Jump to content


Photo

Stay Cool


  • Please log in to reply
48 replies to this topic

#1 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 02 August 2008 - 11:54 PM

STAY COOL

I just finished the first eight days of ?Stay Cool?, my sixth feature for the Polish Brothers. We started shooting just two weeks after wrapping out fifth feature together, ?Manure?. This new movie is much smaller in scale, a romantic comedy about a writer (Mark Polish) with a successful pop novel about high school life to his credit returning to the scene of the crime, so to speak: his home town and high school, to give a graduation commencement speech. He hooks up with old friends and rivalries and finds himself re-experiencing all the pains of his high school years.

We are using the same three RED cameras that were bought for ?Manure?, with the same Zeiss Ultra Primes and Angenieux Optimo zooms rented from The Camera House, and PlasterCity Digital again doing the dailies work. The budget is smaller on this movie however, and we are mostly shooting on locations in Santa Clarita.

In a complete twist from the soft, diffused, brown palette of ?Manure?, this movie is being shot in a sharp, colorful style, sort of a pop art modernist feeling to the suburban locations.

We decided to update the cameras to Build 16 despite it still being in beta status at the time ? it seemed to have stabilized the week before we began shooting so we decided to take a chance. This movie will have a number of low-light scenes and I needed to be able to push an underexposed image around, so the lower noise floor from Build 16 seemed like a good match.

I only had a day to shoot any tests of Build 16, mainly to look at the noise when rating the camera at higher ASA?s. Besides shooting some charts at different speeds, I took a camera out on the road ? along with some T/1.3 Zeiss Master Primes -- at night to shoot some residential streets only lit by their sodium streetlamp fixtures.

I digitally projected the tests at 2K in the D.I. theater at PlasterCity Digital the next day. I found the increase in noise at higher ASA ratings to be fairly mild unless you tried to lift up low-end detail buried in the shadows. But the highlights and midtones were quite clean even at 800 ASA. There was enough exposure at T/1.3, 24 fps, 1/48th shutter, 800 ASA rating, to shoot by available light as long as there was a streetlamp right in that spot. Further away, the light levels fell off dramatically. The image was nice and clean and the sodium streetlamp color was closer to how it looks to you eye (orange-ish) compared to that reddish look created in some Sony HD cameras when shooting sodium streetlamps.

I did discover that no matter what frame rate or shutter speed I tried to select, there is a limit of 1/24th of a second for the longest exposure time per frame, so you can?t just go to 8 fps, for example, with the shutter off, to get nighttime landscapes exposed at 1/8th of a second per frame.

We spent the first seven days of the feature at West Ranch High School in Valencia. It was just built three years ago, so we decided that instead of the character seeing his old familiar building, he would discover that his high school had been demolished and replaced by a new structure, with fragments ala the Berlin Wall placed around campus as decorations.

This location is mainly fluorescent-lit and stark, so we had to find ways of adding color to these spaces.

On Day One, we created a local radio station inside one of the rooms in the high school, which I lit with a mix of tungsten practical lamps and color accents (some red hairlight, cooler fluorescents contrasting with the warm tungsten, etc.) We also did a location move in the afternoon to a nearby chemical lab building which could double as a hospital. I let the main hallway be a very strong cyan from the overhead fluorescents but I lit the connecting rooms in tungsten. In a day scene in one hospital room, I had a half-orange light coming through a window flat we added to the larger room.

Then we came back to the school for the next six days. Some rooms had such small windows that I had to keep to the basic overhead fluorescent look (with the tubes swapped out to daylight-balanced ones), augmented in closer shots with Kinos or by bouncing an HMI Source-4 off of the ceiling.

But on our fifth day, we had a big prom scene in the gymnasium, perhaps one of our biggest and most expensive lighting set-ups for the whole movie. Skipping ?Manure?, this was the third movie in a row I?ve done with a prom dance in a gym to light (?Assassination of a High School President? and ?Jennifer?s Body? had one) and they are never easy ? you can?t really use the overhead metal halide fixtures in these spaces for a dance. We hung about 40 PARCAN?s for the dance area, plus four HMI Source-4?s pointed into a mirror ball, a 1200w HMI follow-spot, two MAC lights, and a stage for a band with rows of PARCAN?s mounted to a truss. In ?Assassination? I ended up, in the last minute, bouncing an HMI Source-4 off of the ceiling of the gym to bring up the general ambience, since the overhead PARCAN?s are quite spotty. And I had to use extras to hide the stand for that Source-4 in the middle of the dance floor. So for this movie, I added eight Chinese Lanterns to the overhead grid, with blue photofloods inside them to create a cool ambient fill for the room. I used Chinese Lanterns because I figured they might look better if seen on camera compared to some other soft units up there.

The following day we shot the graduation scene in an outdoor amphitheatre, and the day after that, having cleared out the gym, we shot a pep rally under the normal overhead metal halides.

At the end of Week Two, which was Day Eight (we started Week One on a Wednesday), we had finished with the school and we beginning our scattered location work. Day Eight was spent at a local IHOP restaurant, with a stakebed move down the block to shoot one small scene in a tattoo parlor. Again, Production Designer Clark Hunter and I kept looking for ways to add more color to these locations. The tattoo parlor was basically a big white box with overhead fluorescents, not some black-painted neon-sign-lit funky place like you?d expect. We also had to deal with a big set of windows at one end of the room for a scene that would be shot from sunset into night for a day scene. So Clark suggested that he paint the top row of windows with letters on a color background so that I could backlight the paint and bathe the back end of the shot in colored light. That worked quite well. The back windows were mainly painted red and yellow, so I added cyan gel to the overhead Cool White fluorescents to provide an overall soft cyan ambience. Then I used a tungsten desklamp to light the foreground work stations.

I won?t be able to post many photos on this one as the last one, but I?ll show you a few actual RED frames (reduced in size and compressed as JPEG?s for the web) to give you some sense of the color approach of this movie.

Posted Image

Posted Image

Posted Image

Posted Image

Posted Image
  • 0

#2 Saul Rodgar

Saul Rodgar
  • Basic Members
  • PipPipPipPip
  • 1682 posts
  • Cinematographer

Posted 03 August 2008 - 12:36 AM

Wow! Looks great. You were not kidding about the colors being a complete departure from the muted palette of Manure.

It is interesting that you guys went with with build 16, despite its beta status. I keep hearing that NO ONE should attempt to use a beta version of the RED software for a big project! I am glad you are taking a calculated risk and prove the nay sayers wrong. Just curious, do you find the image has more noise when the camera is tungsten balanced? From here the orange mid-tones look a bit muddy, but I don't know if is the compression. It looks great anyway.

Good luck with it!

Edited by Saul Rodgar, 03 August 2008 - 12:39 AM.

  • 0

#3 Hal Smith

Hal Smith
  • Sustaining Members
  • 2280 posts
  • Cinematographer
  • OKC area

Posted 03 August 2008 - 12:40 AM

Yummy color, makes me want to buy a big bag of Gummi Bears and an even bigger bag of Jelly Bellys and pig out. :)
  • 0

#4 Richard Boddington

Richard Boddington
  • Sustaining Members
  • 5482 posts
  • Director

Posted 03 August 2008 - 12:47 AM

So David....was the decision made to shoot on Red for Manure and Stay Cool based on budget or the desire to try a new format?

Thanks
R,
  • 0

#5 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 03 August 2008 - 01:01 AM

So David....was the decision made to shoot on Red for Manure and Stay Cool based on budget or the desire to try a new format?

Thanks
R,


A mix of both -- it seemed like a chance to experiment with new technology and it suited the budget.
  • 0

#6 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 03 August 2008 - 01:05 AM

Wow! Looks great. You were not kidding about the colors being a complete departure from the muted palette of Manure.

It is interesting that you guys went with with build 16, despite its beta status. I keep hearing that NO ONE should attempt to use a beta version of the RED software for a big project! I am glad you are taking a calculated risk and prove the nay sayers wrong. Just curious, do you find the image has more noise when the camera is tungsten balanced? From here the orange mid-tones look a bit muddy, but I don't know if is the compression. It looks great anyway.

Good luck with it!


Well, with all the codec errors people were reporting just a week and a half before we started shooting, I was hesitant to switch to Build 16 -- basically the week before production started, people were saying that most of the bugs were fixed and it was stable, so we took a calculated risk but I don't think it was a big one.

The camera always looks cleaner in 5600K lighting and settings, but tungsten is certainly usable, not so bad, noise-wise. Though generally I don't expose tungsten night scenes as darkly as I might, just to keep the noise down until I get to color-correct it later.
  • 0

#7 Max Jacoby

Max Jacoby
  • Sustaining Members
  • 2955 posts
  • Other

Posted 03 August 2008 - 02:37 AM

Hi David

How long is this shoot?
  • 0

#8 Satsuki Murashige

Satsuki Murashige
  • Sustaining Members
  • 3510 posts
  • Cinematographer
  • San Francisco, CA

Posted 03 August 2008 - 02:48 AM

Looks great, David. What's your approach to the processing of the footage in Redcine or Red Alert? Do you lock in your settings in prep much like setting your printer lights with film, and then just expose with a light meter? Or do you have a computer on set where you can check the occasional .R3D file and tweak your post-processing settings scene to scene as necessary?
  • 0

#9 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 03 August 2008 - 12:46 PM

This is a 26-day shoot, like "Manure"... except that in this case, that's more reasonable! So far, our shooting days have been pleasant and normal in length.

We can't get our Pro-Res HD dailies color-corrected for elaborate reasons -- PlasterCity tells me that it takes too long once they add in the step of letterboxing our dailies to 2.40 and adding windowburns -- if we were cutting on an AVID, as Soderberg did for his RED projects, then they would take the footage through an HDCAM-SR step and do color-correction, but some reason, since we are cutting on FCP in a file-based workflow, processing times are too enormous to get color-correction plus the other things we requested.

So since the look of the ProRes files will be based on the camera's metadata, basically I just boosted the color saturation levels in the camera (one setting for the whole movie) plus am playing with color temp setting to get the look on set, nothing more elaborate than that.

But recently the director told me that the cut sequences he saw were muted and washed-out compared to our monitor image and my frame grabs from the R3D files, so something isn't quite right in the workflow chain.
  • 0

#10 Saul Rodgar

Saul Rodgar
  • Basic Members
  • PipPipPipPip
  • 1682 posts
  • Cinematographer

Posted 03 August 2008 - 07:34 PM

But recently the director told me that the cut sequences he saw were muted and washed-out compared to our monitor image and my frame grabs from the R3D files, so something isn't quite right in the workflow chain.


Ah, the workflow. This is the main reason I would remain leery about wholeheartedly embracing RAW based camera acquisition for features. Not that the DP has any say over this anyway. It just seems there are so many steps that need to be taken away from set to bring the footage back looking like what the DP and director wanted or saw on set, that it simply makes it too hard to work with. Film can also be like that, but it is a lot more established as a format so the workflow is more streamlined. I know there are ways to color correct the RED footage and ultimately bring it back to what it was conceived as, but it just seems so labor intensive and iffy it almost defeats the purpose of shooting with RED in the first place.

I imagine a few DP's working on studio pictures (if they ever embrace RED) will feel the heat if the execs are not looking at the right cc'd footage and raise objections to it. That would be awkward, and potentially dangerous for the DP's job security.

Still, RED is a new format so the kinks to the workflow will eventually be worked out, I hope. Once that happens, DP's like David, who have been working on RED pictures from fairly early on will have a better handle on it. But in the mean time I am sure it can be annoying. One will just have to wait for FCP to better integrate the RED workflow to its software.

Question for David: Are you going to be closely involved in the color correction for Manure and Stay Cool?

Edited by Saul Rodgar, 03 August 2008 - 07:36 PM.

  • 0

#11 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 03 August 2008 - 08:10 PM

The main problem is not really with the workflow, it is with the lack of time for me to set it up exactly to my parameters. There are basic settings in all these conversion programs, etc. that affect how the image looks through the chain and I don't have time to spend tracking everything thru to the edit bay and down to what monitors people watch things on. In this case, I have a good working relationship with the director so as long as we see something we like on the HD monitors on set and in the frame grabs, we are not too worried about dailies. But if I were working with strangers, I'd have to spend more time setting up the workflow into something that translated into correct-looking dailies.

Yes, it's a similar problem with film dailies, whereas when you shoot in Rec 709 on an HD camcorder, it tends to look the same in the chain through editing. But ultimately, what really matters is how it looks after the final color-correction and film-out, and then how it looks in the deliverables and when released.
  • 0

#12 Andrew Brinkhaus

Andrew Brinkhaus
  • Basic Members
  • PipPipPip
  • 176 posts
  • Cinematographer
  • Los Angeles, CA

Posted 03 August 2008 - 09:18 PM

It makes perfect sense that subtle changes could occur throughout the workflow and into the editing bay/dailies. Do you adjust your production monitors so that producers/director, etc. are seeing more of a "final product" image as compared to the muted look of the RAW format while shooting?
  • 0

#13 Matthew Rogers

Matthew Rogers
  • Basic Members
  • PipPipPip
  • 175 posts
  • Cinematographer
  • Knoxville, TN

Posted 03 August 2008 - 09:59 PM

It is interesting that you guys went with with build 16, despite its beta status. I keep hearing that NO ONE should attempt to use a beta version of the RED software for a big project!


Just FYI for those of you who don't make it over to Reduser. Build 16 is no longer beta and all RED's now ship with it pre-installed. The only codec issues I ever had with build 16 were not build 16 failing--my RED Drive needed some tweaking and couldn't keep up with data rate required for what I was shooting. I actually just shot a short film with some stuff that should have error'ed, but I didn't have a single error during the 4 day shoot.

BTW David, I'm really digging these stills.

Matthew
  • 0

#14 Tom Lowe

Tom Lowe
  • Basic Members
  • PipPipPipPip
  • 1211 posts
  • Director
  • somewhere worshipping Terrence Malick

Posted 03 August 2008 - 10:32 PM

David, would it be possible to hire a full-time DIT type of person (in addition to the regular DIT) and get a few super-high-end PCs so you could RAW process all the footage on set or on location? Essentially, this way you and the director might have more control over the dailies and perhaps the final look? Or is the amount of data simply too much for a few high-end PCs to handle? Is your on-set DIT or data wrangler making duplicates of everything before it's sent to Plaster City, or is Plaster City doing the backups?
  • 0

#15 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 03 August 2008 - 11:11 PM

David, would it be possible to hire a full-time DIT type of person (in addition to the regular DIT) and get a few super-high-end PCs so you could RAW process all the footage on set or on location? Essentially, this way you and the director might have more control over the dailies and perhaps the final look? Or is the amount of data simply too much for a few high-end PCs to handle? Is your on-set DIT or data wrangler making duplicates of everything before it's sent to Plaster City, or is Plaster City doing the backups?


There's no way -- we shoot so much footage per day that it would take over a day of continuous processing to convert 4K R3D files. PlasterCity recently sent me this info:

"Right now, with 2+ hours of footage being delivered around 1AM, our overnight staff is crunching somewhere between 18 and 27 hours of rendering each night in attempt to have them done by 11AM the next morning."

And that's just to create 1080P ProRes HD files w/ letterboxing and TC windowburns. Build 16 is currently taking them 9X real time to create Quicktimes.

It's a lot like shooting film in some ways -- we won't see the final quality until we finish post and do the final conversion and color-correction.

Now it may be possible to do some simple adjustments to get the dailies to be closer to what we are seeing on the set without getting into color-correction. We are looking into that. You just have to remember that there are a number of steps in the chain and this is all rather new.
We make two back-ups to RAID's and PlasterCity also makes back-ups to LTO tapes.
  • 0

#16 Gus Sacks

Gus Sacks
  • Sustaining Members
  • 287 posts
  • Cinematographer
  • Los Angeles, CA

Posted 04 August 2008 - 12:22 AM

"Right now, with 2+ hours of footage being delivered around 1AM, our overnight staff is crunching somewhere between 18 and 27 hours of rendering each night in attempt to have them done by 11AM the next morning."

And that's just to create 1080P ProRes HD files w/ letterboxing and TC windowburns. Build 16 is currently taking them 9X real time to create Quicktimes.


That's pretty ridiculous, but sounds about right. I was doing some down-resing for a short I Red Tech'd last weekend, and it was taking 12 minutes per every minute of 4k footage to down-res to 1080p Pro-Res. And I wasn't using machines that could do 9x realtime, haha.
  • 0

#17 Tom Lowe

Tom Lowe
  • Basic Members
  • PipPipPipPip
  • 1211 posts
  • Director
  • somewhere worshipping Terrence Malick

Posted 07 August 2008 - 10:15 AM

There's no way -- we shoot so much footage per day that it would take over a day of continuous processing to convert 4K R3D files. PlasterCity recently sent me this info:

"Right now, with 2+ hours of footage being delivered around 1AM, our overnight staff is crunching somewhere between 18 and 27 hours of rendering each night in attempt to have them done by 11AM the next morning."


Jeez Louise. Well, computer technology continues to march on at a nearly exponential rate, so a year from now, that render time should be cut in half?

Perhaps with the SDK out now, someone like Cineform or Adobe will come up with a faster engine for doing all of this.

Then, of course, Epic will be released, and the amount of data will skyrocket again. :lol:
  • 0

#18 Michael Collier

Michael Collier
  • Basic Members
  • PipPipPipPip
  • 1262 posts
  • Gaffer
  • Los Angeles, CA

Posted 07 August 2008 - 01:52 PM

Looking good David! I just gripped a feature for DP Michael Hardwick shot on two reds, and we were on 16. From talking to him and the 1st AC it sounded really stable, even back then (about a month ago).
  • 0

#19 Keith Mottram

Keith Mottram
  • Sustaining Members
  • 824 posts
  • Other

Posted 08 August 2008 - 12:36 PM

There's no way -- we shoot so much footage per day that it would take over a day of continuous processing to convert 4K R3D files. PlasterCity recently sent me this info:

"Right now, with 2+ hours of footage being delivered around 1AM, our overnight staff is crunching somewhere between 18 and 27 hours of rendering each night in attempt to have them done by 11AM the next morning."

And that's just to create 1080P ProRes HD files w/ letterboxing and TC windowburns. Build 16 is currently taking them 9X real time to create Quicktimes.

It's a lot like shooting film in some ways -- we won't see the final quality until we finish post and do the final conversion and color-correction.

Now it may be possible to do some simple adjustments to get the dailies to be closer to what we are seeing on the set without getting into color-correction. We are looking into that. You just have to remember that there are a number of steps in the chain and this is all rather new.
We make two back-ups to RAID's and PlasterCity also makes back-ups to LTO tapes.


okay so two hours @ 9x = 18hours for conversion on a $4000 machine. so why cannot they get 9 machines running and have it done in a few hours. I know that I'm being a little bit blunt, but this is small fry for hardware costs compared with film dailies etc. and as for avid being done quicker i assume that is because they are making the SR dumps via scratch something that could be done in the same way for you if your editor is working FCP. this is either nonsense or a (low) budget issue.
  • 0

#20 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 19759 posts
  • Cinematographer
  • Los Angeles

Posted 08 August 2008 - 01:56 PM

How do you know that they don't already have nine machines but are processing eight other RED shows? I know that's unlikely... but my point is that it's not like my one show is the only one that the post house is dealing with on their machines and with their staff.
  • 0


CineTape

Paralinx LLC

Willys Widgets

Metropolis Post

Abel Cine

Opal

CineLab

Aerial Filmworks

rebotnix Technologies

Gamma Ray Digital Inc

FJS International, LLC

The Slider

Rig Wheels Passport

Media Blackout - Custom Cables and AKS

Tai Audio

Wooden Camera

Glidecam

Broadcast Solutions Inc

Ritter Battery

Technodolly

Visual Products

Aerial Filmworks

Metropolis Post

Opal

Technodolly

CineLab

Broadcast Solutions Inc

Ritter Battery

Rig Wheels Passport

CineTape

The Slider

Glidecam

Gamma Ray Digital Inc

Tai Audio

FJS International, LLC

Wooden Camera

rebotnix Technologies

Willys Widgets

Visual Products

Abel Cine

Media Blackout - Custom Cables and AKS

Paralinx LLC