DV capture resolution question

Hi All,

Can someone please answer what should be a pretty straightforward Q. Outside of the graphics resolution, is there any image quality benefit to capturing DV source footage into an uncompressed timeline. i.e. Kona 8 or 10 bit.

THX

No. Graeme Nattess has studied this and found there is no benefit. And we never argue with Graeme.

Michael Horton
-------------------

John Foley already answered this question in the earlier thread.


www.derekmok.com

Thx Mike.

Hi Derek,

Yup, but the answer was not totally clear...

Best,

M

Let me try again!

DV has a 4:1:1 color space. SD has a 4:2:2 color space. The difference is is kind of like doing a graphic in Photoshop with 72 DPI vs 300 DPI. Color space is the number of pixels that are individually colored differnet to make up a finer texture of image on one frame.

4:1:1 is a sample rate for digital video. It should make sense that 4:1:1 is somewhat less than 4:2:2 when each digit is an oversample number.

So, 4:2:2 is the normal sample rate of carrier:luminance:chrominance. In 4::1:1 there are fewer digital bits of information to work with.

Now that you have captured DV at 4:1:1 (compressed on tape 5:1) you can not improve that content. Placing that content in a 4:2:2 codec will, however allow you to add other graphic pieces in the 4:2:2 color space, so the total output will be somewhat better, if additional graphic elements are introduced on top of the DV content.

Thanks John!

M

John,

The improved color space would make for better color correcting also, yes?

Cameron Young

<<<better color correcting>>>

Yes, I was going to bring that up. There is no absolute quality improvement or benefit of capturing better. It does work out well, however, if you're going to do heavy multiple effects. Aunt Betty at the beach will never look any better than she does now, but she will degrade far less rapidly if you're applying effects onto an uncompressed timeline versus a DV one.

Contrast that with the 5-6 times more computer storage needed for uncompressed.

Koz

Hi Koz,

Would this require a full re-capture into the uncompressed timeline or a copy/paste/render into the new un-compressed timeline, after which we could then add effects & color correction?

THX,

Marc

<<<Would this require a full re-capture into the uncompressed timeline>>>

I'm going to bow to the nitty-gritty experts on this one. It doesn't make the slightest quality difference when you do the conversion, but IMHO, it's trickier to do it after the capture. You have to make sure you bring the raw capture clips into your 8-bit timeline. If you let Final Cut "help you" and start converting or exporting at the wrong time, you could end up with multiple compressions/decompressions which is always a bad idea.

Anybody?

Koz

all you need to do is to copy paste your DV timeline to an Uncompressed one.
FCP will source the DV files, of course, but all renders will be done in the uncompressed settings of the timeline.

Q: do you need to add a "Colour Smoothing" filter? (Natress or Apple)

A: think that;s how it works, but no personal experience.
maybe that;s only for pulling a key???


nick

<<<A: think that;s how it works, but no personal experience.
maybe that;s only for pulling a key??? >>>

Blue or Green Keys are pulled using the color information in the video signal. The colors have been intentionally damaged in DV to make it electrically smaller. Smoothing helps by taking the "gritty noise" out of the color signal. The key doesn't get any more accurate, but it looks a lot better.

If you don't have to color key, I would not smooth anything. Your original show is brittle and easily damaged. Any effects you apply will damage the video slightly. Less is better.

Koz

sidebar question: does FCP do a better job of upconverting than compressor?

I'm just in the process of my son's middle school graduation. About 10 minutes worth of DV. Lighting in the chapel was through a huge stain glass window, and the predominate color was BLUE, of all colors. So, need to do some good CC. I'm running a test clip through compressor up to DV Pro 50 to get to the 4:2:2 space. Visually, the only difference I've seen so far has been better detail in the blacks......damn, I digress!

I could be answering my own question, but, is it better to just run clip selects through compressor and edit, or edit in DV and drop sequence onto new timeline for upconverting?

Cameron Young

have you tried CC'ing the DV shot in a DV timeline?

<<have you tried CC'ing the DV shot in a DV timeline?>>

Oh, yes, that's fine, but with any more complex correcting (ie:selective color), I found that the image gets "blotchy" (for want of a better word) in places, so I want to get up to a 4:4:2 color space and see how that works. Hey, it's only 10 minutes of my son at graduation, so it's not a big deal. Basically, I'm taking Shane Ross's workflow of up-rezzing DV to HD. But, instead just going up to DVC Pro 50 for a better color space. I'm going to take a couple of clips, one CU of face, and one with movement to compare in the different spaces with CC'ing.

Cameron Young

Just to cover the bases, you understand when you "upconvert" from your DV quality color to uncompressed quality color (4:1:1 -> 4:2:2), you don't make your original show sharper, you just make it possible to see and manage the original fuzzy color much more accurately.

While you're inside the 4:2:2 colorspace, anything you add or change will be with new sharper colors.


As a gross oversimplification, 4:2:2 tells you the quality of the black and white, red, and blue. Picture your wife's skirt in black and white. If you wanted to give her a red skirt in 4:2:2, you could color in the skirt with a box of sharpened crayons. In 4:1:1, we hand you cans of spray paint.

There is such a thing as 4:4:4. Productions use this for advanced effects. In that case, there is no difference between the color and black and white information. They all have the same extremely high quality and sharpness.

Blue or green keying from a DV quality show means you're trying to extract fine detailed outlines and accurate edges from a spray painted picture. Yes, it's possible to do that, but it's a lot of work and can be expensive compared to doing the same job in a higher quality format in the first place.

Koz

<<Picture your wife's skirt in black and white.....we hand you cans of spray paint>>

What a great analogy, Koz, pretty much answers THAT question.

Thanks!

Cameron Young

Logging and capturing from a video shoot using fcp then sending producer home with a copy of captured files on a separate drive to make edit decisions. Can the captured footage be replayed using fcxpress?

The captured footage, unless modified in some way, will just be quicktime movies, so should be fine. This is a seperate thread, by the way. You should start a new thread if you have a new question, so more people notice it and can help/benefit from it.