Which is better Method of capture from DVCAM?(help settle a debate)

I've got a guy telling me that the Sony DVCAM deck does a better job of uncompressing than FCP can do in my Mac.

So he is telling me that it is better to go Component out of the DVCAM deck into the capture card, capturing it as uncompressed video. He says we get 4:2:2 with this method vs 4:1:1 when captured through firewire.

He also says we are way better off working with uncompressed even though we rarely have more than 4 video layers, 2 streams of video, minor amounts of transitions/dissolves and only do basic image correction.

I've had other people tell me that capturing from DVCAM through firewire with our capture settings at DV NTSC is the better route.

We finish right back to DVCAM

Your friend is sort of right. I'd think you'd be better off capturing through a card in ProRes, keeping a ProRes master, and going back through the card to component if you can to record back in DV.

All the best,

Tom

Depends on your workflow. If you are making straightcuts only, then uncompressed or pro res won't give you much. If you pulling keys and other heavy graphics environment then Pro Res is the way to go. Uncompressed is an old workflow these days (imho).

CHL

Chi-Ho Lee
Film & Television Editor
Apple Certified Final Cut Pro Instructor

We are currently on an ancient system, but we will be upgrading when the MP line is updated. Looking forward, if our source is DVCAM and we finish to DVCAM what do you recommend our capture/sequence settings be? We have Sony DVCAM decks and are looking at the LHi card. We use static graphics and rarely more than 2 video streams at any given moment. Image correction is basic as well. We are looking at either component capture or FW capture, Uncompressed NTSC or not really sure what codec

Thanks, T

Quote

I've got a guy telling me that the Sony DVCAM deck does a better job of uncompressing than FCP can do in my Mac.

Not really true. DV is not a new and mysterious technology. It's an ancient format, as these things go. Software decoding of DV works just fine.

Quote

He also says we are way better off working with uncompressed even though we rarely have more than 4 video layers, 2 streams of video, minor amounts of transitions/dissolves and only do basic image correction.

Technically true, but ? meh. When you edit DV, it's not a lossy process, because DV is an I-frame-only format. If you cut two DV shots together and export them, all the computer does is copy the frames from the first shot, then the frames from the second shot. Nothing gets decoded or re-encoded in the process.

When you do stuff to DV, however, you're compressing it again. Because the frames have to be uncompressed, stuff-done-to, then recompressed. That's two generations of DV compression, which isn't great.

But here's the thing: Assuming you do your workflow right, those two generations are all you're going to have to deal with. And you'd have those same two generations regardless, because (as you said) you're going back out to DV tape at the end. DV in, stuff-done-to, DV out. It doesn't matter what order you do those operations; they're the same operations.

Now. Does it really not matter? No. There are mathematical differences. DV is an 8-bit integer format, so you can get round-off errors if you do color correction or compositing with it. But honestly, you're not doing a lot of color-correction or compositing with DV in the first place, because the format's simply not capable of much in that arena. It falls apart. So while you can try working in uncompressed, I'm willing to bet actual money that it'd be very difficult, or even impossible, to spot the difference after you've gone back out to DV tape again.

So it doesn't matter, even though it really does matter, because really, it doesn't really matter.

I think.

>he is telling me that it is better to go Component out of the DVCAM deck into the capture card,
>capturing it as uncompressed video. He says we get 4:2:2 with this method vs 4:1:1 when
>captured through firewire.

DV is chroma 4:1:1 subsampled on tape (4:2:0 for PAL). You will never get a true 4:2:2 image because the picture has already been decimated on recording. However, when you play the footage out through a component or SDI output on the deck, the deck will chroma smooth the image and play it out as a 4:2:2 signal, so the picture can look perceptually better.


Graeme Nattress has an article on it:

[www.nattress.com]



www.strypesinpost.com

This is of course only on a live recording through SDI. The recorded tape is already 4:1:1 and compromised.

All the best,

Tom