What is MJPEG

[Karl Beem]

and how does it differ from MPEG?

Karl

[fluggo99]

MJPEG is short for Motion JPEG. It's essentially a lot of JPEG's lined up in a row. JPEG's are useful to encode pictures because they remove some nonessential details from a picture to save disk space.

MPEG is similar, but takes it a step further. MPEG watches to see what changes between frames, instead of posting a complete picture of every frame, like MJPEG. So, the first frame might be a full picture, then the second only carrying the differences from the first, and so on.

MPEG and MJPEG have their ups and downs, and are used very differently. MPEG takes a long time to encode, but results in smaller files, so it's typically used for final products such as VCD's and DVD's. It's also useful on the web. MJPEG creates larger files, but is fast and keeps separate frames, which helps immensely in editing.

[Karl Beem]

Can sound be gotten from MJPEG files? I ask because I've only heard of MJPEG in relation to video cards.

[Michel Carleer]

Karl,
When we talk about MJPEG, we talk about an AVI file which is made of a video and a sound stream. Video is compressed in MJPEG. Sound can be compressed too but usually one simply store sound as uncompressed PCM. Of course, one can always choose to capture only video and no sound but usually an AVI file comprises both.
Michka

[DrMordrid]

One other difference is that MJPeg (and DV) is much easier to edit than MPEG1 or MPEG-2 and gives higher quality in the final product.

DV and MJPeg consist entirely of I frames; data that is actually captured frame-by-frame.

Most all MPEG-1 and MPEG-2 files use P and B frames as well, frames created by interpolating data from I frames that can be many frames apart. This gives MPEG files a high compression and small file size, but makes editing problematic.

Unless the editing software converts all the P and B frames to I frames internally you can't do frame accurate cuts. Even if the I frame conversion is done you can still get artifacting if you add effects since each effect causes one more generation of data re-compression. Re-compression in a format as highly compressed as MPEG's are is the kiss of death quality wise.

IMHO it's better to capture and edit in I frame (DV or MJPEG) then convert to MPEG-1 or MPEG-2 for distribution only.

Dr. Mordrid