dv loss quality

[matriplex]

I captured video from my mini dv with firewire. and now I have dv avi type 2.
if I edit the video on any program (I am using vegas6) (edit:cut, add effect etc.)
and then I encode the video with any dv codec do I lose any video quality?

if I lose how can I have the same quality?

[mats.hogberg]

Yes, you'll lose some quality, as DV isn't lossless, but since the compression is slight, I think you'll have to save out many generations before threr will be any visible difference. After all, DV is made to be editable (and reencodable). Only way to completely avoid quality loss between generations is to save as uncompressed AVI (~5 times the file size of DV) or use some lossless codec like huffyuv.

/Mats

[matriplex]

thanks for reply

disk space is not important. if I encode as uncompressed.avi more than one times do I lose quality?( for example I edit today and encoded as uncompressed.avi and then I import it again tomorrow and edit again and again but I always encode as uncompressed.avi)

[Cornucopia]

You'll not have ANY loss due to multiple encode-decode generations, or to compression, if you encode uncompressed or losslessly (HuffYUV, Lagarith, etc).

You WILL have "loss" whenever you do "processing" beyond simple edits: compositing, resizing, filtering, FX, titles. I put loss in quotes, because even though it may not be modifying the whole picture, and it IS doing what you're wanting, it has to do math on the whole screen and that math has its rounding errors, guessing errors, limited color depth, etc. However, since you're wanting those titles, composites, etc., this type of "loss" is usually considered VERY minor.

Scott

[edDV]

Vegas and other DV native editors minimize losses by maintaining the original camera data as long as possible through the edit. Generation loss only happens in those frames that are processed by filters or effects.

Take simple cuts editing as an example. Since DV contains all frames (unlike MPeg) you can cut and rearrange DV video in any way. If the timeline is processed out to a new DV-AVI file all that is happening is selected frames are copied in the specified order and all resulting video is still first generation just as it was encoded in the camcorder.

If you add a 15 frame transition, those 15 frames need to be decompressed and processed to create the effect. The processed frames are then encoded back to DV format and those second generation frames are used when the output is exported to a new DV-AVI file. In this example all the frames in the editied program would be first generation except those 15 frames which are second generation.

If you add fancy effects or filters, the frames affected will be decompressed for processing and will loose a generation.

As said above, DV format re-encodes well so any losses at second generation are probably difficult to detect. If the project is organized so that all processing is done from first generation DV camera transfers, the quality will be very high. If you process the camera data to second generation, and then use that as a source for another process, then you are dropping another generation and the potential for visible losses increases.

DV format can withstand several decode-encode generations through straight copies. Losses will result from the processing that is done while uncompressed for the effect or filter. Multi-generations of filtering are going to be the cause of visible loss of quality due to math rounding errors and other artifacts.

Decompressing the entire timeline isn't necessary and may cause a generation loss for no reason. Some MPeg2 encoders may decompress everything anyway so that would be a wash. Some MPeg2 encoders will transcode DV to MPeg2 without full decompression of the luminance since the DV luminance (I frames) are already DVD MPeg2 compatible. Motion analysis will be done and interframe compression (P and B frames) calculated first and only if necessary will further intraframe (I frame) compression be done.

PS: If the specified DVD bitrate allows presevation of DV I frames, that could mean that the luminance I frames on the DVD are still first generation camcorder originals.