4:4:4 - 4:2:2 - 8, 10, 12 bit : I need some explination

Spot|DSE wrote on 11/23/2004, 4:12 PM

You can do 8 bit 4:2:2 or 4:4:4, or 4:2:0, or 4:1:0, or any other combinations of compression schemes. 10, 12 bit aren't related to the compression of the vid, ie; the codec. 10 bit simply means it breaks up the image into smaller chunks so it can have greater color values per chunk. In other words, the higher the bitrate, the smaller the sampled chunks are, and therefore the finer they are. In audio, we use 16 bit or 24 bit, and then sample it fairly often to get the most accurate "picture" of how the sound is converted from analog to 1's and 0's.
adamwilt.com has some great answers for this, I'll also plug the " Basics of DV" book from VASST. $20.00 and it covers the whole concept of 4:1:1/4:2:2/4:2:0. Keep in mind, it's not in Andromeda's best interest to show good DV images. And there is such a thing. It's about lighting and exposure.

FrigidNDEditing wrote on 11/23/2004, 4:24 PM

Thanks spot, I've been making the transition from Photography for several years, and you along with the rest of these wonderful Vegas Forum users have been a great help to me. I do know that there is such a thing as a good DV film, I just know that if I can have the ability to capture 4:4:4 or 4:2:2, for a thousand or two more, I am not necessarily going to pass it up.

Thanks again.

Spot|DSE wrote on 11/23/2004, 4:32 PM

There are several hardware capture devices that can capture 4:2:2 and/or 4:4:4. Then you've got the camera to consider too. You need a cam that shoots 4:2:2 as well. The DVX 100 is a 4:1:1 cam, so having a 4:2:2 or better capture card/compression scheme doesn't really benefit you at all. The only other place you need to worry about compression, is when you're discussing filters, transitions, and such. If you can avoid low-range compression of generated media using filters, transitions, and such, then you're better off. The benefit gained by rendering filters, transitions, etc for DV is small overall. Some DVC Pro cams are 4:2:2, analog cams can generally be 4:4:4, and you can take the component out from any analog cam, and a very precious few DV cams, and plug that into a hardware converter that manages uncompressed SD input, and get uncompressed from that. See the www.convergent-designs.com website for an example.
You do get a richer, deeper picture from 4:2:2:, but that's at the camera. Shooting 4:2:2 at the cam, and then downsampling to 4:1:1 is virtually indistinguishable. Yes, it's removing information, but since it was captured at a better compression rate, it spits back out as really sweet video.

FrigidNDEditing wrote on 11/23/2004, 4:40 PM

Ok, so what I understand is that even though they are capturing off of the chips almost directly, because the DVX's chips aren't 4:2:2 or 4:4:4 as they insinuate. (which makes me wonder what, if any, improvement would be gained by this) the footage wouldn't be anywhere near the equivalent of a real 4:2:2.

scissorfighter wrote on 11/23/2004, 5:43 PM

Nonlinear 4, by Michael Rubin, is a good book that explains this well, along with a host of other topics. It does a good job of explaning the historical foundations of film and video from a technical standpoint so that you understand the modern concepts of video editing better. Well worth the read, especially if you're transitioning to nonlinear editing from photo.

http://www.amazon.com/exec/obidos/tg/detail/-/0937404853/qid=1101260382/sr=8-1/ref=pd_ka_1/002-9929483-7061634?v=glance&s=books&n=507846

Ryan

SonyEPM wrote on 11/23/2004, 5:47 PM

If you are shooting with an NTSC DV camera like the DVX100, you are acquiring to an 8 bit 4:1:1 source, and if you are editing in Vegas, you are processing in 4:4:4 RGB...but don't get confused by any of this. Make it look good on the format you are delivering on, make it sound good, and ALL will be cool- you have the tools in hand.

Also: You can broadcast DV/DVCAM, its done all the time, every day, by major networks no less.

Also 2: HDCAM is an 8bit format, Sony XPRI is an 8bit editor, looks great by any measure. Filmout from 8bit HDCAM- its done all the time, Lucas has done it, Robt. Rodriguez has done it. 8 bit source handled right can be used for ANYTHING.

FrigidNDEditing wrote on 11/23/2004, 6:20 PM

I appreciate the attention, and the time you've take to answer a quistion of mine, but I don't know that the answers were entirely answered. Does anyone know if the chip can capture above 4:4:4 before it's sent through the A/D converter. or at least I think that's where they said it was being taken from.

I realize that this isn't panasonic AGDVX100 panasonic forum.

Spot|DSE wrote on 11/23/2004, 7:28 PM

There isn't an "above 4:4:4".
First I realize you're trying to learn as much as you can, but you're somewhere getting numbers confused while trying to run rather than walking. This isn't easy stuff...lots to learn.
Second, there is a LOT, if I could use 72 pt type, I'd use that and say again a LOT of BS and hype about the DVX100 and other 24P cams out there. My take is that it's the rationalization for not shooting on what they really want to shoot on, so a lot of crap gets made up, reworded, assumed, interpololated, and just generally made up about all sorts of cams, processes, etc.
Before you worry about compression, worry about glass. You can also always rent a higher grade cam, but the biggest difference one can make overall, is glass. Then CCDs. Then the storage/compression format. And last, the NLE and related hardware.
Like a bad microphone, if you don't have the glass, you ain't got crap on tape. (or more accurately, you DO have crap on tape with bad glass)

FrigidNDEditing wrote on 11/23/2004, 7:55 PM

I'm sorry, but I mistyped somewhere along the line there, I didn't mean above 4:4:4, I meant above 4:2:0 or 4:1:1. that's what I was trying to find out. I do understand the aspect of the whole glass thing (Photo for about 5-10 years does that). So now that you know what I meant and I sound a little less ignorant (and a little more like I can't operate a keyboard properly) could I ask yet another couple of questions.

Well, first can the DVX produce above 4:1:1 or 4:2:0 at the chip level, and second would you consider a better purchase to be the DVX or XL2 in terms of glass. Not taking into account the super High quality (and still out of my price range) lense that Canon makes for the XL mount.

Thank you again for your guys patience.

Nat wrote on 11/23/2004, 8:00 PM

Considering the timeframe, I wouldn't buy a DV cam now, I would rather invest in HDV, you can still shoot DV with it but in a couple of years people will want HDV over DV...

Spot|DSE wrote on 11/23/2004, 8:15 PM

The DVX100 can produce it, yes. Getting it out of the camera is another thing. You'd have to modify the cam.
I absolutely concur with Nat, buying a DVX right now is sorta silly. If you really, really want one, head over to the DVInfo.net, because there are a few for sale. People are running, not walking to the HDV format.
Any camera can produce 4:4:4 output, but getting a component out is another story. And it doesn't always look good. We shot my Vegas 4 Editing DVD's with the JVC 500 coming off the component output. Didn't look as good as my XL1 with the broadcast lens on it. Because the Alladin capture system used to capture the footage off of the JVC was junk, IMO.
No matter what, you're gonna get a great image off the HDV, so for all the costs you're talking....that's the route I'd go, were it my money being spent.

FrigidNDEditing wrote on 11/23/2004, 8:38 PM

Well Spot, IMO your opinion is probably worth a lot. Perhaps I should just go to HDV. I was interested in the DVX due to the fact that there's one that I can get my hands on now for a measly $25/day if I need to do a shoot, and this would give me two identical cameras. Which would be a dream come true (considering that my life, up until this point has been three different cameras from 3 different companys being mixed to a live screen with different sized chips, and so on, and so on, and so on. This makes me grow very weary of doing live color correction constantly.

Sometimes I grow very weary of it. So that was my main desire for the DVX100.

I guess now I start looking into HD and spend a good bit more money.

Any good suggestions?

Spot|DSE wrote on 11/23/2004, 8:55 PM

FX1 or Z1. There is a good reason I didn't jump on the JVC bandwagon, I've only seen one project done with that cam that I thought was decent, so in my opinion, it's a poor excuse for what has become HDV. In fact, some people heard me rant at NAB about how HDV was gonna suck due to MPEG 2 and other issues. I'm happy to say I was wrong. Now that I've shot it, edited it, experienced it on screens ranging from 62 feet wide to 12 feet wide in my own home, not to mention the several screens I've worked with it on....It is the future, no doubt. This SO MUCH reminds me of when the VX1000 came out, I was using an Ikegami 55 at the time that a friend who worked for the State of Utah had access too. What a very different world.
This brings up a problem though...
2 camera shoots, you can't mix HDV and DV very well, not if you're gonna stay HDV. DV doesn't upsample anywhere near what HDV downsamples to.

FrigidNDEditing wrote on 11/23/2004, 9:12 PM

I just wouldn't use the other, it's not an absolute necesity to use 2, but I wanted it. However right now at least I probably wouldn't do HDV. or I would have to down sample everything. I don't really know too much about HDV, so I'm running even more blind here.

FrigidNDEditing wrote on 11/23/2004, 9:18 PM

Here's the thing, HVR Z1 is the only way I could imagine going in terms of inputs connectors etc...
anyone have a price on it, I don't want to know, I'm sure.

Spot|DSE wrote on 11/23/2004, 9:29 PM

About 5K street price is expected for the Z1U.

vicmilt wrote on 11/25/2004, 4:33 AM

any projected delivery dates, yet??

farss wrote on 11/25/2004, 7:05 AM

If Frigid is worried about color balance then shooting on Hi8 or Imax isn't going to solve the issue. Same goes for exposure, sure having more latitude means you can be more off the mark and pull things back but thats not the way a pro would look at it.
You don't shoot 35mm thinking "Oh well F8 should be close enough, what the heck we can wind it up when we print it". All that extra latitude and resolution comes at an incredible cost and those that can afford it use every iota of it to their best advantage. If that was the right way to look at it how is it that the pros can turn great stuff out no matter what they shoot on?
Sure I'd love to shoot 4:4:4 and love to have more latitude, but not so I can be sloppy, quite the opposite.

Just to answer a few technical points. CCDs are analogue devices basically, the concept of sampling is irrelevant at that part of a camera. The analogue values are sampled at either 4:1:1 or 4:2:2 or whatever during the A->D conversion, the bit depth at that point maybe something between 10 and 14 bits. That then goes into I think DSP chips to process the signal so it can fit into the available bandwith and also for compression. More bits here is good, better processing can give better control and pictures but no matter what it still ends up as 8 bit data, be it DV25 or CineAlta.
There's one or maybe two exceptions in video cameras but that is an enormously expensive path to go down, not least because of the HUGE amounts of data to be stored and then you've got to again process huge amounts of data in post.
All said and done I can think of a lot of things to worry about, not least of all the other 70% of what makes for good video.... the audio!
Lets face it most cameras are pretty horrid in that department and it's something that wouldn't add a lot of cost to the camera to fix, sure you can always use something else to record good audio but why should you have to?
And oh yes, Vegas has been able to edit "Hi Definition Audio" since day one :)
Bob.