LUT Education

[Marc Wielage]

Although that might sound good, considering the current level of understanding and competence, I have grave doubts as to who exactly is going to be the person doing such manipulations. Because you're no longer talking about cinematography, you're talking about digital image manipulation that really, really, honestly, doesn't belong on a set where you're far more concerned with making the day, dealing with things that go wrong, dealing with getting the best performance, and wrapping the location. I honestly don't understand the desire to do time consuming, meticulous work in an environment where, by definition, you have the least time available than at any other point in the process.

And that's the wisest thing said in this entire thread. Many, many DPs have told me: "the last thing I want to be involved with is worrying about anything beyond what's in front of the camera, and what the next set up I have to deal with on the shoot." They're just trying to make the day -- not worrying about setting a complicated look for every shot. They do need the DIT to tell them, "hey, I think we have a technical problem" (like corrupted data, a clipped channel, bad focus, etc.), but beyond that, they know they can stay on top of the look and get anything they want at the end of the process.

Note that David Fincher and Jeff Conenweth famously didn't use a D.I.T. on Gone Girl, only an assistant who dealt with everything afterwards for dailies and the workflow. I assume they got a decent working LUT put together to use for dailies every day, but I don't think it was anything super-complicated or involved. I think both are secure enough in the look of the camera to know how far they can push it to get what they need in post, and that to me should be enough most of the time.

 

[Hrvoje Simic]

Although that might sound good, considering the current level of understanding and competence, I have grave doubts as to who exactly is going to be the person doing such manipulations.

There are two main candidates.
Both usually sit in the grading room and one is an extension of the other.

Because you're no longer talking about cinematography, you're talking about digital image manipulation that really, really, honestly, doesn't belong on a set where you're far more concerned with making the day, dealing with things that go wrong, dealing with getting the best performance, and wrapping the location. I honestly don't understand the desire to do time consuming, meticulous work in an environment where, by definition, you have the least time available than at any other point in the process.

This was neither suggested nor is it the goal.

Cinematographers think about the image before cameras start rolling. Sometimes months before.
Plenty of time and focus to design the aesthetic.

 

[Mark L. Pederson]

They're just trying to make the day -- not worrying about setting a complicated look for every shot.

AMEN.

And herein lies the ironic, big disconnect on this entire subject. I have been on many, many sets where one single 3D LUT would have empowered the cinematographer to achieve on set monitoring and dailies much, much closer to his / her creative intent - with much less "complicated" fuss that happens in many "tents of negativity" on sets today.

I don't think it was anything super-complicated or involved.

I have a hard time with calling one 3D LUT more "complicated" or more "involved" than another 3D LUT. A high precision 3D LUT (64^3) is 262,144 numbers. Period. Using one over another is just a different tiny file of of 262,144 numbers.

That's like saying one Fuji film stock is "more complicated and involved" than one Kodak stock.

It takes the same effort to load - and the same effort to understand what each film stock does or does not do aesthetically and technically with respect to capturing light / color and being "processed" by a standardized method.

What is CURRENTLY somewhat complicated and involved is creating and adjusting 3D LUTs - and that is just lack of the right software toolsets. ... and education.

And some of us believe in this future enough to make the effort.

 

[Nick Shaw]

I definitely think looks used on set should be very broad strokes, and defined beforehand, rather than any attempt to tweak per setup during shooting.


On a show I worked on recently we set up two basic looks, one warm, one cool, with high, medium and low contrast versions of each. These were created in a controlled grading environment during camera prep. There was discussion of having a wider range of looks, but it was (wisely IMHO) decided to limit it to these few. In fact only a sub-set of those looks was actually used during the shoot.

I have a hard time with calling one 3D LUT more "complicated" or more "involved" than another 3D LUT. A high precision 3D LUT (64^3) is 262,144 numbers. Period. Using one over another is just a different tiny file of of 262,144 numbers.

I think the issues of complication has to do with what the pedigree of the LUT is. People often find a LUT somewhere they like the look of, then they tweak that look in a grading system and save the result as a new LUT. After a few itterations of this, there can be some peculiar things going on in the transforms, with colours being pushed one way by one LUT, then pulled in the opposite direction by a subsequent tweak.


You need to check the result of your LUT on a range of images. Because a 3D LUT is capable of doing things to very specific colours, which may be quite different to what they do to adjacent colours, if a particular colour does not exist in your test image, you have no idea what the LUT may do to that colour.

 

[Mark L. Pederson]

I think the issues of complication has to do with what the pedigree of the LUT is. People often find a LUT somewhere they like the look of, then they tweak that look in a grading system and save the result as a new LUT. After a few itterations of this, there can be some peculiar things going on in the transforms, with colours being pushed one way by one LUT, then pulled in the opposite direction by a subsequent tweak.


You need to check the result of your LUT on a range of images. Because a 3D LUT is capable of doing things to very specific colours, which may be quite different to what they do to adjacent colours, if a particular colour does not exist in your test image, you have no idea what the LUT may do to that colour.

GREAT post Steve.

These are the things we want to achieve in terms of education and information in a thread like this.

And it's worth noting that a grading system is purpose built to color grade - it's not purpose built to create, manipulate and visualize 3D LUTs.

 

[Nick Shaw]

GREAT post Steve.

Wrong Shaw!

Nick, not Steve.

See Shaw identification reference image above!

 

[David Battistella]

I like the conversation.

While the creative possibilities for 3D LUTs are undefinable, there are very salient points about their usage. One LUT to rule them all does not really exist (but I might argue that REDlogFIlm and DC4 are very close when it comes to RED camera's) and that has been the result of some very diligent and hard work, day in, day out by very skilled colour scientists at RED. I'm very grateful for these very colour accurate offerings from RED.

This said, I have to agree that going into any shooting situation the number of LUT's (stocks) should be limited to a narrow "palate" for the production.

I'm a tiny bit old school in that I have to say I was quite happy to hear the Gone Girl workflow meant no DIT on set. This tells me that people did their homework and went to work without any freakish paranoia about how the film was going to look or feel.

I hope the video village days are more or less behind us and I really wish I would hear more about people lighting with a light meter and creating contrast ratios and using their eyes and the knowledge of the sensor response than becoming overly fixated with the monitor image.

Also, as a side comment, I think the look of things is being taken more and more out of the cinematographers hands (except at the top levels) and being pushed into post. Even more reason that the looks need to be built on lighting ratios, style and contrast rather than the colour on the monitor. Truly, the look belongs to a team of people and the cinematographer's role is to capture the light needed to create that feeling on the neg/r3d. (because production design, art direction and performance are massive factors in successful cinematography).

Battistella

 

[Mark L. Pederson]

Wrong Shaw!

Nick, not Steve.

See Shaw identification reference image above!

LOL! Sorry Nick - posted before completing my first coffee.

GREAT post Nick

 

[Hrvoje Simic]

I think the issues of complication has to do with what the pedigree of the LUT is. People often find a LUT somewhere they like the look of, then they tweak that look in a grading system and save the result as a new LUT. After a few itterations of this, there can be some peculiar things going on in the transforms, with colours being pushed one way by one LUT, then pulled in the opposite direction by a subsequent tweak.

I put this route into "misuse" category.

Proper use would be avoiding look iterations based on consecutive inter-dependent transforms and designing a creative LUT from the start for each specific purpose.

You need to check the result of your LUT on a range of images. Because a 3D LUT is capable of doing things to very specific colours, which may be quite different to what they do to adjacent colours, if a particular colour does not exist in your test image, you have no idea what the LUT may do to that colour.

Agreed. One way of getting the feel about its effect is by designing a transform yourself, another is directly working with the person who did.

 

[Medavoym]

Although that might sound good, considering the current level of understanding and competence, I have grave doubts as to who exactly is going to be the person doing such manipulations. Because you're no longer talking about cinematography, you're talking about digital image manipulation that really, really, honestly, doesn't belong on a set where you're far more concerned with making the day, dealing with things that go wrong, dealing with getting the best performance, and wrapping the location. I honestly don't understand the desire to do time consuming, meticulous work in an environment where, by definition, you have the least time available than at any other point in the process.

I fully agree with this. I don't understand this either.

I've witnessed countless projects where people went crazy trying to manipulate the image on set, loading custom LUTs, custom camera looks, LUT boxes and all the arsenal -- and the final movie was #$%W. They just wasted valuable time on set instead caring about WHAT THE ACTUAL FILM IS MADE OF.

The majority of projects where people on set cared about performance, camera moves, compositions, doing enough takes (which, by the way, you never ever have enough time for it as it is!) - the basic things where you make or break a movie (custom looks etc being at the very bottom of the list while shooting) - turned out great.

Of course, one could make the case that you can have both. In theory, yes. In practice, almost always you can't have everything - you have to have priorities when it matters.

Quick excerpt from an interview by the cinematographer of "Rust and Bone" (courtesy of American CInematographer). I think good filmmakers think like this:

--------------------------------------
You didn't use pre-established LUTs?

Fontaine: No, I think wanting to see the finished film on set is absurd, as if the film were already locked in. I don't know what the film will look like when I'm shooting. I have some clues and desires, but I also want to be able to be surprised. For me, it's an advantage to have an uninteresting image on the monitor on set. I feel like I'm looking at a video assist. I don't understand why, suddenly, people now expect to see the finished image on set.
--------------------------------------

It's easy to fall into the trap of trying to be too technical. People say you need to be both an artist and a technician. In reality, people fall in love with tools and forget the art part. Judging from the quality of films we see nowadays, I wish it was the exact reverse.

P.S. I'm not saying "Don't use LUTs". I'm saying, pick your battles. Use a custom LUT for your entire feature (on set, not in post). Or two LUTs (one for day, on for night). Or one warm, one cool. Keep it simple. Go out and shoot a fantastic feature where you care about your actors, about the colors on set (production design), about your camera moves, about not losing important shots, and so on. Not about the unimportant $hit. Go back to the image in post where you can discover and finesse and polish to your liking.

 

[M Most]

AMEN.

And herein lies the ironic, big disconnect on this entire subject. I have been on many, many sets where one single 3D LUT would have empowered the cinematographer to achieve on set monitoring and dailies much, much closer to his / her creative intent - with much less "complicated" fuss that happens in many "tents of negativity" on sets today.

So it seems, as is often the case, that you and I are actually saying pretty much the same thing. It just takes us 25 or so posts to get there. :thumbup1:

 

[Mark L. Pederson]

So it seems, as is often the case, that you and I are actually saying pretty much the same thing. It just takes us 25 or so posts to get there. :thumbup1:

Yes.

But ... the point I'm trying to hit with the big hammer is that a 3D LUT can do WAY more for a cinematographer than just change the color space for a display - and that is how it is currently used the most.

Further ... and it will not be the right choice for all projects - LIVE adjustments to a 3D LUT are coming - and you can be afraid of them and think they are bad idea if you want - but it's wickedly powerful stuff.

Can't wait to show this in action

 

[David Battistella]

I get the positives.

I could see this being a great benefit to photographers. As the look of a single image is very important. But in the context of a film, it's harder to decide on final looks until the film is completely assembled. Even though the 3D LUT is metadata and can be changed, the looks feel and tone of a film is really dependant on how it is assembled.

What works while looking at a shot (for much longer than it appears in the film) on set, does not always work in the edit.

But I do like the idea of painting personalized looks, etc, and that a DOP can have a "personal" LUT set, differentiating himself from the masses of LUT magicians.

Still, I'd rather light something like Rob Richardson, Storaro, Ballhaus, etc., than dick around with a LUT too much.

But then again, getting a good gaffer and Colorist involved in LUT creation is an interesting proposition, where "look cooking" with lighting, LUT's and filtration could be giving us the cinematographers of the future...that is, until the CGI guys get a hold of it.


Battistella

 

[Michael Tiemann]

David, Mark, and others! I'm awakening this post with some good news: I've managed to implement GIO Scope functionality using IPP2, a LUT, and a CDL.

Technical info here (plus link to a video demonstrating the functonality): https://manifoldrecording.com/2017/12/20/gio_lut/

Cross-posted to the other mighty LUT thread, in case that's actually the better one.

My own Christmas present to myself!

 

[Mike Nagel]

Thanks Michael, will check this out !

 

[Audy Erel]

David, Mark, and others! I'm awakening this post with some good news: I've managed to implement GIO Scope functionality using IPP2, a LUT, and a CDL.

Technical info here (plus link to a video demonstrating the functonality): https://manifoldrecording.com/2017/12/20/gio_lut/

Cross-posted to the other mighty LUT thread, in case that's actually the better one.

My own Christmas present to myself!

This is amazing.. thanks for all the hard work, Michael! I always use GIO Scope for every shoot I've done. It's the most versatile tool ever. Kudos!