Initializing... drag & drop files here
Supports: MJPEG
Both formats come from the same decade and both are still everywhere, but they solve opposite problems: MJPEG keeps every frame independent and pays for it in size, while MPG throws away the redundancy between frames and pays for it in editability. The short answer: convert to MPG when the destination is legacy hardware, disc authoring or an archive that expects the MPEG family, and stay on MJPEG when individual frames are the product. If the destination is anything modern, neither is the right answer.
| Property | MJPEG (source) | MPG (output) |
|---|---|---|
| What the file is | A bare stream of JPEG frames, no container | An MPEG program stream, a real multiplexed container |
| Specification | JPEG coding applied per frame | MPEG-1 Part 1 (ISO/IEC 11172-1) or MPEG-2 Part 1 (ISO/IEC 13818-1) |
| Compression style | Intra-frame only — every frame stands alone | Interframe — most frames store only what changed |
| Duration stored | No; ffprobe reports Duration: N/A |
Yes, written into the stream |
| Seeking | Poor to impossible without an index | Supported; this is what the container is for |
| Size on a static scene | Large; each near-identical frame is stored in full | Much smaller; the repetition is what gets removed |
| Frame independence | Complete — pull any single frame cleanly | Lost; most frames depend on their neighbours |
| Audio | None, ever | Supported by the container, if the source has any |
| Still made by | IP cameras, webcams, machine vision, microscopes | Disc authoring, broadcast archives, legacy playback chains |
| Patent status | Unencumbered | MPEG-2 patents expired worldwide on 3 January 2024, except Malaysia |
.mjpeg because it has no container to parse.| Where the file is going | Pick | Resolution to use | Why |
|---|---|---|---|
| Video CD authoring | MPEG-1 | 352×240 NTSC or 352×288 PAL | The 1993 VCD standard from Sony, Philips, Matsushita and JVC specifies MPEG-1 video at 1,150 kbit/s |
| DVD authoring | MPEG-2 | 720×480 or 720×576 | DVD-Video peaks at 9.8 Mbit/s of video; authoring software wraps the stream into a disc structure |
| Legacy player or broadcast archive | MPEG-2 | Keep the original | Widest support across old hardware, and now patent-free |
| Anything current | Neither | Keep the original | H.264 in MP4 is smaller and plays on far more devices |
.mpg. No sign-up, no watermark.Not in any way that affects the file. Both name an MPEG program stream, and the short form is a leftover from three-character DOS extensions that a lot of software still expects. Renaming one to the other is harmless. MJPEG to MPEG writes the long spelling and documents the container in more depth.
On typical camera footage, yes — often dramatically, because MJPEG stores every frame in full while MPEG-2 stores mostly the differences, and a fixed camera watching a quiet corridor is almost entirely difference-free. The exception is high resolution: MPEG-2's bitrate floor rises with pixel count, so a short 4K clip can come out larger than its MJPEG source. Reducing Video resolution is the fix.
Different, not better. Per frame, MJPEG is usually cleaner because nothing is predicted from anywhere else — that is why machine vision uses it. MPEG-2 spends its bits far more efficiently across a whole clip, so at any realistic file size a full recording looks better as MPG. Judge it by what you need: one perfect frame, or a whole clip that fits.
MJPEG cuts cleanly at any frame, since every frame decodes on its own. MPEG-2 groups frames and most of them cannot be decoded alone, so editors either cut only at group boundaries or decode and re-encode around each cut. If editing is imminent, keep the original or convert to an editing-friendly target; if the file is finished and just needs to play, MPG is fine.
No. Interframe prediction is the entire reason the file shrinks, so frame independence is what you trade away. If you need both a real container and independent frames, wrap the footage while keeping intra-only coding instead — MJPEG to MKV offers MJPEG in its codec list for exactly that.
Because MPEG-2 has no constant-quality mode; it long predates CRF-style rate control. Quality Preset, Constant Quality and Constraint Quality therefore all resolve to the same resolution-derived minimum bitrate, and switching between them changes nothing measurable. Use Video resolution to move the result, or Constant Bitrate to set the rate yourself.
No. A raw MJPEG stream contains video frames and nothing else, so there is no audio to carry into the container. The Audio Codec list appears because MPEG program streams support audio in general, but with a silent source it has nothing to encode.
You can — MPG to MJPEG runs the reverse — but understand what you get. The MPEG-2 encode already discarded detail permanently, and going back re-expands each frame into a full JPEG without recovering any of it. You end up with a much larger file that looks like the MPG. Keep the original .mjpeg if you may need it.
Files are uploaded over an encrypted connection, converted on our servers, and deleted automatically a few hours later. No sign-up, no watermark, nothing shared or made public.