just finishing a D.I. for my last film

[David Mullen ASC]

As I have been shooting "Big Love" for HBO, I have also been doing the D.I. color-correction for "The Astronaut Farmer" on my off days, mainly the past few Saturdays. It has been interesting.

 

Previously, my only experience with a D.I. was transferring "Shadowboxer" (35mm anamorphic / Fuji) to HDCAM-SR and doing a tape-to-tape color-correction using HD monitors (with the last pass tweaked in a D.I. suite with DLP projection.) We scanned on a Spirit and color-corrected using a DaVinci.

 

Then for "Akeelah and the Bee" (35mm anamorphic / Kodak) we did the "opticals" (fades, dissolves, etc.) digitally. The footage was scanned at 2K on a Northlight and color-corrected in a D.I. theater using a DaVinci.

 

This time, for "The Astronaut Farmer" (35mm anamorphic / Fuji), I got the studio to let me scan at 4K, even though all subsequent work was done at 2K. We are doing the work at the new D.I. facility on the Warner lot that did "Poseidon". Scanning at 4K on a Spirit, then downrezzing to 2K, and color-correcting using a Baselight while projecting the image with a 2K DLP projector (I think -- they also have a Sony 4K projector there but I belive we were using the 2K DLP one.)

 

It's been a relevation using a color-correction system like a Baselight rather than a DaVinci in terms of the ability to draw shapes around areas to correct and track them frame by frame -- much more sophisticated than the typical Power Windows features that the average DaVinci suite is hooked up with.

 

The 4K-to-2K scan of the 35mm anamorphic material is very clean and sharp for the most part. We have some problematic shots, some underexposed 2nd Unit shots, other little mistakes, but for the most part, 35mm anamorphic photography scans great for D.I. work because of the larger negative area. Also, the lower contrast of the Fuji Eterna 500T stock seems better suited for D.I. work than the higher contrast of the F-64D and F-250D stocks we used, in terms of flexibility in color-correcting.

 

We had a lot of weather problems that caused some mismatching in our day exterior stuff, and with the digital color-correction tools, we have matched that stuff better.

 

There are a few shots where we probably have pushed digital color-correction a bit heavily, for various reasons, some for the sake of stylization. For example, we shot in the fall and the cottonwood trees on our ranch location went from green to bright yellow in the course of production. At first, we started turning the yellow trees back to green digitally, but then decided the yellow was so pretty that it was better to keep the fall look and turn the green trees yellow (I'm talking about the same trees on the location, to get them to match scene to scene.) We are doing one efx shot where to establish a time change, we will dissolve from a big tree being yellow to green (a fall to spring time jump in the story.)

 

We also have one sunset scene on a porch where I was not able to shoot at sunset (too many pages) and could not frame out the midday backgrounds, plus when I tried to light the actors with a Dino to simulate the setting sun, they complained about looking into a bright light, so I had to move it way off to the side. So I had a problem of a scene on a porch that looked midday (but in the shade) intercut with a reverse angle of them silhouette against a setting sun. So we used a lot of mattes and windows to somehow time the foreground and background separately to make it all look warmer and more late afternoon. It's stuff like that where I get the most nervous about digital edge artifacts from all the manipulation.

 

Plus with D.I., you just seem to get into situations where the actor's pancake make-up just doesn't render as a true skintone no matter how you time it. I don't know why it is more obvious than when doing a straight print off of the negative.

 

But overall, the D.I. has been a major creative tool in improving shots, matching shots, fixing dirt & dust problems, even some minor focus problems. On some close-ups where the focus was barely off the eyes, I have been able to draw a window around each eyeball and sharpen them just barely enough to get a sharp glint back into the eye and make it seem sharp. And with the high resolution of Primo anamorphic lenses, there have been a couple of shots where I have had to digitally soften faces, but have been able to window the eyeballs to keep them sharp. I have also used selective softening to hide bad hairpiece joins and other make-up problems.

 

The movie ends with a spectacular sunset shot that looks fake, with fiery red clouds, like we digitally enhanced it -- but it was actually like that in real life!

 

What's great about timing in a D.I. suite is that we look at the image projected on a huge screen, and for really fine work, we zoom way, way into the image. This way we can tell quickly if we are overdoing something like sharpening or softening.

 

We will be doing the film-out in the next few weeks.

[Joe Lotuaco]

Very, very interesting information. How do you keep track of such a complex post workflow? Or maybe the better question is what goes into your decisions in choosing your post workflow? With this whole digital revolution that seems to be going on, it seems like there are an infinite number of options available to you.

 

Also, with all the experiences you've had working with DI, do you think this will affect your decisions on future projects? Do you think you'll be finding yourself saying "we'll fix it in post" more?

[David Mullen ASC]

Very, very interesting information. How do you keep track of such a complex post workflow? Or maybe the better question is what goes into your decisions in choosing your post workflow? With this whole digital revolution that seems to be going on, it seems like there are an infinite number of options available to you.

 

Also, with all the experiences you've had working with DI, do you think this will affect your decisions on future projects? Do you think you'll be finding yourself saying "we'll fix it in post" more?

 

We have a post supervisor who worked on "Poseidon" so a lot of that area is covered. We toured some D.I. facilities and talked -- my main thing was to get the scanning done at 4K, and the supervisor was supportive of that.

 

We shot the movie with the idea that it might not get a D.I., and I still feel that the decisions we made to get the best possible negative for printing (35mm anamorphic, overexposed negative, etc.) all paid off when doing the D.I. -- in other words, you can't get around the "garbage in / garbage out" rule, you still have to deliver good images.

 

In fact, it's very different than a telecine session in a way. The tools are more sophisticated but the artifacts are much more visible on a huge screen, so a little underexposure, for example, which is fine for straight-to-video work, is visibly grainier on the big screen, so accurate exposures still matter. The best-looking shots from a technical standpoint are still the ones that I manipulated the least.

 

But it's great to be able to "dodge & burn" a frame like a still photographer. For example, I shot a twilight dialogue scene on a hilltop and was able to bring down the sky and bring up the faces in a subtle manner using digital tools -- on the other hand, the scene still would have looked great if just printed straight.

 

Another example was a big dance scene at a fairground at night, lit with hanging Chinese Lanterns and string lights -- I was able to put a soft vignette so that your eye goes to the main characters dancing among dozens of other people in the scope frame.

 

Where the "fix it in post" mentality comes in are more things like getting a lab report that a major emotional dialogue shot has a scratch running down one side, and knowing that you don't have to reshoot it but can just paint out the scratch later.

[Phil Rhodes]

Hi,

 

I probably installed that Baselight.

 

Phil

[Emmanuel Lariviere]

Hey, David. How are things looking on the D.I. for "Solstice"?

[David Mullen ASC]

Hey, David. How are things looking on the D.I. for "Solstice"?

 

I guess I'll find out once they lock picture...

[Michael Most]

In fact, it's very different than a telecine session in a way. The tools are more sophisticated but the artifacts are much more visible on a huge screen, so a little underexposure, for example, which is fine for straight-to-video work, is visibly grainier on the big screen, so accurate exposures still matter. The best-looking shots from a technical standpoint are still the ones that I manipulated the least.

 

That's not surprising. Part of the difference is that film projection is much lower contrast overall than a video image on a CRT. They typical color correction technique/trick of crushing the blacks in video does not work in a DI situation - in fact, in most cases, it can't even be done, because the combination of "super black" imaging capabilities and a proper film lookup table - which has a toe and knee - prevents it. Since the blacks in a film print are never completely black, you see things like grain that can be "crushed out" - a more polite way of saying "clipped" - on video. And, as you say, that larger image makes these things more visible.

 

But it's great to be able to "dodge & burn" a frame like a still photographer.

 

Yes, but you have to have enough time available to do it. I'm curious as to how much time they budgeted for - and how much they actually allowed - on this picture, and how much work was done prior to your supervised sessions.

[David Mullen ASC]

I wouldn't be surprised to hear that we are over-budgeted on the number of days, but luckily they haven't been on my back about that. Basically we had about 10 days budgeted I think, but I wouldn't be surprised if it's been more like 15 days total, partially because of me having to work on my days off, one Monday and two Saturdays, with the colorist working by himself and with the director during the week.

 

We have rewatched some reels many times, while others had holes while we waited for efx shots to come back, or titles over picture (beginning and end). Sometimes I went back and re-timed shots where I felt that there were too many artifacts from the approach used, playing around with the keying / windowing better, so I'm glad that Warners wasn't on my case to hurry it up, although I would have been happy to have one more pass at things.

[David Sweetman]

Warners wasn't on my case to hurry it up...

 

You mentioned earlier that you were on the Warner lot; if WB is funding this picture, does that mean it will see widespread distribution? I can't wait for this one and definitely want to catch it in theatres. Even if it's only from the online perspecive, I feel like I've been 'following' it.

[Chayse Irvin ASC, CSC]

David, when your doing an Anamorphic DI, how does work in regards to the scan? Do they scan the film with an anamorphic projector type optic infront of your neg to the 4k scanner and if so on the film out are you going to be gaining any of the benfical anamorphic picture quality once you get scanned or is that information lost?

[David Mullen ASC]

The movie was produced on a budget by Warner Independent Pictures, but the Warner brass now seems to like it enough to release it under the regular Warner Brothers label, which means more money spent on advertizing hopefully and a wider release. I don't know the details though.

 

They scan the anamorphic frame as is, so it's basically Academy width but nearly Full Aperture height -- similar to scanning 4-perf Super-35, or 4-perf in general, except that there won't be any cropping to achieve 2.35.

 

The image is unsqueezed only for viewing purposes, but the scan and corrected digital files are left with the squeeze to retain maximum vertical pixel resolution.

 

I'm going to take a guess, based on the fact that the anamorphic projection aperture has a 1.195 : 1 aspect ratio (with a 2X squeezed image, thus 2.39 : 1 unsqueezed) which means that the "2K" files for film-out will be 1828 x 1530 or so.

 

This is unlike Super-35 photography framed for 2.35, which gets cropped and then stretched by the Arrilaser back out to the anamorphic projection aperture height. So there is still some reason to shoot anamorphic even when doing a D.I., although you have to weigh the disadvantages of dealing with anamorphic lenses.

 

Of course, an all-4K D.I. would retain even more resolution from the original, but tests have shown that scanning at 4K and downrezzing to 2K retains more fine detail than scanning at 2K, so if I can't get a 4K D.I., it's the next best thing. Take a look at this article:

 

http://digitalcontentproducer.com/mag/vide...cial/index.html

[Chayse Irvin ASC, CSC]

This is unlike Super-35 photography framed for 2.35, which gets cropped and then stretched by the Arrilaser back out to the anamorphic projection aperture height. So there is still some reason to shoot anamorphic even when doing a D.I., although you have to weigh the disadvantages of dealing with anamorphic lenses.

http://digitalcontentproducer.com/mag/vide...cial/index.html

 

Awesome. Can't wait to see the film when its out. Thanks!

[Michael Collier]

it seems as though you work almost as many days after editorial than you do on set (maybe an exageration) do you find that DI is moving the number of days worked up? Also I am guessing WBs takinga liking to your project is directly linked to the DI (which you said was unplanned) and the fact they aren't on your back about it. I suppose with Akela and Astronaught will probably get you some bigger films in the future? I can't wait to see what else you have in the pipeline (I always learn a lot from your in production posts. They are like an ASC article, but much more in depth and interesting.)

[Mike Williamson]

Thanks for posting all the info, David, it's great to hear your take on doing a DI with a decent budget. Looking forwards to seeing the finished film!

[Paul Bruening]

Hello David,

 

Is 2K really enough for a DI? Is it significantly better at 4K to justify the increased costs in storage and rendering- not just in practical budget terms but in objective asthetic terms? Would that 4K actually manifest when release print, projected at my local MALCOntent 8-plex?

 

Thanks,

Paul

[David Mullen ASC]

Hello David,

 

Is 2K really enough for a DI? Is it significantly better at 4K to justify the increased costs in storage and rendering- not just in practical budget terms but in objective asthetic terms? Would that 4K actually manifest when release print, projected at my local MALCOntent 8-plex?

 

Thanks,

Paul

 

Technically, 2K might be fine for digital projection release and release prints made right off of the output neg, but if the output still has to go through an IP/IN stage like many movies, then it would end up slightly softer than something shot in standard 35mm and contact-printed through all generations.

 

Obviously you see a ton of 2K D.I.'s in theaters these days and have to ask yourself if it is OK or should you push for the higher 4K standard. Personally, I think D.I.'s of anything shot in 35mm need to be all-4K to maintain the inherent resolution of the original, but for now, 2K is more common. That's just the reality of the situation.

 

4K is four-times the data of 2K.

 

Truth is that you can't really make decisions of what quality level to work at based on worst-case or mediocre projection scenarios, unless you have no choice.

[Max Jacoby]

Everyone seems to agree that there is a visible improvement from 2K to 4K, as evidenced by the link that David posted. The only people who like 2K are VFX people, because 4K means so much more work for them.

[Michael Most]

Everyone seems to agree that there is a visible improvement from 2K to 4K, as evidenced by the link that David posted. The only people who like 2K are VFX people, because 4K means so much more work for them.

 

I wouldn't put it that way. I would say that using 4K for visual effects work causes the systems to slow down to the point where you can't work interactively, so everything slows down to a crawl, and problems aren't noticed until much later in the process, with yet more time involved to fix them. When you've got a delivery deadline (and every project ultimately has one) this can become a significant bottleneck, and in the end, the quality of the final work may suffer in comparison to what you could have had if you had more runs at it, as you would working in 2K.

 

Your description sounds like it's whining that's the issue. That's not the issue. Getting the work done to everyone's satisfaction in the time allotted is the issue. Regardless of the pronouncements of those who don't deal with such things, 4K is still beyond current technical capabilities in terms of efficient throughput and interactive manipulation, especially in the area of compositing. That's not to say it can't be done, but without considerable horsepower - even more than that currently possessed by such high end systems as Flame and Inferno - it can't be done efficiently. This is also true for color correction, which is why except in very rare cases, "4K DI" work is done using either 2K or 1K proxies, with only the final rendering done using the 4K files.

[Dirk DeJonghe]

I couldn't agree more with Mike Most.

 

Also the difference in quality going from good 2K to 4K is noticeable, but much smaller than the four-times increase in workload would suggest. Producers don't want to pay (mostly) the price premium.

 

I use Baselight for grading, most of the work is done on proxies, instant playback on full res is possible be rarely needed for color correction work. Interactivity and speed are factors that are more valued on a grading session than 2K over 1K proxies.

[Will Earl]

Only adding to this, just to agree with what Mike and Dirk have already said.

 

Every VFX house in the world would love to work with 4K material, however the technology isn't at a stage where we can manipulate that data effectively.

[Evangelos Achillopoulos]

Hello to everyone

 

I understand that the 4k images have more refined grain capture than 2k but more resolution? is something that we have to be careful.

 

If the captured 2k image being seeing on the article that Mr. Mullen points, was upres with Photoshop, none will find the aliasing problem that is being seeing on the lamp image. And moreover on the other images I can notice ?sharpening edge artefacts?.

 

So I?m not convinced that more scanning resolution will provide more resolution from film since there are a lot of limiting factors (High speed film more grain etc.).

 

I think the best is to scan in 4k but immediately down res to 2k in order to work, this will give a quad over sampling of the image and it will render better noise reduction approximately 3 - 5dB less noise.

 

But not more resolution, this is what I see in the images I have seen in the article and in the websites of Cintel and others.

 

As for the DI process, yes this is a major step forward for the art of cinematography.

 

Regards, from the sunny Athens, to everyone from a new guy to this forum