still images sampled across the clip.
Type the moment into the Time (seconds) field. It accepts decimals, so 5.75 is 5.75 seconds in and 2.100 is 2 seconds plus 100 milliseconds. There's no scrubber preview, so if you miss the moment, nudge the value by a few hundredths of a second — at 25–30 fps each frame is only about 33–40 ms wide — and convert again.
Because the clip is interlaced. AVCHD camcorders commonly record 1080i, where one frame is woven from two fields shot a fraction of a second apart; on a moving subject those fields disagree and you see combing. Pick a frame where the subject isn't moving, or extract from low-motion footage. Clips recorded in progressive 1080p don't have this issue.
Use lossless for screenshots, diagrams, logos, or anything with hard edges and text — it's pixel-exact and averages about 26% smaller than PNG. For a photographic frame from real-world footage, leave Lossless? off: lossy WebP at "Very High" looks identical to most eyes and runs 25–34% smaller than the same frame saved as JPEG, so it's the better choice for the web.
No. You're extracting raw pixels from a single video frame, so the output is a fresh image without the camcorder's recording timestamp or any GPS tags the clip carried. If you need that information, read it from the original .mts/.m2ts before converting.
In our testing, a 1920×1080 frame from AVCHD footage saved as lossy WebP at "Very High" quality lands around 150–300 KB depending on scene detail — noticeably smaller than the same frame as JPEG and a fraction of an uncompressed PNG. Turning on lossless or keeping fine detail (foliage, textures) pushes it higher.