Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Ask David Mullen ANYTHING

Obvious I create a "look" on a set by how I shoot the image, I mean I don't believe in taking the time to do complex post color-correction work to the image while on the set shooting, other than basic stuff like white balancing, or picking a color temp, etc. -- i.e. anything simple. There is a time & a place for color-correction, which is in a controlled environment with proper viewing conditions. You really want to maximize shooting time when you are on a set with the clock running. However, I'm all for creating looks during prep that can be applied to dailies or on-set monitors.
 
DIT value

DIT value

I have served as a Digital Imaging Technician for a number of top cinematographers and have learned so much about the craft - thanks to all of you if you are reading this post, you have helped me be a better DP on the shows where I wear that hat myself.

I contend that the first priority as a DIT is to know how your DP likes to work. For example, based solely on David's comments in this thread, I would do my best to gain an understanding of his goals during prep and for the most part leave him alone on set - assuming what I am seeing on the monitor jibes with what he was going for.

This is where the judgment of the DIT is crucial - When do you seek out the DP and address issues you see? There are two primary areas I watch like a hawk:

1) Exposure levels. I will often casually walk over to the DP and privately mention clipped areas or elements that play in scene that are in the noise floor. Often times the DP will quickly tell me they don't care if a particular part of the image is blown out, etc - this is when you nod and walk away, this is NOT the time to explore the vagaries of digital clipping, etc.

2) Anything unexpected. Different look than discussed in advance, coverage that doesn't look like the master, odd skin tones, moire, etc.

I basically think of the DIT as watching the DP's back.

Note: one of my favorite tricks is to roll about 10 seconds of footage on the stand ins (crew if there are no regular stand ins) with the final or at least roughed in lighting and run that clip back to my van (or cart) and pull up the RAW image in Scratch (or RedCineX) and check it over. In most cases everything is fine and there is no reason to bother the DP, but if I see something that would be difficult to fix later I will drag them back to my station and have them review the issue.

Final note: be aware of what your DP is dealing with at a given time and what is likely to be asked of them in the next 3 minutes before you talk to them. If you choose your interface times wisely you can do your job without slowing the day down - if you are up in the DP's face at the wrong times you can irritate the crap out of them and reduce their effectiveness at collaborating with the director/actors/gaffer, etc.

Timing and demeanor is critical. Keep virtually all of your discussion with the DP private. Do not second guess the DP around others - I made that mistake one time and I will never do it again. Don't be afraid to state a bold opinion, you are a creative partner, but you must do it in the right way at the right time.

OK, off my soap box. My apologies to David for such a long post on his Q&A thread but I firmly believe that refining the relationship between the DIT and the DP is a part of the evolution of how to consistently achieve desired results with digital RAW media.

Cheers - #19
 
Thanks for sharing Blair, we have some equally intense DIT's here in OR and they all worship you and your kit:) Think this mantra certainly will apply to young up and comers as much as it will to seasoned 35mm DP's, still virgin/resistant to digital.

Conversely, I would now love to hear what David's take (from a Cinematographers perspective) desires/admires in a DIT.
 
Any of the high-end primes are great, but if you mean "one PL-mount lens" instead of "on PL-mount lens" it mainly depends on how much zoom ratio you need, whether you need more wide or telephoto, and how much size & weight do you want to deal with.

I've been using a 15-40mm Ang. zoom lately that is fairly small & light for a 35mm zoom, but it's not very long. You may want something in that popular 18mm to 90mm range.
 
Hi David,

How exactly are LUTs applied to monitors?

My only experience has been with the Panavision GDP (Genesis Display Processor) box. Generally there is a computer processor of some sort that is feeding the monitor with a display LUT.

I don't know if anyone now makes an HD monitor that can be loaded with custom LUT's directly. I assume all pro monitors have some sort of internal LUT working inside to calibrate the signal for display but I'm not sure which allow you to easily create or manipulate or load a custom display LUT, like to turn a LOG image into one with Rec 709 gamma.
 
Hello, David. What's your opinion of Aaton's choice of DNxHD and DPX recorded output for the Penelope-Δ?

_dsc3490-text-web_201648_f.jpg
 
I don't know if anyone now makes an HD monitor that can be loaded with custom LUT's directly. I assume all pro monitors have some sort of internal LUT working inside to calibrate the signal for display but I'm not sure which allow you to easily create or manipulate or load a custom display LUT, like to turn a LOG image into one with Rec 709 gamma.

The CineTal LCD monitors have been able to load LUT's from USB sticks for quite some time, and have been used on many sets specifically because of that ability.
 
Hello, David. What's your opinion of Aaton's choice of DNxHD and DPX recorded output for the Penelope-Δ?

I don't have an opinion one way or the other -- I'm waiting to see what the actual quality is, and secondly, how it goes through dailies, color-correction, etc.

I assume the camera isn't converting/debayering 4K RAW into 10-bit 4K RGB DPX, but is storing 4K RAW in some sort of DPX wrapper. Otherwise, what's the point of tripling the amount of uncompressed data to record? So I have no opinion on what that will look like.
 
Hi David.

I know you have a lot of experience with filmouts from digital material including F900 and RED. In terms of workflow, how involved are you normally in the process once the recording media leaves the camera?
What does a normal post workflow from a file point of view looks like in the higher end of the spectrum, specially with a F900 until it gets to filmout?
Pretty much all facilities want the movie to be in a DPX sequence for the filmout right? At least as far as I know DPX is the standard for all film recorders out there, but correct me if I’m wrong.
But I would think it only gets converted to DPX in the very end of the process, possibly just before color grading or even after that, only after all is finished.
So roughly speaking, how a show shot on HDCAM would flow, for example HDCAM in camera > HDCAM SR dubs> online > DPX?

On a different matter, when it comes to green screen shooting, it’s my understanding that you should avoid customized gammas like Hypergammas and cine gammas. Do you agree with that? If so, what gamma should be used, Rec-709?
Still on the green screen matter, green spill is obviously one of the worst problems in green screen. Since spill is usually a bigger problem in the darker parts and shadows, do you feel like using a negative value for the low key sat function in the f900 is a good idea? Have you used that? When do you think you could run into trouble with it or how could over doing it hurt more than help in the case of a green screen shot?
I have also heard knocking it down a bit into the negative numbers is a good way to help with noise in general.
Thanks David.
 
All the F900 movies I shot went through a standard HD tape-to-tape conform and color-correction, not file-based, in Rec 709 color space since the F900 does not have a Log mode.

So basically once we created a color-corrected HD tape master, it went through a conversion LUT (Rec 709 to Log) for film-out. So essentially it was treated as an HD broadcast project in terms of post. The only twist being that half of them were framed for 2.40 and we had to make various masters for home video / broadcast that were letterboxed 2.40, full-frame 16x9, and pan & scan 4x3 (they were all shot 16x9 and protected for the full signal.)

So far, my Red features were shot for theatrical release in mind and have gone through a file-based D.I. workflow. Because of political issues between the parties involved, the final post on the two Red features I shot do not involve the participation of the director and myself -- the producer has taken complete control over the post. So I believe they are being finished at Lightiron doing a real-time conversion from the R3D files for final color-correction. I don't know what the final master will be at, probably 10-bit Log RGB DPX, maybe 2K.

I'm shooting a TV pilot right now on the M-X Red and I believe Encore Video in Hollywood is doing a file-based post involving RedRocket conversions from the 4K HD R3D files to 10-bit DPX 1920 x 1080 files in RedGamma.
 
My experience with green screen on the F900 has been to basically follow Panavision's advice and just turn everything off (except the shutter of course) -- detail off, the color matrix off, knee, black gamma, etc. And then light the best green screen you can, evenly, and expose it properly with as much green saturation as is possible with lighting. Color Matrix / Gammas can't really add information that the sensor didn't capture and Rec 709 basically throws away a certain bit of information.

Of course, the other problem with the F900 is the HDCAM recorder, you'd get better results if you could record the HD-SDI output to 10-bit 4:2:2.

With the Genesis, I just shoot Panalog for greenscreen in 4:4:4 mode.

I usually don't mess with Black Gamma / Black Stretch on video cameras anyway, I don't feel that they lack in shadow detail, just overexposure detail, and I feel that monkeying around with the bottom end just leads to noise in the shadows.

Knee sat shouldn't really affect the greenscreen much since you aren't exposing it so high that any of the green is in the area affected by knee controls.

I haven't used the F35 yet so I can't comment on whether HyperGamma versus S-Log is better or worse for greenscreens.
 
Thanks David.

I haven't used the F35 yet so I can't comment on whether HyperGamma versus S-Log is better or worse for greenscreens.

The F900R has the hypergammas, I think the f900/3 may too. I was actually talking about the F900R. I should have specified it.

My experience with green screen on the F900 has been to basically follow Panavision's advice and just turn everything off (except the shutter of course) -- detail off, the color matrix off, knee, black gamma, etc. And then light the best green screen you can, evenly, and expose it properly with as much green saturation as is possible with lighting. Color Matrix / Gammas can't really add information that the sensor didn't capture.

Interesting. Detail off is a given but other things like knee and matrix is a bit surprising, but kind of makes sense.
So basically shooting with the sensor at it’s flattest. I will give that a try when I run some tests.
About the shutter, I have seen people recommending shooting with a lower than 180 shutter to minimize motion blur which is bad for keying. But I think it should make up for a funky look. I would think that wouldn’t mean going to a 45 but I don’t know. Anything lower than 180 may look too of, even if it will make keying easier.

and Rec 709 basically throws away a certain bit of information.



I thought Rec 709 was the standard and you couldn’t get away from it? Meaning if you are not using a hypergamma it means you are in Rec 709.


Of course, the other problem with the F900 is the HDCAM recorder, you'd get better results if you could record the HD-SDI output to 10-bit 4:2:2.

That should definitely help. But uncompressed takes a lot of processing power. Have you done it this way? I guess even if not uncompressed, as long as you go to a lossless codec or the so called visually lossless ones it should still be better than HDCAM tape. There’s Apple Prores for example. Not sure what the results would be like but its 10 bit 4:2:2.


Knee sat shouldn't really affect the greenscreen much since you aren't exposing it so high that any of the green is in the area affected by knee controls.

David, I think you misunderstood me. I did not mean the knee sat, but rather the low key sat function which is a function that controls the saturation in the darker parts of the image. I thought that could help with green spill since spill is usually a bigger problem in the darker parts and shadows and lowering the low key sat would lower the saturation in the shadows. Less saturation would mean less green in the shadows, at least theoretically. I was just wondering if you have used that function or thought it to be a good idea.
 
All the F900 movies I shot went through a standard HD tape-to-tape conform and color-correction, not file-based, in Rec 709 color space since the F900 does not have a Log mode.

So basically once we created a color-corrected HD tape master, it went through a conversion LUT (Rec 709 to Log) for film-out. So essentially it was treated as an HD broadcast project in terms of post. The only twist being that half of them were framed for 2.40 and we had to make various masters for home video / broadcast that were letterboxed 2.40, full-frame 16x9, and pan & scan 4x3 (they were all shot 16x9 and protected for the full signal.)


I'm just debating right now how to post a F900R project.
One idea is to get it out of HDCAM as soon as possible so it's all file based. Since it would not be mastered to HDCAM this seems like a good idea.
I just wanted to avoid too many conversions. I know stuff going into compositing is going to have to be converted into DPX files but the editing of the normal stuff can't be DPX. So I just wanted to know what type of workflow you had used before. But tape to tape is not an option for us.

So far, my Red features were shot for theatrical release in mind and have gone through a file-based D.I. workflow. Because of political issues between the parties involved, the final post on the two Red features I shot do not involve the participation of the director and myself -- the producer has taken complete control over the post.

Sorry to hear that, this sucks. I hate when things like that happen.



.
 
Last edited:
Thanks David.
The F900R has the hypergammas, I think the f900/3 may too. I was actually talking about the F900R. I should have specified it.

About the shutter, I have seen people recommending shooting with a lower than 180 shutter to minimize motion blur which is bad for keying. But I think it should make up for a funky look. I would think that wouldn’t mean going to a 45 but I don’t know. Anything lower than 180 may look too of, even if it will make keying easier.


I thought Rec 709 was the standard and you couldn’t get away from it? Meaning if you are not using a hypergamma it means you are in Rec 709.


David, I think you misunderstood me. I did not mean the knee sat, but rather the low key sat function which is a function that controls the saturation in the darker parts of the image. I thought that could help with green spill since spill is usually a bigger problem in the darker parts and shadows and lowering the low key sat would lower the saturation in the shadows. Less saturation would mean less green in the shadows, at least theoretically. I was just wondering if you have used that function or thought it to be a good idea.

When I did greenscreen work for "Tara" the efx supervisor wanted me to ue the Panalog gamma on the Genesis because he was used to compositing film scans, which are Log.

The idea being that he preferred whatever gave him the maximum dynamic range that the camera could capture. So I would think an efx person would prefer Hypergamma over Rec 709 gamma for the same reason.

When I say "Rec 709" I am referring to the gamma, not the color space. Rec 709 describes both. Hypergamma is probably still using Rec 709 color space. But I don't know how the F900R labels its normal gamma -- non-Hypergamma -- option.

All of this really comes down to what the efx person wants to work with, but if you are shooting on the F900R for theatrical work, I'd be using HyperGamma if there were no Log capability. So you want the efx work to also be in HyperGamma and stay that way through compositing.

Lowering the saturation of the low key areas may help a little with green spill but you'd be much better of getting rid of it through lighting, etc. In fact, for all you know, the efx compositor's tools for getting rid of green spill may work better if the green saturation in the low end is not reduced.
 
Mike, just a little advice from someone who does a shitload of (amateur-level) compositing: just do exactly what your efx supervisor suggests and, failing that, shoot clean.

Green screening can be weirdly difficult for unexpected reasons, but at least with an un-messed-with image (no funky gamma or post processing applied) the compositing artist doesn't have to work backward to get a decent image for keying.

The very LAST thing I would do is use in-camera saturation reduction, particularly in the shadows. First of all, that's like the simplest thing to do in post so if a compositing artist wants it, he can do it in like ten seconds and would probably choose to do so after pulling a key. But the biggest issue is that most keying software does best with an evenly lit green screen and struggles pretty badly when there are big dark patches to key out. If you set your camera to lower saturation in dark areas, the difference between dark skin tones and dark green screen sections will be less cleanly delineated and your key may require an enormous amount of roto work to salvage. Don't bake in any weird post-processing. Just keep it clean. Using a 1/2 minus green on the backlight and separating the talent from the cyc is the best spill-killer and then there are lots of post plugins to fix the rest.

Yes, shooting with a fast shutter does not look natural, but most compositing software (Nuke, etc.) has accurate built-in motion blur generators based on motion vector technology. I assume your compositing artist has this capability no problem. Whether you use it depends on the scene. For fast action it may be necessary to shoot with a fast shutter and correct in post; for normal everyday keying a normal shutter is fine. Everyone works differently. Just do what your efx supervisor says and you should be fine.
 
Last edited:
Hi David: first let me say...Joey Lauren Adams. Wow, how lucky are you. OK, the real question, seems like Titles and Credits are starting to become a widely recognized medium in their own right, along with separate awards etc. As a DP, you are in charge of the overall cinematic look, but it is unclear to me whether that look also includes the titles and credits since they are not part of the principle photography. So what role do you in particular, and a DP in general, have in their creation? Thanks
 
Hi David,

Long story short, I'm helping out a friend at the last minute and shooting his final year high school project for him tommorow morning. It's 11PM now, and it's absolute no budget, but he's managed to score some tungsten fixtures. I don't have any of my equipment or gels with me (I'm visiting my parents at home), and we need some CTB for the tungstens, and also a polarizer.

There's NOWHERE around to buy it here, and I can't go and get my gear (it's too far a journey), is there anything I can ghetto together to use as CTB? Litrally anything that will cool the light? I know it's unlikely, but much appreciated anyway!
 
Back
Top