Up-Rez Question: miniDV to DigiBeta...

I have two questions that I can't seem to find the
answer to:

1. I have a feature project in FCP that was shot on
miniDV and needs to get to Digital Betacam. My
question is how do we successfully export/prepare the
footage before layoff to up-rez it to the proper
quality? Any info online about the matter was very
vague and refered to upping the chroma or exporting
tiffs (which would be tedious for a feature). How can
we upres this footage in FCP or After Effects?

2. I have to also export a mov or avi file to specific
specs listed below, but when I select export quicktime
conversion and set the codec to uncompressed all other
options disappear. How do I make sure I'm meeting
these specifications? And would making this file
essentially upres the miniDV footage to 4:2:2 and
solve the above problem?


Name: 10-bit Uncompressed NTSC
Description: Uncompressed 10-bit NTSC Video with audio
pass-through
File Extension: mov or avi
Audio: multi-track passthrough
Video Encoder
Format: QT or AVI
Width: 720
Height: 486
Pixel aspect ratio: NTSC CCIR 601/DV
Crop: None
Frame rate: 29.97
Frame Controls: Automatic
Codec Type: Uncompressed 10-bit 4:2:2
Multi-pass: Off, frame reorder: Off
Pixel depth: 24
Spatial quality: 100
Min. Spatial quality: 50
Temporal quality: 0
Min. temporal quality: 50

First, assuming your whole show is in NTSC, you're not uprezzing anything. The line structure and number of pixels is the same between the two standards. There is that 6 scan line thing between the two, but that just ends up black in DBeta.

You need a DBeta machine connection to your Mac and press record on the tape machine and play on the Mac. You can't do this with straight FireWire. You need actual extra hardware that can talk SDI picture, AES/EBU sound, and RS422 machine control so the tape machine knows what's going on.

You will never get to 4:2:2. Your show, because of the way it was shot, has very much reduced color sharpness compared to straight DigiBeta. DV is either 4:1:1 or 4:1:0 or one of those. Reduced quality color. That may be what the people who wrote the specs were getting at. They didn't want anybody shooting DV and trying to pass it off as high quality video.

More later.

Koz

--Name: 10-bit Uncompressed NTSC
--Description: Uncompressed 10-bit NTSC Video with audio
--pass-through
--File Extension: mov or avi
--Audio: multi-track passthrough

That means native sound which is 48 KHz, 16 bit, Stereo, unless you tried to produce 5.1 surround. Then, you're on your own.

--Video Encoder
--Format: QT or AVI
--Width: 720
--Height: 486

Your show right now is 720:480. You're missing those six magic scan lines. The only way I know to get around this is to create an uncompressed 10-bit timeline and render your show to it. You may not like the results. Final Cut will have to completely take the show apart and put it back together again. It will likely get fuzzy and it will certainly take up enormously more room. Uncompressed goes by at about 1M per **frame**.


--Pixel aspect ratio: NTSC CCIR 601/DV
--Crop: None
--Frame rate: 29.97
--Frame Controls: Automatic
--Codec Type: Uncompressed 10-bit 4:2:2
--Multi-pass: Off, frame reorder: Off
--Pixel depth: 24
--Spatial quality: 100
--Min. Spatial quality: 50
--Temporal quality: 0
--Min. temporal quality: 50


I suspect some of this was written by someone guessing at it. A lot of those numbers are only significant in MPEG2 compression.

Koz

I was under the impression that laying off straight to DigiBeta would cause a noticeable quality loss due to the lack of extra chroma info- much like increasing the scale of a clip which causes blurriness and pixelation.

Some other forums (http://dvinfo.net/conf/archive/index.php/t-29557.html) posted sloppy vague instructions on upping the chroma using Shake or After Effects.

i know there's a progra called Instant HD by Red Giant, but is there a way to do it in FCP or AE? Will simply exporting my sequence as Quicktime conversion using the Uncompressed 10 bit 4:2:2 do the conversion* for me?

*By conversion I mean basically what Instant HD does- calculating what the missing chroma info should be and adding it in.

Maybe this will help explain what I'm trying to do: [www.redgiantsoftware.com]

How come when I go to export my timeline as quicktime converion, when I open the size options and its set to current it says 720x486? I noticed this happened when I upgraded FCP a while ago... any connection?

And yes I assume they're guessing, their final output will be things like DVD, VOD, TV etc...

Personally, I don't see the point in up-rezzing miniDV only to bring it back down to DVD... it just seems like more conversions means more potential of quality loss. but i could be wrong. I guess the distrbutor is used to buying higher quality projects that originated on HD or 35mm.



> Your show right now is 720:480. You're missing
> those six magic scan lines. The only way I know
> to get around this is to create an uncompressed
> 10-bit timeline and render your show to it. You
> may not like the results. Final Cut will have to
> completely take the show apart and put it back
> together again. It will likely get fuzzy and it
> will certainly take up enormously more room.
> Uncompressed goes by at about 1M per **frame**.

<<<Personally, I don't see the point in up-rezzing miniDV only to bring it back down to DVD... it just seems like more conversions means more potential of quality loss.>>>

I read their spec sheet as their trying **really hard** to make sure you *don't* shoot in DV, but rather some better quality format. I can lay off a VHS tape to those specifications, but it won't look very good. You can always force whatever you have to conform to those specs, but had you shot that way to begin with, there would be no conform or conversion. It would automatically fit.

If you have an uncompressed timeline, either 8-bit or 10-bit, it will say 720:486. It assumes "perfect" 601 video by your settings, and perfect video doesn't have a DV frame size. DV comes pre-damaged by all the extra compression and stripping away things that you don't absolutely need to view the picture.


I would not have assumed that a DigiBeta would look funny if you did a straight layoff from a machine fitted with appropriate hardware. All the hardware we have "knows" how to fill in the missing pieces no matter which way you're going. Yes, DV has half-sharp color both directions and it really doesn't matter how many bits you use carry it.

It's possible to get software packages to "simulate" better color, but that 'making video where there didn't used to be any' business can be tricky or unstable. I would try the layoff first without all the additional fuss.

Koz

i usually just export to DVCam, then get that bumped up to Digibeta.

the main thing with the DV colour space is that because of it;s limitations you dont want to do your colour correction there.
then again, while it;s possible, you don't really want to do the colour correction on FCP anyway.

your best bet is to offlay your DV with as few fx as possible to digibeta.
then do the CC in a decent suite.

it all depends on the scale of your project of course.
i've graded and finished shows in FCP on DV.


nick

I have also gone DVCAM to Digibeta via hardware and the quality is fine.

There is a guide to upressing using compressor here : [www.proapptips.com]

This one deals with Dv to DVCPRO HD, and just as a heads up, it takes a LONG time to do upressing. I recently did a test with 17 seconds of vision and it took more than an hour to finish. Results were very good, but I'd much rather go real-time with a hardware transfer if possible.

For a 3-screen museum piece this fall I color-corrected sources shot on MiniDV, wich were recaptured 10-bit uncompressed from a DSR-1800 via SDI using Kona 2. It was then output to DVCPro 50 QuickTime for display on large LCD's, but could have gone to any format.

Reslution doesn't change of course, but color correction/enhancement worked great and the exhibit was a success.

- Loren
Today's FCP 4.5 / 5 keytip:
Toggle Audio Scrub with Shift-S !

The amazing FCP KeyGuide?: your power placemat.
Now available at KeyGuide Central.
www.neotrondesign.com

Quote
nick
your best bet is to offlay your DV with as few fx as possible to digibeta.
then do the CC in a decent suite.

wow, i could have swore that most networks are converting to fcp. i even remember a story about the bbc is now all apple. sundance channel uses apple fcp and if i am not mistaken so does fox news.

so what do you mean by a decent suite?

""" What you do with what you have, is more important than what you could do, with what you don't have."

> > > Knowledge + Action = Wisdom - J. Corbett 1992 """"

a decent colour correction suite.
DaVinci etc.

Thanks! This video was extemely helpful!


> There is a guide to upressing using compressor
> here :
> [www.proapptips.com].
> mov

So what happens to the footage when you use the quicktime conversion export and set your codec as Uncompressed 10 bit 4:2:2? That doesn't upconvert your footage and calculate and fill in the missing chroma info?

It seems like this video says you can make your quality better:

[www.proapptips.com]

As well as the supposed use of Instant HD v1.0 by Red Giant... but I'm not sure that' s why I'm asking.

[www.redgiantsoftware.com]



> You will never get to 4:2:2. Your show, because
> of the way it was shot, has very much reduced
> color sharpness compared to straight DigiBeta.
> DV is either 4:1:1 or 4:1:0 or one of those.
> Reduced quality color. That may be what the people
> who wrote the specs were getting at. They didn't
> want anybody shooting DV and trying to pass it off
> as high quality video.
>
> More later.
>
> Koz

Most of what you are doing with these upconverts is changing the dimensions of the picture. Think of taking a thumbnail picture off the internet and trying to use it to fill a whole webpage. You'd need to smooth off a lot of nastiness in the process.

I'm no expert on this, and possibly some chroma smoothing is going on, but you can see in Shane's tutorial that the colour does not really improve. There's no way for the software to enhance colour that does not exist in the first place.

Like Nick says, the most professional result will happen by going to digibeta as cleanly as possible from your source and then doing colour work in a dedicated colour correction suite.

Quote
Jude
Like Nick says, the most professional result will happen by going to digibeta as cleanly as possible from your source and then doing colour work in a dedicated colour correction suite.

is this to say that a dedicated color correction suite can some what up rez dv footage to look better?

if so what are some color correction suites

""" What you do with what you have, is more important than what you could do, with what you don't have."

> > > Knowledge + Action = Wisdom - J. Corbett 1992 """"

Well, DaVinci is probably the most lauded. You can find out more about their various versions here:

[216.79.18.60]

And here's a picture of one of their 2K systems



Keep in mind that these are not cheap and they need very specialised knowledge to run them. Instead of trying to force your mini dv footage into something it's not, you could probably save a lot of time and money by shooting with a better camera in the first place.

I don't mean that in a nasty way. It sounds kind of nasty when I read it. What I mean is, it's a more sensible approach to start with better quality than trying to fake it later.

yep, i dont even have to ask or surf to find out what it cost. i am sure that system is more than the cost of a hvx200. dont even need a price to know that:-)

""" What you do with what you have, is more important than what you could do, with what you don't have."

> > > Knowledge + Action = Wisdom - J. Corbett 1992 """"

"is this to say that a dedicated color correction suite can some what up rez dv footage to look better?"

just to be clear on this:
you are NOT up-rezing your image going from DV to Digibeta in terms of image size.
the resolution of the image is the same.
however, you are moving into a colour space with more information.
so some shots may look better.
(for instance i off-lined a rock concert in DV, and then on-lined from the original Digibeta.s
the big blocks of reds and blues looked way better on digibeta, and this was without any colour correction at all)
a lot of shots you wont even see the difference.
and if your shown is black and white there's no difference at all. it;s only the colour space that has more info.
(ok, i'll admit this is an assumption on my part. it stands to reason, but if someone wishes to correct me, please do)

the MAIN benefit of going to a dedicated cc suite is that you'll be having the work done by a pro.

you haven't said much about your project.
what form it;s in, (have you graded it already in FCP?)
& why you have to go to digibeta.

not much has been said about your second question.
i think that really all those settings add up to be a simple export to 10bit uncompressed.
you'll get a .mov out of fcp a lot easier than a .avi

nick

Hey, Nick, I think you're right about:

"The MAIN benefit of going to a dedicated cc suite is that you'll be having the work done by a pro.?"

There are secrets which pros have that you won't hear openly. I noticed at post production facilities the technicians know how to correct a scene that's a little tweaked (for lack of a better word :-)

<<<is this to say that a dedicated color correction suite can some what up rez dv footage to look better?>>>

Nice try. When you take a picture of a tangerine in DV, the outline of the fruit is full resolution and sharpness, but the orange color itself is carried by a can of Krylon spray paint. No matter what you do later, the tangerine orange color will never get sharper than the signatures from a freeway overpass.

What will improve in a higher quality format is the ability to do post (color correctons). If you need to create any objects or do color surgery, that's much better and "invisible" if you do it in 4:2:2 rather in DV quality. But if you need to change that tangerine, you're going to be stuck with the original fuzzy color.

As has been pointed out in this thread, you can sometimes make the color worse and it will look better. This is the smoothing and interpolating process. It's never more than a patch job, tho, and it may get you through until you can get a better camera for the next shoot.

Koz

"There are secrets which pros have that you won't hear openly. I noticed at post production facilities the technicians know how to correct a scene that's a little tweaked (for lack of a better word :-)"

uh, yeah.. that's their job.
not so much secrets (most of them are more than wiling to tel you what they're doing) as experience.
but maybe that's the real secret

<<<most of them are more than wiling to tel you what they're doing)>>>

Totally. However they tend not to do that as a matter of course because of the Blank Look Problem.

"I'm going to take the green gamma up a touch to get rid of that magenta cast in this scene. Then we'll slide the soft clipper on red a bit to make the sunset look better."


What.

Koz

And voila the picture looks better :-) You're right, Koz. I've seen it. I've been told by Post houses, "bring the edited movie in and we'll put it through our ______, you'll like the results. They even have dust busters :-)

I don't know about ghost busters though. LOL After over-doing the color corrections, sometimes the picture looks fuzzier. I think the pros know how to use the controls better -- and, yes, from experience (doing it everyday for years).

I like to edit my movie and take it in to one of these facilities and ask them if they can give it that Gone With the Wind look, you know what I mean? The Color by Deluxe look :-)

i am just following this thread to know more on color correction. i just have a simple inquiry about somthing that mike said.

is 4:2:2 hd quality or is it just under hd. i am thinking it goes:

best - hd
better - 4:2:2
good - dv / sd
poor - ?

i know i am probably wrong

""" What you do with what you have, is more important than what you could do, with what you don't have."

> > > Knowledge + Action = Wisdom - J. Corbett 1992 """"

> is 4:2:2 hd quality or is it just under hd. i am thinking it goes:
> best - hd
> better - 4:2:2
> good - dv / sd
> poor - ?

Hmm, that's kind of like asking "Are oranges better than fruit?"

The confusion probably came of Koz' statement of " if you do it in 4:2:2 rather in DV quality". What he meant was "doing it in a 4:2:2 format rather than a 4:1:1 format like DV". 4:2:2 is a chroma-sampling value, not a format. For example, one of the HD formats, DVCPro HD, has 4:2:2 chroma sampling. Another, HDCam (not SR), has 3:1:1. Yet another, HDV, has 4:2:0.
However, one SD format, DigiBeta, has 4:2:2 chroma sampling, while DV is 4:1:1.

There's some good information here:

[www.kenstone.net]

So you're talking about a bunch of different standards. Better chroma sampling isn't the only factor in a "better" image -- you also have to take into account pixel count (frame size). In this particular case, though, pixel count is a moot point because the footage has already been shot, and there's no point in trying to represent a small frame size (720x480 DV NTSC) using a large frame (eg. 1920x1080 HDCam) because you won't gain any more information on the image. That had to have happened at the shooting stage, as Koz mentioned.


www.derekmok.com

Yes. Too many shortcuts. Sorry.

Television video doesn't transmit the electrical signal "Red" (for example) because generally, there's no room for it in a TV channel, so the broadcasters played a few games.

Television since the thirties has been in black and white. It still is. You can turn on a black and white TV and it will still work (not in HiDef, of course). To get color, the broadcasters added color separation signals.

How much of the actor's gray dress can be separated into red.

There are two separation signals R-Y and B-Y. R-Y carries two colors. Sorta red and kind of cyan. B-Y carries greenish and bluish. For electrical reasons they didn't come out exact. Sorry.

If you produce perfect, sharp, clear, immaculate video, that would be 4:4:4. That's the quality or sampling of Y (black and white) R-Y and B-Y. That is sometimes used in special effects work where color accuracy is top priority.


It turns out you don't have to do that. Pictures look just ducky with fuzzy color, so they transmit the two color signals at half quality. 4:2:2. Almost all grown up video works like this. Broadcast, DigiBeta, HiDef, etc.

Then they discovered you could easily produce consumer grade video by wrecking the color channels even more. This produced 4:1:1 and 4:2:0. The metaphor falls apart here, too. 4:2:0 does not mean they leave out blue. It means they process the colors in a way that saves room but whose sampling doesn't come out even.


It also turns out you don't notice the difference until you see them side by side. We were doing DV quality DVDs for the Commercial People until we got tired of the phone calls and went to "uncompressed" 4:2:2. Then the phone calls stopped. The commercial people, of course, were able to compare our work with "the real thing" video on their servers and they were not pleased with our DV work.


That's the problem you have. You have damaged fuzzy color (and maybe other problems) and no amount of post will fix it.


This brings to mind an old joke about Kodak offering a "Focusing Solution" to clear up all those shots where you forgot to focus the camera. It only came in the 3-1/2 gallon size, you had to mix it up all at once, and it only lasted for ten minutes.

Koz

now i am really confused

i shoot with a dvx100b which i think is 4:1:1 by what i have read i thing that i can not get 4:2:2.
i think i am rite there.

i have heard so much on 4:2:2 being better that i am now trying to figure out how to get it.

considering the cam i am using, is there no way of improving its quality?

if each number represents an amount of sampling why cant a capture software program just re-sample the :1:1 part twice?

i have been on a mission to find a way to get better footage with what i have, plus a couple of ad-ons under 500 bucks. it seems to me there must be a way of improving the look of what i get from my dvx100b.

""" What you do with what you have, is more important than what you could do, with what you don't have."

> > > Knowledge + Action = Wisdom - J. Corbett 1992 """"

The standard that you are shooting in - the format - mini-dv- is 4.1.1. If you want better you need to get a better camera. Ooooor - this is a wild shot - I vaguely remember that there was a new bit of software that could grab the video from your camera before it went to tape (this is where the compression happens). It was on the Digital Production Buzz, so maybe Michael remembers what it was called?

Have a look at something like the (coming soon) Red camera. Now that's going to be an awesome rig for the price, I reckon.

Of course you can do colour correction to clean up your image even with 4.1.1. It won't be as good as a more expensive setup, but then that's why it's more expensive. Just setting up the full range of contrast can do a lot for a mini-dv image.

We shot a feature on miniDV and edited/color corrected in FCP. We only spent a grand on the movie but have a distrbutor interested. On their lis tof deliverable they want masters of the movie on DigiBeta. Just trying to figure out the best way to get there with making the movie look like crap. Someone told me the missing color info causes or can cause problems in the picture resulting in a less than perfect QC report. Hope that helps?


> you haven't said much about your project.
> what form it;s in, (have you graded it already in
> FCP?)
> & why you have to go to digibeta.

Just get it transferred to digibeta tape. You can't make the colour info come from anywhere, and most likely, if they have already approved of the film, it will be fine anyway.