When Do We Eat?

[David Mullen ASC]

Went to a screening last night of the movie, which I haven't seen on film yet. 24P HDCAM Panavision, cropped to 2.39. 35mm scope projection on a moderately large screen at the Pacific Design Center.

 

Well, after seeing some projected print dailies from "Akeelah and the Bee", it's more obvious to me than ever how much better 35mm anamorphic is over HD-to-scope, just in terms of detail, depth of colors, well, everything! I always knew that but last night recomfirmed why I'd rather be shooting 35mm...

 

The movie looked fine, don't get me wrong. And it played well for the audience. And since it is mostly a talking-head movie, there were times when I actually preferred the more diffused shots to the sharper ones. FotoKem did the transfer to film and while they may have gotten away with some more sharpening, perhaps it was better they kept it a little less sharpened-looking because it was more flattering to the actors plus almost no one at the screening thought it was shot digitally. It was relatively artifact-free other than some inconsistent noise in colors due to compression/color-correction that always seems to happen with HDCAM.

 

But one of the odd things about digital color-correction versus film timing is that while you'd think that digital color-correction would lead to better shot-to-shot matching because of the greater control over the image, the reverse is actually true for the typical feature that color-corrects in less than a week. Because every shot is basically eyeballed by you and the colorist. Sure, you can use frame grabs & split-screens to match a shot to a previous shot, which helps, but it is still done by eye.

 

With film, assuming you exposed it more or less consistently, a few printer light variations within a scene is not going to alter to overall look of the image. The stock won't be more saturated in one shot and less in the next, or certain colors saturating more than others shot-to-shot. It has a BASE consistency of look designed into the stock that is hard to deviate from, for better or worse. In the digital realm, whether you are talking about D.I.'s or digital photography, you CAN vary the look and thus the problem is more about matching than anything else. And your eyes get tired day to day in the color-correction suite and you really only have time for one complete pass. Plus I think digital projection masks some color variations that are more visible in a transfer to film, where film can resolve those subtle differences.

 

The plus side to an all-digital post though is that the transitions (wipes, dissolves, etc.) are all perfect, no changes in look compared to cutting in opticals. And the digital efx cut in seamlessly.

[Guest Kai.w]

With film, assuming you exposed it more or less consistently, a few printer light variations within a scene is not going to alter to overall look of the image.  The stock won't be more saturated in one shot and less in the next, or certain colors saturating more than others shot-to-shot. It has a BASE consistency of look designed into the stock that is hard to deviate from, for better or worse. In the digital realm, whether you are talking about D.I.'s or digital photography, you CAN vary the look and thus the problem is more about matching than anything else.  And your eyes get tired day to day in the color-correction suite and you really only have time for one complete pass.  Plus I think digital projection masks some color variations that are more visible in a transfer to film, where film can resolve those subtle differences.

<{POST_SNAPBACK}>

 

Made a similar but slightly different experience when I graded a Students Short (HDCAM) in 2001 which was later printed to 35mm on an oxberry CRT printer. At the first screening we saw both the filmprint and the digitalprojection. The filmprint gave it some "texture" which covered alot of the mistakes I did at grading resp. compositing (was one of my first "big" projects ;-)

However I felt like the digital projection was much less forgiving than the filmprint. The colors appeared clearer (revealing any mismatches) the lack of grain the stability of the image, etc., all nice "artifacts" of film we have become so used to. In the digital projection the image was so "perfect" (in terms of lack of these artifacts) everything else that was'nt stuck out like a sore thumb...

Considering this one must really appreciate what Lucas and ILM managed to deliver with Episode II. Because this one was perfect.

As you said, decent DIs take would take some time...

 

-k

[David Cox]

I wonder what they used to do the grade? You see with the software we use (mistika), once you have a "look" for one shot, you simply cut and paste all the processes to the next shot. So as long as the two neighbouring shots have the same original properties, the two sets of processes will (have to) do the same thing so there is no (human) matching to be done.

 

However, if you push a grade so that your result is a long way from your starting point, you amplify the natural differences in the shots. But if the original footage was constant then there's no real excuse for what you have seen.

 

My guess would be that they have used a system that does not allow them to copy processes from one shot to another, and so they have had to start from scratch each time. Either that, or the original footage was very different between cuts and they just have had a hard job matching up.

 

David Cox

Baraka Post Production

http://www.baraka.co.uk

[Elhanan Matos]

I believe Fotokem uses Quantel IQ's for their digital post.

[John Sprung]

> My guess would be that they have used a system that does not allow them to copy processes from one shot to another,

 

What kind of system wouldn't be able to do that? Everything we've used for as long as I can remember does it.

 

 

 

-- J.S.

[David Cox]

What kind of system wouldn't be able to do that?  Everything we've used for as long as I can remember does it.

-- J.S.

<{POST_SNAPBACK}>

 

Well exactly - thats whats confusing me. After all, if you take a single grade setting and apply it to the whole film, thats the same as an optical regrade. So theres still no excuse for shot-to-shot changes unless the original footage was very different between set-ups. I haven't seen the film so I'm guessing, but maybe there were shots cut together from two angles shot seperately (classic one camera conversational shoot) and the light was very different for the alternate angle. The grader then battled to make them match? But I am guessing!

 

Of course it might be the post graders nightmare of the DOP saying "don't worry, I graded it in camera!"

[Phil Rhodes]

Hi,

 

> once you have a "look" for one shot, you simply cut and paste all the processes to the next shot

 

As you can for practically any grading software worthy of the name.

 

Phil

[Sam Wells]

Of course it might be the post graders nightmare of the DOP saying "don't worry, I graded it in camera!"

<{POST_SNAPBACK}>

 

Yes but in a sense this is what David is saying ["It has a BASE consistency of look designed into the stock that is hard to deviate from, for better or worse"] and I think there is truth in that.

 

-Sam