Please consider also that recompression only occurs in Vegas for stuff you've changed (added filters, effects, transitions etc). All untouched frames are merely copied from source to target files without recompression. So, for the most part, the initial compression that takes place in the camera is the only compression that occurs until you employ edits made up of anything other than cuts.

Of course, if your target medium is DVD, then MPEG-2 compression is required for all frames.

I believe that the in-camera compression is a derivative of MPEG-2, but only I-frames. Hopefully someone can elucidate.

Go here to get more technical details about DV.

As mentioned, DV uses a mild form of DCT compression to reduce the video data by about 5 times. The resulting video stream is about 25Mb/sec. Adding audio brings the total bitrate to about 29Mb/sec.

Compression is done only within each frame, so the frames are completely independent of each other. Artifacts from compression are almost completely invisible, and DV video is ranked as being at least as good as BetaCamSP, which is the format your local TV station has used for years for newsgathering.

Technology marches on, however, and the latest compression format, HDV, manages to put a high-definition video signal onto conventional miniDV cassettes at the same rate as regular-definition DV. It uses a variation of MPEG2.

Even that format is not the leading edge. Microsoft's WMV9 and MPEG4(264) improve on MPEG2 by about a factor of 2. Amazing.

No the in-camera compression is pretty much the same as what Vegas uses, I mean precisely the same format, that's why there is no alteration of the data during capture or PTT. So if you capture into Vegas and just PTT without any editing you've just made a copy of the tape.
The DV25 codec bears almost relation to mpeg encoding, mpeg-2 involves spatial and temporal compression, DV uses only temporal compression, very similar to jpeg.

Sorry Farss - which bit are you saying "no" to?

"It is fairly similar to MPEG compression but is made up entirely of I-frames"

Taken from http://www.dvcollections.com/support_dvcompress.html

Yes,
fair enough statement from what I know. The significant bit being though that it's entirely I frames which means there's no temporal compression. Although mpeg can be composed entirely of I frames it is an unusual way to encode as it doesn't achieve any significant compression.
The important difference is that the frames of DV are independant, with normal mpeg-2 encoding on say a DVD the frames are not independant which makes it difficult to edit.

AND just to restate the basics
dv is a lossy compression, done in camera @ 4:1:1
imported via firewire into Vegas is a direct transfer of data w no loss
when edited in Vegas and , print to tape, ntsc dv output,
it is output at the same w 0 loss from the original.