[Yuri Reznichenko]

Hi there.

The problem is, as far as I learned, is in how quicktime treats sizes in mpeg-2.

For example, if you make a mpeg-2 from a 720×576 (rectangular) PAL movie, you’ll see 720×540 square in both quicktime’s and final cut’s properties of the mpeg-2 file (the size is quite odd because pal is usually made square by stretching the longer side to 768).

if you make an 16×9 SD PAL movie (720×576, 1.3 aspect), you’ll get 720×404 (again, square).

when you import mpeg-2 files into fcp sequence, final cut pro looks at the dimensions quicktime provides – and automatically scales and distorts the movie back to normal size again, resulting in severe visible quality loss (worst of all is when your footage is interlaced).

What I usually did is I opened mpeg-2 in mpeg streamclip (which treats everything the way it should – right size, rect pixels) and exported movies to a low-compression quicktime format which works natively in fcp (mpeg-2 footage is very rare so that workflow was o.k. for me).

Now I did a little research with a movie consisting of alternating black and white lines 1px wide, both horisontal and vertical. After conversion to mpeg-2 and looking at the distorted array of horisontal lines I went to A/V controls (+K) and manually rewrote the dimensions to what they should be. The image returned to normal.

Quicktime won’t allow you to save the changes back to the mpeg-2 file, but you can save as .mov (either reference or self-contained with the original mpeg-2 stream inside – no recompression). this is easily imported back to final cut and no distortion or scaling happens (you’ll still have to render everything, fcp doesn’t work natively in mpeg-2).

while this would be ok for one or several files on occasion, this is not a good workaround if you have several hundreds of them. in this case you might think about a different output format on your other system.
and, again, mpeg-2 is a delivery format, it’s not meant for editing.