M2TS to EPS Converter

Putting an AVCHD Frame Into a Print Document: What This Covers

An .m2ts is what a Blu-ray disc or an AVCHD camcorder writes. An EPS is what a print workflow wants placed into a layout. Bridging the two is a legitimate job — a frame grab for a programme, a still from event footage for a printed brochure — and it works, but two things about it need explaining before you run it.

The first is that an EPS made from a video frame is a raster image wrapped in PostScript, not a vector drawing. The second is that the file will be much bigger than you expect, for a reason that is entirely predictable once you know it.

How to Convert M2TS to EPS

1. Upload Your M2TS File: Drag the .m2ts onto the drop zone or click "Add Files". AVCHD recordings are large; use the source's own timeline to note the moment you want before uploading.
2. Pick the Frame Under Frame Selection: The first frame is the default and is very often a poor choice on camcorder footage. Use Specific Frame with a time in seconds, or Multiple Screenshots with a capture rate, to get something usable.
3. Set Resolution Percentage If the EPS Must Be Smaller: This page's Image resolution group offers only Keep original and Resolution Percentage — there are no width and height boxes here. Percentage is the one size lever, and it is effective.
4. Convert and Download: Click Convert. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours. No sign-up, no watermark.

Why the EPS Is So Large

The frame is extracted as a lossless PNG and then embedded in a PostScript program by ImageMagick. Reading the produced file back gives the exact structure:

Uncompressed ASCII hex means every pixel is written as six hexadecimal characters. There is no compression stage to claw any of it back, so the file size is a direct function of the frame's pixel count:

A single 1080p frame is a 12.6 MB EPS. Dropping Resolution Percentage to 50% quarters that to roughly 3.2 MB, because the cost scales with pixel count rather than with picture content.

The 72 dpi density has a layout consequence too: at one point per pixel, a 1920 × 1080 frame declares a bounding box 26.67 inches wide by 15 inches tall. Your layout application will place it at that physical size and you will scale it down on the page. Placing it at 6.67 inches wide gives you roughly 288 effective dots per inch, which is a sensible print resolution.

The Interlacing Problem Is Worse Here Than Anywhere Else

AVCHD's most common recording mode is 1080i, and there is no deinterlacing anywhere in this pipeline. An interlaced frame contains two fields captured milliseconds apart, woven together line by line, and both of them end up in the still.

We demonstrated this rather than asserting it: a test stream whose two fields were solid blue and solid red produced an extracted frame whose scan lines run blue, blue, red, red, blue, blue down the picture. On real footage that shows up as horizontal tearing on every moving edge.

Print is unforgiving about this. On a screen at a small size the combing can pass unnoticed; at 288 dpi on paper it is obvious. Three things help:

• Choose a static moment. Use Frame Selection to land on a shot with a locked-off camera and little subject movement. Combing only appears where something moved between the two fields.
• Check before you commit. Convert to M2TS to PNG first, look at the frame at 100%, and only then produce the EPS. It is far quicker than inspecting a 12 MB PostScript file.
• Deinterlace upstream. If the shot you need has motion in it, deinterlace in a video editor and export a still from there.

Frequently Asked Questions

Will the EPS scale up without losing quality?

No. The EPS contains the video frame as embedded pixels, so enlarging it magnifies those pixels exactly as enlarging a JPEG would. Nothing here traces the image into vector paths. EPS is being used as a container for a raster, which is a perfectly normal use of the format but not the one its reputation suggests.

Why is a single still bigger than a chunk of the video it came from?

Because the still is uncompressed and text-encoded while the video is compressed across frames. At roughly 6.08 bytes per pixel written as ASCII hexadecimal, a 1080p frame is 12,606,819 bytes. Several seconds of AVCHD at a typical bitrate can weigh less than that one frame.

How do I make the EPS smaller?

Resolution Percentage, which is the only size control this page exposes. Because the byte cost is proportional to pixel count, 50% gives you a quarter of the file and 25% gives you a sixteenth. There is no quality slider to trade against, since nothing in the EPS is compressed.

What resolution will the printer actually get?

That depends on how large you place it. The EPS declares 72 dots per inch, so a 1920-pixel-wide frame arrives 26.67 inches wide. Scale it down in your layout: at 6.67 inches wide you get about 288 dpi, at 5 inches about 384 dpi. The pixels do not change, only the physical size they occupy.

Why can I not enter an exact width and height?

The EPS path exposes proportional scaling only — Keep original and Resolution Percentage, nothing else. The bounding box is derived directly from the pixel dimensions at one point per pixel, so an arbitrary stretch would put the declared page geometry at odds with the embedded image data.

My frame has horizontal stripes across the moving parts. Can the converter fix it?

Not on this page. There is no deinterlacing filter in the pipeline at all, so both fields of a 1080i frame land in the still. Pick a frame with less motion using Frame Selection, or deinterlace the footage in an editor and export the still from there.

Why does my .m2ts play in some tools and not others?

Because M2TS uses 192-byte BDAV packets — a four-byte arrival timestamp followed by the 0x47 sync byte at offset 4 — rather than the plain 188-byte packets of a regular transport stream. Tools that only understand plain TS framing reject it. That is also why renaming an .m2ts to .ts rarely helps; the packet layout differs, not just the name.

Will Illustrator or InDesign open the result?

Yes. It is a conforming EPSF 3.0 file with a valid bounding box and both place it normally. Illustrator will show a placed raster image rather than editable paths, which is exactly what it is.

Can I take several frames in one pass?

Yes — set Frame Selection to Multiple Screenshots and choose a capture rate anywhere from one frame every tenth of a second to one every ten seconds. Bear the size in mind: at 1080p each EPS is around 12.6 MB, so a fast capture rate over a long clip produces a great deal of data very quickly.