Reset to defaults
Need help?Extract a PNG Still from AVCHD Footage
Pull one frame out of a camcorder recording and save it as a PNG. AVCHD clips arrive as .MTS or .M2TS files from Sony, Panasonic, Canon, and JVC cameras, carrying H.264 video at up to 1920×1080 — enough resolution that a well-chosen frame makes a usable photo. You pick the exact moment by timestamp, and PNG stores it without the blocking artefacts a JPEG re-compression would stamp on top.
How to Convert AVCHD to PNG
- Upload Your AVCHD File: Drag and drop the
.MTSor.M2TSclip, or click "+ Add Files" to browse. Queue several clips if you want a still from each. - Set the Frame Time: Under Frame Selection, Specific Frame is already chosen and Time (seconds) starts at
0— the very first frame of the clip. Type the moment you actually want, decimals included:12.25lands a quarter-second past the twelve-second mark. - Choose a Quality Preset and Resolution (Optional): Quality Preset starts at Very High (Recommended); move it to Highest for a true-colour lossless PNG. Image resolution starts on Keep original, or scale down with Preset Resolutions, Resolution Percentage, or Width x Height.
- Convert and Download: Click Convert and save the PNG. 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.
PNG vs JPG for a Camcorder Frame
Both start from the identical decoded frame; what differs is what the encoder does to it next.
| Property | PNG (this tool) | JPG (AVCHD to JPG) |
|---|---|---|
| Compression | Lossless at the Highest preset; a palette quantiser at the presets below it | Always lossy |
| Visible artefacts | None at Highest; slight colour banding on gradients at lower presets | Blocking and ringing around edges and gradients |
| Typical size, 1080p frame | Several megabytes at Highest, much smaller once quantised | Usually well under a megabyte |
| Colour depth | Up to 8 bits per channel true colour | 8 bits per channel, chroma-subsampled |
| Transparency | Supported by the format (a video frame has none to carry) | Not supported |
| Re-saving repeatedly | Safe at Highest — no generational loss | Degrades a little on every re-save |
| Best for | Editing, archiving, screenshots with text or graphics | Email, web thumbnails, sharing where size wins |
What Each PNG Setting Actually Does
| Setting | Starts at | What it changes |
|---|---|---|
| Frame Selection → Specific Frame | Selected, Time (seconds) = 0 | Which single frame is grabbed; type any timestamp, decimals allowed |
| Frame Selection → Multiple Screenshots | Capture Rate = 1 second per frame | Samples frames across the whole clip instead of grabbing one |
| Quality Preset | Very High (Recommended) | Highest writes a true-colour lossless PNG; every preset below it runs a palette quantiser that trims file size at a small cost in colour fidelity |
| Image resolution | Keep original | Native camcorder frame size, or scale with presets, percentage, or explicit width and height |
| Colors | Original, or a reduced palette | Caps the palette at 256 down to 2 colours, with an optional Dither toggle — useful for flat graphics, harmful on photographic frames |
| Compression level | 6 (range 1–10) | How much effort the encoder spends packing the file; higher takes longer |
| Compression speed | 4 (range 1–10) | Trades encoder time against turnaround on large frames |
Frequently Asked Questions
Is the PNG I get actually lossless?
Only at the Highest Quality Preset. That setting writes a true-colour PNG whose pixels match the decoded frame exactly. The default Very High preset — and everything below it — runs a palette quantiser that reduces the number of distinct colours to shrink the file. On a photographic camcorder frame the difference is usually invisible, but if you are archiving, print-preparing, or planning further edits, move the preset to Highest first. Note that neither setting can undo the lossy H.264 compression the camcorder applied at capture.
Which frame do I get if I leave the settings alone?
The frame at zero seconds — the very first frame of the clip. On camcorder footage that is often the least useful moment, since recordings tend to start with the camera settling or the operator's hand in shot. Set Time (seconds) to the moment you actually want; the field accepts decimals, so 4.5 is four and a half seconds in.
Why does my frozen frame show jagged comb lines?
Because the clip is interlaced. AVCHD's most common consumer modes are 1080/60i and 1080/50i, where every displayed frame is woven from two fields captured a fraction of a second apart. Freeze a moment where something moved between those fields and the mismatch appears as horizontal teeth. Nudge Time (seconds) a few hundredths until you land on a still moment, or work from progressive footage — 1080/24p, 1080/50p, and 1080/60p clips never comb.
Will the PNG be full 1920×1080?
Yes, as long as Image resolution stays on Keep original. AVCHD records 1920×1080, 1440×1080, or 1280×720 depending on the camera mode, and the extracted still keeps whichever the source used. There is nothing above that to recover: the AVCHD specification has no 4K mode, so 1920×1080 is the format's ceiling.
Can I get every frame instead of just one?
Switch Frame Selection from Specific Frame to Multiple Screenshots and set a Capture Rate — anything from one frame every ten seconds down to a frame every 0.1 seconds. Each sampled frame comes back as its own PNG rather than as one combined file. If what you actually want is the moving footage in a shareable container, convert the clip with AVCHD to MP4 instead.
Does the PNG keep the recording date or GPS tags from the camcorder?
No. You are extracting raw pixels from inside a video stream, so the output is a fresh image with no camera timestamp, no GPS coordinates, and no lens metadata. If that information matters, read it from the original .MTS/.M2TS file before converting — camcorders store it in the stream and its companion index files, not in individual frames.
Why is my PNG so much bigger than a JPG of the same frame?
Because at the Highest preset PNG stores every pixel exactly, while JPEG throws away detail the eye is unlikely to miss. A detailed 1080-line photographic frame simply needs a lot of bytes to be described losslessly — as a rough floor, 1920 × 1080 pixels × 3 bytes of true colour is about 6.2 MB before any compression at all. If size matters more than perfect fidelity, use a lower Quality Preset here or switch to AVCHD to JPG.
My camera made an AVCHD folder, not a single file — what do I upload?
Camcorders store AVCHD as a directory tree, typically PRIVATE/AVCHD/BDMV/STREAM/ on solid-state models, where the actual recordings are the numbered .MTS files inside STREAM. Upload one of those directly; the top-level AVCHD folder is not a media file. Clips copied to a computer are often renamed .M2TS, which is the same H.264 transport stream and works here too. For files that carry the raw camera extension, MTS to PNG is the tuned route.