Benefit to going component out?

[Ken Hodson]

I need to be brought up to speed in this area.
As I understand it, uncompressed capture through component out will help smooth chroma for blue screen work. What are the other benefits, deficiencies?
I have read that some codecs, avid and matrox do chroma smoothing. Does this make componet input void?
Do I loose resolution going through a y/c?
Can this tecnique only be performed live? Meaning once it is on tape it will have DV compression regardless of wether it is captured componet out or not.
If you have some understanding of this, I would sure be glad to hear what you have to say.
Ken

[Ken Tanaka]

Once footage has been recorded on DV tape it's already been compressed.

What camera are you thinking of using for this operation?

[Ken Hodson]

Well there are a few right now- $5000 and under, possibly used.

[Graeme Nattress]

The smoothing of chroma can be done with 3rd party filters - for instance, if you're running FCP you can use my G Nicer filter, which does a great job of intelligently upsampling the chroma. In most NLE's, you can apply a small blur to the chroma channels of the image.

Unless you're going out to a higher format - ie digibeta, then chroma bluring/upsampling will only help on blue/green screen shots.

Capturing via component will, in general, cost you more, use more HD space, and reduce the quality of the video slightly.

Sure, you can digitise the video output of the camera live, but to what? A digibeta deck? Would be cheaper and easier to shoot digibeta in the first place, or DVCPro 50.

The best solution for DV is firewire, with a chroma-upsampler, as necessary for blue or green screen shots.

Graeme

[Ken Hodson]

"Capturing via component will, in general, cost you more, use more HD space, and reduce the quality of the video slightly"

So why do people do it?

"Unless you're going out to a higher format "

Yes I wouldn't consider it if I was going to put it back on a poor format.
Ken

[Ken Hodson]

Can anyone explain the differance as far as capture quality when using:
Analog Component Out - or - Composite Video Out - BNC
I understand Component uses four seperat BNC connections, opossed to one for composite BNC, but how different are they?
Is there a limit to resolution?
Ken

[Jeff Donald]

Resolution isn't necessarily different. But the signal to noise ratio will probably be much higher with component video. A quick visual check is to feed a monitor the component signal and turn on the blue gun and repeat with the composite signal. Much less noise will be observed in the component signal.

Resolution is mostly dependent on the original format. But it may be limited by any individual device in the post production chain (origination to finished media).

[Rob Belics]

Am I not thinking correctly here? Component video has always been the preferred method. With composite you have all the filtering required to insert color burst and bandwidth. I'm talking ntsc here.

Without the filtering, you get the full bandwidth of each rgb component. Yes, eventually it will get filtered in the final end product.

[Martin Garrison]

Component is a better analog transfer than than either s-video or composite. Firewire is an exact copy of what's on your tape.

The reason some people use component out of the camera directly to hard drive; is mainly for chromakey purposes. The downside of DV is it's chroma sampling. If you capture uncompressed from component outs on a camera, none of the color information picked up by the ccd's will be lost. Depending on the camera you may even capture higher resolution than DV is capable.

[Graeme Nattress]

"If you capture uncompressed from component outs on a camera, none of the color information picked up by the ccd's will be lost. Depending on the camera you may even capture higher resolution than DV is capable."

That's only the case if you capture live while you're shooting, not if you capture off the recording on DV tape after the shoot - just thought I'd make sure that's clear.

I still reckon the best solution is firewire capture, then using either a DV aware keyer that does it's own chroma-upsampling, or to use a seperate chroma-upsampler for the particular shots that need it, along with a normal keyer. This keeps everything digital and there are no losses using this method other than render time!

Graeme

[Ken Hodson]

Graeme- What about the 5:1 compression I would be avoiding?
I will most likely be using a GY-DV500/550/5000 which I have been told will offer higher resolution and 4:4:4 colour space if I capture uncompressed.
Are you saying firewire transfer with some form of chroma-upsampling is going to give me the same quality?
Ken

[Graeme Nattress]

"What about the 5:1 compression I would be avoiding?" - you only avoid the compression if you don't record onto DV tape.

"GY-DV500/550/5000 which I have been told will offer higher resolution and 4:4:4 colour space if I capture uncompressed.
" - will only give better quality than DV if you capture directly to a high quality uncompressed codec (most uncompressed codecs are 4:2:2 BTW) while you're shooting. You can record to tape as a backup if you want. Alternatively, hook the camera to a digibeta deck and record onto that, but that's getting silly - you may as well hire a digibeta camera at that point.

If you record onto DV and digitise via component you do not get a better picture - it doesn't magically become uncompressed.

For all practical purposes save tying your camera to your computer's uncompressed digitiser card by a long component lead and recording direct to your HD as you shoot, you will get the best results via a normal firewire transfer with some chroma-upsampling when needed.

[Ken Hodson]

Graeme- " If you record onto DV and digitise via component you do not get a better picture - it doesn't magically become uncompressed."

Yes, that is understood and does not need to be mentioned again as that is not what I am asking.

"For all practical purposes save tying your camera to your computer's uncompressed digitiser card by a long component lead and recording direct to your HD as you shoot, you will get the best results via a normal firewire transfer with some chroma-upsampling when needed"

Are you serious? Can you back this up?
How can 5:1 compression, lower resolution, and 4:1:1 colourspace compete with uncompressed and 4:2:2 colourspace?
Ken

[Graeme Nattress]

Just to make sure I'm getting things straight:

You're shooting with a DV based camera.

No matter how you capture from that camera, wether it's by firewire, component, wether you're captureing live or from the DV tape, the resolution is going to be pretty much identical. One way you're analogue, the other you're DV - they're pretty much going to balance out.

The only way to capture uncompressed from the camera and have it make sense, is to capture live while shooting, thus avoiding the DV tape and firewire entirely, and shooting direct into the computer. To most people, this would be a highly impractical solution to movie making. This has the potential to give the best quality, but it's why I say, "For all practical purposes... ...you will get the best results via a normal firewire transfer with some chroma-upsampling when needed" I don't think the above method is at all practical - you're situation may be different.

"How can 5:1 compression, lower resolution, and 4:1:1 colourspace compete with uncompressed and 4:2:2 colourspace?"

5:1 compression only really has problems when doing effects work - it's easy enough to bump the DV to uncompressed for such work, and you're digital all the way.

Any other issues with the compression for none effects or heavy colour correction work are not going to be helped by going uncompressed unless you're going out to digibeta or some other higher format. Indeed, it may make more sense to bump your DV to DVCPro50 codec and edit with that - it's practically as good as digibeta.

Upsampling the 4:1:1 can and does work. I sell such an upsampler, so I'm biassed, but I have clients who swear by it. In my own tests I found it could look pretty much indistinguisable from 4:2:2.

Ken - what format are you going back out to - what uncompressed capture card are you using?

Graeme

[Glenn Chan]

Are you serious? Can you back this up?
How can 5:1 compression, lower resolution, and 4:1:1 colourspace compete with uncompressed and 4:2:2 colourspace?

What Graeme said: If you record onto DV and digitise via component you do not get a better picture - it doesn't magically become uncompressed.

If you captured uncompressed in the first place (while you were shooting) then you might be able to get some extra quality. However, Graema points out this is silly:

"GY-DV500/550/5000 which I have been told will offer higher resolution and 4:4:4 colour space if I capture uncompressed.
" - will only give better quality than DV if you capture directly to a high quality uncompressed codec (most uncompressed codecs are 4:2:2 BTW) while you're shooting. You can record to tape as a backup if you want. Alternatively, hook the camera to a digibeta deck and record onto that, but that's getting silly - you may as well hire a digibeta camera at that point.

You can get a computer to capture uncompressed on location instead of using a digibeta deck (the computer can be cheaper than the deck). It's still kinda silly though. You can get very good keys from DV already (using the right software).