Yes, Firewire capture is lossless. The DV stream comes in over firewire, a direct copy of the data on tape, and gets put in a Quicktime wrapper so that any application can see it and use it. If you were to play that quicktime back out to tape, it gets unwrapped and the same digital data goes back out to tape again.
...but remember, if you alter the clip in any way - crop, colour correct, add titles - it will have to be recompressed by FCP before going back to tape. As DV is 5:1 compression, you will loose a little quality there. It doesn't mean that th footage will now be compressed 25:1 but that it has been compressed by 5:1 twice, picking up digital compression artifacts on each pass.
I tend to find that one recompression is acceptable with DV, but any more and it starts looking ropey, IMO.
That's right to an extent, but it all depends on what you do to the image. If you change it drastically, it can compress and look great because it's just like a new image going through the compression. A subtle change might look worse in that respect. If you scroll the image and it's not by a multiple of 8 pixels, you're doing some of the worst you can do. If you just crop, the recompression might quantise back to the same values and not degrade the image much at all.
I agree with Graeme, because he is better at this stuff that me!
I was interested what you said about scrolling and 8 pixels Graeme, why is that?
I certainly colour grade all of my footage because I think it looks MUCH better that way, even with the extra compression step.
I was merely saying you should be aware of recompression. If you output through compressor to MPEG2 DVD, I THINK FCP compresses straight from your raw original footage so it will be the best quality humanly, well computerly, possible. Please correct if I am wrong.
That's because of the size of the macroblocks that the DCT compression uses. If you don't move the image and recompress it, that does the least harm as the DCT coeficients, the ones that get quantised are still with their own blocks, and their lack of precision (the bit that does the compression) may just get them sent back to their original values on recompression. However, if they move into new macroblocks, then there is no way that they can get back to their old values as everything gets recalculated differently, and greater losses occur.
Good colour grading more than makes up for any compression IMHO.
If you export direct to compressor from FCP, it will re-render all your effects to uncompressed, and send that info to compressor. I've not seen it make much difference though, other than more crashes and slower MPEG2 compression. If you want that kind of benefit without the problems, do the final render to uncompressed, save out an uncompressed movie and go to MPEG2 from there.
I don't mind the wait for compressor after spending so long on post-production. I leave it over night normally. I thought that was the case. Do you think that a DVD done this way looks better than DV to tape master? ie is MPEG2 GOP better than intraframe, higher bitrate DV?