DIGITAL INTERMEDIATE ROUTES

[nitinsagar]

I have shot a film os super16...my post house is suggesting me to grade the film on spirit..take a HD D5 out..confirm the edit on Fire...send for film out.

 

is this route good enough? :unsure:

[David Mullen ASC]

HD-D5 is a step below working in 2K data but it might be acceptable for you -- you need to look at some tests.

 

"City of God" used HDCAM as an intermediate (and picked up a slight video-ish quality as a result) but it probably looked the way the filmmakers wanted it to look. HD-D5 is at least better than HDCAM...

[nitinsagar]

is it possible for you to quantify it in terms of percentage??

[David Mullen ASC]

No, I can't. Basically you'd have a wider dynamic range and color depth by working in 2K data, less with HD-D5 and even less with HDCAM. Plus you start picking up compression artifacts.

 

But how MUCH of a difference there is is a matter of taste. You have to shoot tests and compare them to decide.

[GeorgeSelinsky]

I saw a print of Polanski's Pianist which was done through a DI, I believe it was a 2K process although I didn't check (I can't imagine Polanski settling for an HD DI). I personally did NOT like the very slight but detectable video-ish quality of some shots. It bothered me esp. when I saw the wide shots, I could tell something was "wrong". For a second I was even wondering if they used some odd lenticular screen to project the film on. This was a film that was depicting events made back in the 1940's, so that sort of quality didn't suit the story at all.

 

I personally hope that 4K DI's will become the standard very soon. I also hope that DI's will cost much less overall, too :D

[Max Jacoby]

I saw a print of Polanski's Pianist which was done through a DI, I believe it was a 2K process although I didn't check (I can't imagine Polanski settling for an HD DI). I personally did NOT like the very slight but detectable video-ish quality of some shots.

Oh that was a horrible DI. While I was watching the film, I kept on asking myself if this was shot on video (despite the filmgrain present). There are some exterior sunlight scenes where skintones had a disgusting brownish video-look. Even if it wasn't a HD DI, it still looked like one.

[AmatEscalante]

Is there such a thing as uncompressed HD-D5?

I have called a few post-houses and none do uncompressed D5.

Is 4:4:4 D5 format uncompressed?

 

Some info would be very appreciated.

[David Mullen ASC]

No, Panasonic's HD-D5 is 4:2:2 and mildly compressed (like Digital Betacam). The compression is probably less limiting than the 4:2:2 color space.

 

The only 4:4:4 HD videotape format is Sony's new HDCAM-SR which either gives you 4:2:2 with no compression or 4:4:4 with very little compression.

[Mitch Gross]

There currently is no uncompressed HD tape format. It simply doesn't exist. You can save uncompressed data on hard drives but not to tape.

[John Pytlak RIP]

You can also save uncompressed data to film. ;)

 

And if stored properly, play it back in 50 or 100 years. B)

[AmatEscalante]

Thank you very much everyone. This info really cleared things up.

 

And many congratulations to you Mr. Mullen, for your new ASC status, I just noticed. I wish the best for you and the work you do.

 

 

Good Day

[Marc Wielage]

Is there such a thing as uncompressed HD-D5? I have called a few post-houses and none do uncompressed D5.

Well... there sort of is. Toshiba's D6 format, though not widely used, is uncompressed, though I'm not sure whether the color is sub-sampled (like 4:2:2) or not.

 

But the bottom line is that real 2K output is widely available at fairly affordable prices now. I know of a DI feature done recently in LA for around $150,000, which is almost half what it used to cost in 2K. The key lies in making very few editorial changes, going into the color-correction session with a clear intent of what you want to get out of it, and getting the work done quickly.

 

It can be done -- albeit with some compromises -- but I think it's a reasonable workflow, even for a low-budget feature.

 

--Marc Wielage

colorist

[nitinsagar]

$150,000 is a lot of money specially for in india ...in this amount of money we can DI nearly 3 films....that to a 2K scan grade and film out.(which seems like a dram to me right now)

In India DI is still in infancy stage,but it's looking up."Producers" still (and obviously) weigh that extra cost to no substantial value addition..so my fear is low technical standards will again become benchmark ....but something is better than nothing. B)

[Erin Harvey]

What about when you shoot Super 16? I've been making the rounds at Cinesite, Technique, and Efilm for a feature that I recently shot. We were zeroing in on Efilm until I read this post. They were going to try and save us some money by using D5 instead of raw data, but they insisted that it would NOT be a quality issue at all. The guy we spoke to told us that in 16mm, you're never getting a real 2K scan anyway. It's always an HD scan, and then it's a matter of how you move the media about. Is that true? I believe what he said is that our negative would be scanned on a 2K machine, but then the data would be stored on D5 becuase it would be a cheaper way of bringing the data in-house than by using the raw data. But again, he said the quality would be the same. Is having a scan stored on D5 different than having a negative telecine'd to D5?

 

Thanks for any info. And the producer needs to make the final decision about which post house to use by tomorrow, so quick replies would be greatly appreciated.

[David Mullen ASC]

D5 is 4:2:2 10-bit, mildly compressed, while 2K data is 4:4:4 uncompressed RGB data. With 16mm, you may not see enough of a quality loss to bother you, but resolution is similar if the scan was done using a Spirit -- it's more an issue of color space. You'll have more latitude to push the image around working with uncompressed 2K data but if you don't plan on extreme digital color-correction effects, you may be fine with D5.

[Erin Harvey]

Thanks for the great info David - on both boards. I definitely don't plan on pushing the image around much. Is that the real quality difference, one of freedom to manipulate the image? Or would you still say that there is an inherent loss of image quality simply by transferring data through D5?

[Jayson Crothers]

I suggest you speak to IO Film in North Hollywood; small facility with top notch people - they're new and are strictly a DI house, not a company that added DI's to their service list - they're doing interesting (and good) work in an efficient and VERY affordable manner over there - high marks all around.

[David Mullen ASC]

By definition, there's LOSS when converting a film image into 4:2:2 YUV video with compression. The real question is whether it's noticeable and whether it limits your ability to manipulate the image.

 

It's a little like asking is it worth using 200 ASA film instead of 500 ASA film or using a prime lens instead of a zoom. Obviously we all don't shoot on 50 ASA 65mm film on prime lenses and then scan it at 15K resolution -- we make decisions on just how much technical quality we really need (and can afford.)

 

You should get them to color-correct and record out to film a one-minute comparison test. It's just something you have to judge for yourself.

[Erin Harvey]

I wish I had the time to do the film out tests you suggested, but we've come up against unexpected time barriers to have this film released in the time span that the producers would like. So all of a sudden, I'm forced to choose a facility today!

 

 

Okay, here's what it boils down to for me...

 

I can choose one facility that can give me raw data, but I don't like that colorist's work so much on 35mm projects I've seen.

 

-or-

 

I can choose another facility that will transfer my data via D5, but I really like that colorist's work on 35mm projects I've seen.

 

 

I'm leaning towards the latter, thinking that a good colorist could deliver me a better product within the 4:2:2 space than the other might be able to deliver within a 4:4:4 space. Wish I could have the best of both worlds, but it doesn't look like a possibility given our budget.

 

But the colorist I'm not so excited about is no hack. Known as the top colorist at one of the big three: Efilm, Cinesite, Technique. But again, I have seen a couple of films come out of that facility where I could see that the image was digital, even though I had no idea before going into the theater. The opposite is true with the other two facilities. Am I making the right choice?

[David Mullen ASC]

Sort of depends on how much control you think you can exert over the colorist. You don't know who made the decisions that affected the look of the final projects that you didn't like the look of. And the facility that did the D.I. work you liked in the theaters, are they offering you the same equipment that those films used? If not, then how can you be sure that the colorist will make up the difference?

 

It's a half-dozen one way or the other so flip a coin and live and learn.

[Marc Wielage]

The guy we spoke to told us that in 16mm, you're never getting a real 2K scan anyway. It's always an HD scan, and then it's a matter of how you move the media about.

That's not true at all. You can absolutely do a 16mm scan at 2K on Northlight scanners. This is an option that's only come out recently, but it does work -- they were demonstrating it at NAB a few weeks ago.

 

Also, I've output Super 16mm from Spirits at 2K res, and the image is acceptable. It's certainly greater than HD resolution -- bearing in mind that there are limitations in color bandwidth and image stability with the Spirit.

 

--Marc W.

[Eric Steelberg ASC]

You can also save uncompressed data to film. ;)

 

And if stored properly, play it back in 50 or 100 years. B)

I keep saying the same thing! I'm terrified of what's going to happen 10 years from now when it's hard to find equipment to play back what was shot last year.

 

I hope Kodak continues to find ways to help DP's make the film choice more financially attractive to Producers. I'm looking forward to seeing the tests that Russ Alsobrook just shot with the Panasonic cameras and film.

[Marc Wielage]

I keep saying the same thing! I'm terrified of what's going to happen 10 years from now when it's hard to find equipment to play back what was shot last year.

I think as long as there are data archives being made of each film, you won't have a problem. There will always be ways to convert old files into new file formats, without loss. The DI facilities refer to this as "data migration" -- moving features from one format to another -- and it's something you should ask about when doing any project.

 

As one example, just the other day I had to ressurrect a 1989 Pagemaker file, and bring it into a 2004 version of inDesign. Oh, it took me about 20 minutes, but eventually I was able to get it all back, and all the artwork, photographs, and so on were exactly where they should be. Absolutely no loss. Same with getting back old Microsoft Word files from the mid-1980s. No problem.

 

Also, as long as the studio archives 35mm IP's and Negatives of the material, you can always work with that. Given good storage conditions, I think a decent IP should last at least 50 years, probably a lot longer. I've personally transferred 50 year-old IPs of some titles, and they actually held up pretty well, considering the circumstances.

 

I would worry far more about video formats. Look at NBC's use of the MII component video format, for almost a decade, in the 1980s. There are thousands of hours of programming archived on that format, and getting MII tapes to play back today is a real problem, since nobody makes the machines anymore. I suspect it will be very, very difficult to play back a D5 or HDCam videotape in 50 years, simply because I don't think the machines will be available -- and that's assuming the cassettes survive OK.

 

Note that if you do shoot 1080 24p, there's no harm in asking the facility to create a data archive of those files. I believe the data tapes only cost about $75 each, and they'll hold up to one reel of film. I think archiving an entire feature as data could cost as little as $500. Again, you'd have to make the facility responsible for migrating the data to new formats as time goes on, but the data should survive for the forseeable future -- potentially, longer than 35mm film.

 

--Marc W.

[Marc Wielage]

I can choose one facility that can give me raw data, but I don't like that colorist's work so much on 35mm projects I've seen. -or-

I can choose another facility that will transfer my data via D5, but I really like that colorist's work on 35mm projects I've seen.

Colorists have different personalities, and we're all different people. My gut feeling would be, avoid doing the job on D5, because the bandwidth isn't great enough to preserve all the information for a really top-notch DI.

 

Ask the first facility if they have any other colorists available, and audition them. Bear in mind that, much of the time, colorists work under the thumb of the cinematographer (and/or any other creative filmmakers present in the session). What you see of their work may not necessarily represent what the colorist wanted to see. When in doubt, ask the colorist why the film looked a certain way. If you hated it, tell them; maybe there's a reason why it looked that way.

 

If a client tells me, "I want this shot pea-green," far be it from me to argue with them. Colorists can make tactful suggestions up to a point, but the bottom line is, it's the DP's (and the director's) film, not the colorist's. If the filmmakers deliberately want the film to look a certain way, our job is to give that look to them. I've often said, "our job isn't to make the film look good -- it's to make the client happy." Ideally, you can find a way to do both, but the latter is far more important.

 

At the same time, if the colorist's opinions are asked, we'll be glad to give them. And I will speak up if I think the filmmakers are about to "drive off the road" visually and get into an area that's going to cause technical problems. This is more of an issue in video, where you get some foreign distributors (particularly Canal Plus in France or Network 10 in Australia) who''ll raise warning flags with ultra-dark sequences. In those cases, I routinely ship a memo with the master videotape saying, "this scene is dark due to creative intent by the director!" Nine times out of ten, that shuts them up. I say, if you can see it on the monitor in the control room, it'll play at home for regular viewers on DVD.

 

Much of the time, though the toughest chore is finding a look and then maintaining it seamlessly throughout a sequence. But the "look" isn't the colorist's -- it's the filmmakers.

 

--Marc W.

colorist

[Eric Steelberg ASC]

I would worry far more about video formats. Look at NBC's use of the MII component video format, for almost a decade, in the 1980s. There are thousands of hours of programming archived on that format, and getting MII tapes to play back today is a real problem, since nobody makes the machines anymore. I suspect it will be very, very difficult to play back a D5 or HDCam videotape in 50 years, simply because I don't think the machines will be available -- and that's assuming the cassettes survive OK.

I was actually referring to video formats primarily but I worry about the integrity of digital storage mediums as well. You're examples are still relatively young. We can still re-transfer Citizen Kane and Chaplin films because of film's integrity. Let see what happens when somone plugs in a hard drive that's been sitting in storage for 80 years (assuming they can find a compatible plug.) I think archiving is a severely overlooked problem right now.