What File Format To Use For RAW Footage

[N-I-C-K]

I don't have a firewire card on my pc so can't transfer DV from my JVC Minidv camera. So I use my video capture card. I hook my camera via s/video and I've got an adapter to connect the red & white phonoleads to my line-in socket on my sound card.

Now the capture software with came with my tvcard/videocard allows me to select any format (i.e. mpeg,mpeg2,mpeg4,avi-divx/xvid).

At the moment i've set it to capture at PAL DVD MPEG2 variable bit rate of up to 8000. This creates a .mpg.

Then I go onto using Ulead Video Studio Pro to edit the video, add titles e.c.t.. I don't use VideoStudio's build in DVD Authoring functions because I like to use DVD Lab. So I render the final video again as PAL DVD MPEG2.

Now i've got 2 mpeg files.

xxxxx-unedited.mpg
xxxxx-editedfinal.mpg

(well you get the drift.)

I'm just wondering if I'm doing the right thing in capturing the video in the format i'm doing. Would capturing it in Lossless AVI with the PAL DVD Resolution bring me better quality and then render the final edited version as PAL DVD MPEG.

What do you do?

Thanks for any advice.

[Richard_G]

Yes, you would get better quality by capturing as avi, editing in that format and then encoding to mpeg. However, lossless avi creates huge files and requires a very fast hard drive sustained data rate to prevent dropped frames.

You'd be far better off just fitting a Firewire card. They are cheap and, in most cases, just work. Have a look here http://svp.co.uk/products-solo.php?pid=407

[N-I-C-K]

Do video editing software prefer AVI over MPEG..

[edDV]

Do video editing software prefer AVI over MPEG..

Yes, uncompressed AVI or DV (if editor is DV native).

Best would be DV format capture (over IEEE-1394), a DV native editor and do the DV to MPeg conversion at the end. Using this method, unfiltered DV frames reach the encoder as first generation (same as on the camcorder tape). Video data rates never exceed 25Mb/s (~3.8MB/s) so can be handled by most computers.

If DV format capture is for some reason impossible, capture S-Video using the huffyuv lossless codec. The function of huffyuv is to use the computers CPU to effectively lower the data rate to a point where a fast computer hard drive can keep up. Then edit and filter as uncompressed YCbCr and then encode. This path isn't "lossless", the greatest loss comes when DV format CbCr digital components are analog encoded to NTSC (or PAL) for S-Video and then decoded and A/D converted in the capture card. Luminance also goes through D/A followed by A/D.

Software encoding to MPeg2 during capture compromises quality and filtering/editing forces a MPeg decode then re-encode unless cut on I frame. In sum this is a lower quality process. If using a MPeg native camcorder, special editing processes are needed to preserve maximum quality. The typical editor will decompress all Mpeg frames to RGB or YCbCr and then re-encode at the end. MPeg generation loss is substantial.