As already pointed out, unchanged footage will go back to the recorder unchanged. But.
Be aware that the signal recorded to the heads *is* actually an analog signal and as such is susceptible to some degradation given enough errors. Real world tests done a few years ago showed the cameras then to be able to go about 5 generations before _any_ loss was noticeable. At that point the material quickly became garbage.

So, yes edited DV is basically first generation material identical in all respects to what came out of the camera. Unless, for some odd reason, you try to go down several generations.

Thanks for the replies. I'm glad to see the functionality is in Vegas. I was mostly just curious - I'm not too concern with the quality loss since compressed DV is still pretty high quality. I don't write back to tape, so that is not a concern either.

Right now, I'm doing all my editing in Vegas, saving the resulting 2 hours to compressed DV, then using TMpgEnc to make DVD compliant MPEG2. I was hoping to just use Vegas' MainConcept encoder and save myself the need to save out to compressed DV first, but the quality I'm getting with TMpgEnc is significantly better then with MainConcept's encoder. Compared to TMpgEnc, the MainConcept encoded videos are washed out and lack contrast. I'm using the highest quality settings in both programs and the same bitrates. My source is 35 year-old 8mm family videos, transferred to VHS 10 years ago, then transferred to my system via a DV capture. Surprisingly, the quality is not that bad, but I imagine it is hard on an MPEG encoder since there is jerkiness and brightness/color changes between each frame of the video. I’ll give the MainConcept encoder another chance when I start doing cleaner videos.

Mmm, I noticed that lack of vibrancy with the MainConcept encoder. And, like you, I prefer the Tmpgenc encoder - I think it's a bit quicker too.
It's a shame because it means an extra step (rendering an AVI from VV3), which you might expect would bring in more opportunities for degradation.
I've also tried the LSX MPEG 3.5 encoder. What a pain that is! For a start, it doesn't recognise the AVIs produced by VV3 (or at least VV3's own codec) - I had to re-render using the 'Microsoft Video 1' option before it would deign to look at it. It doesn't adapt the audio either, so if your original AVI has 48KHz audio and you're wanting to make an SVCD, you have to re-render the AVI so that it has 44.1KHz audio before presenting it to the Ligos encoder.

I wonder when we'll get a really good MPEG encoder? With the growing trend for home DVD production there will be an increased incentive for developers to deliver the goods. At the moment, I think we're ill served.

-Pete

'Be aware that the signal recorded to the heads *is* actually an analog signal and as such is susceptible to some degradation'

Am I reading this correctly? Would it be accurate to say that the info laid down on the tape in analog fashion is all numeric, and, as such, is either readable or non-readable? Is not the advantage of digital information that, when it is there, it's all there?

Is not the digital info we pass from cam to computer and back again the same as copying any other sort of computer file such as a spreadsheet or word processing document?

I'm not trying to be contrary, but this was always my impression of the justification for digital formats.

I have also assumed that the degradation wreaked by VV (or other editors) at rendered points occurs once only . . . i.e. in order to see generational degration, one would have to render some FX upon frame sequence 'x', print that to tape, recapture, and then apply FX to sequence 'x' again, render and print, recapture, etc.

Am I correct in this assumption?

Would it also be correct to assume that one could capture raw footage, perform one's editing, apply FX, render a finished product, output that to tape for archival purposes, and rely upon recapturing to get a 'perfect' digital copy of that previously rendered project which could then, if unmodified, be output back to tape to create a second replica of the archived project?

From what I have read previously, this was always my assumption about digital video. What I read in this thread appears to be at odds with those previous assumptions of mine.

What say ye all?

Caruso

I do find that the Main Concept encoder to be very close to the TMpgEnc for my needs and the problem of washed out colors and contrast is being address by SonicFoundry by their next update as I understand it.

from SonicEPM 1/30
Some of you have noticed washed out color when converting DV to MPEG-2 or streaming formats. You might try adding the "Levels" filter to the entire project, settings:

input start: 0.063
input end: 0.922
output start: 0
output end: 1
gamma: 1

With an anecdote like 30 generations without loss.. and the fact that I've never ever recall anytime watching any material recorded in DV format showing any artifacting.. would we all agree that DV is magnitudes better than MPEG-2 as codecs go? I cringe all the time when watching DVDs when macroblocks pop up left and right. Just wondering if there was a consensus on this opinion..

>Would it be accurate to say that the info laid down on the tape in analog fashion is all numeric, and, as such, is either readable or non-readable? Is not the advantage of digital information that, when it is there, it's all there?


No, the nature of mag tape recording is such that there will always be dropouts. DV machines are very good at data recovery, but if it reaches a certain point you get total garbage. Unlike say, VHS where you can still see a picture behind all the noise.

>degradation wreaked by VV (or other editors) at rendered points occurs once only

Right. As long as it's not rendered it hasn't dropped from what was captured.

>Would it also be correct to assume ... capture,write,capture,...

Remember each time the signal passes the heads of the recorder it's being converted to analog, so there may be some loss. Again, until it gets to be a lot you won't see it and then you suddenly won't have anything at all.

Rarely a factor in real world DV usage, but it prevents us from saying DV is 100% lossless. DV copies are certainly 1000% better than VHS copies, but some people take great delight in pointing out this 'failing' of so-called 'digital' video. Of course, by this argument you could only achieve true digital with punched paper tape!

In my mind, drop outs and degradation are separate issues. If you have a dropout on a floppy disk, then, your excel spreadsheet file will be corrupted, or, perhaps error checking of some sort will kick in so that the dropped data is re-written to a good spot on the disk.

I understand about dropouts.

My question has more to do with generational quality. Assuming the media doesn't fail physically, is it correct to expect that data can be passed back and forth between the cam and the computer without loss in situations where no additional rendering is required?

Thanks again.

Caruso

has anyone done a side-by-side comparison on tmpgenc and the VV3 mainconcept mpeg1 codec for standard VCD ...

take a look at my thread here:

http://www.sonicfoundry.com/forums/ShowMessage.asp?MessageID=90584&Page=1

can anyone else do a side-by-side test and let us know if they get similar results??