Best LUTs for Weapon?

[jake blackstone]

I think Phil means it´s easy to translate so it fits to any grading suite, if there is a lut that can not be transfered so it fits the grading suite, then I would say there is likely something wrong with the lut or the suite. But often it´s not a one click solve. But mostly less than 3, like open, transform, save.



"Lattice" does the transforms between most formats.

Yes, you can do transformations between LUT's.
But firstly, I wouldn't want to use a LUT created by someone I don't know, sorry. I'm not a fan of using some home made LUTs, as I'm not terribly keen on chasing down some weird noise, clipping etc. Yes, of coarse it's different, if LUT is supplied by a company, that employs a color scientist. And in that case, I doubt very much, that they would supply a LUT, that would need to be translated in order to work in the grading suite.
Second, as I said, not all systems work with all LUTs. Nucoda requires .cms, Baselight works with Truelight LUTs and Lustre requires .3dl. Yes, most will work just fine with .cube, but it depends on the version of software. You assume, that everyone SHOULD be able to deal with LUT translations and everyone SHOULD own Lattice. I disagree...
Third, if you're lucky enough with just one type of camera in your timeline, great! Unfortunately, modern timelines are full of different cameras, especially B-Cameras- BMD, GoPro, GH4 etc. What do you do then? More LUTs? With ACES, there is still no need for LUTs, just select an appropriate IDT and you're done. Obviously, mixed source timeline doesn't really apply to DIT work, but wouldn't it be prudent to start thinking about it? Yes, I know, presently LUT workflow IS an accepted workflow, but companies, like FilmLight are slowly changing that.
Once more, Only ACES is an OPEN format, that is available to ANYONE. No need for proprietary secret sauce here...

 

[Gavin Greenwalt]

RED's "neutral" is too green for my tastes. This "don't use a LUT hire a colorist!" false dichotomy is nonsense. RED *has a look* as its starting point. I got home a couple months ago and my fiance was watching a movie on Netflix. I said "Shot on Red." it took 2 seconds to see the telltale look. It was MX, not Dragon which made is especially easy, but RED has a look. Arri has a look. You can grade both of them but you could grade Kodak or Fuji and people still had preferences. Film LUTs have a unique perspective on the world. Digital sensors and color science have unique perspectives on the world. I run nearly 100% of my projects through FilmConvert's luts because they just provide a starting point that I like. It's not a naturalistic perspective on the world but nothing ever is and it's completely disingenuous to suggest that DragonColor2 + REDLogFilm is any less of an aesthetic choice than Vision3, neither is perfect, both are subjective interpretations. So yes, hire a colorist if you have the time and budget. But don't pretend that a 3DLUT you develop or purchase or download off the internet is anything less than the native 3DLUTs that RED includes in the menu. Those are the result of one person's perspective on what's the best starting point and their perspective and objectives couldn't possibly align with everyone else's tastes and objectives.

 

[Michael Hastings]

It sounds like the answer to your problem is that you have one camera but you want another. But most people who shoot Red, want Red. Red's signature look is precisely why I buy them. It seems a massive waste of time to shoot on one camera with the sole intention of grading it to look like another.

First of all, it isn't that I have one camera and want another - I have a $15K camera that has better resolution (which I appreciate) and (for my purposes) a better form factor than the $40K+ Alexa -

AND

It has been said specifically in this thread that users have created LUTs that essentially matches that Alexa look and if that is the case then I could have the best of both worlds.

AND

Further it has been stated in the Panavision thread that the camera as shown at the Expo had a look that was highly desirable to the cinephiles (if that's a word
:wink5 - seemingly beyond what is normally seen from the Red yet it has been generated using the same hardware, software, and OLPF that we already have - presumably via some customized LUTs.

AND it seems that some are suggesting that I should do my homework and not be so lazy - and I would reply that I have 30 years of experience in a lot of areas that maybe you all don't have (and have shared a lot of that knowledge) and it isn't necessarily the best use of my time to become a digital color expert as well - but it is doesn't make me a lazy cinematographer to see if I can get that information in another way - I asked for it to be shared but also offered to pay. I guess I also wonder why there hasn't been more done on this front by RED themselves.

Red's signature look is precisely why I buy them.

From what I hear in talking to a lot of Redusers, I find this somewhat unusual. Not that the look is bad, but usually the choice is more based on some combination of price, resolution, size than preferring the look - they often have some criticism of the look over other cameras they have used. Some of that is probably frustration with getting the look they want which is the whole point of asking about these LUTs - Can they make things easier?

It seems a massive waste of time to shoot on one camera with the sole intention of grading it to look like another.

Again, it's not the "sole intention of grading it to look like another" it's having the best of both worlds, since it seems that is possible. And getting the situation where THE WHOLE COMMUNITY has an easy workflow that removes some of the stigma that is out there - whether the LUTs come from RED or some Reduser - would be a good thing. While it doesn't necessarily affect RED right now since they are selling more cameras than they can make, there is a bias out there that affects RED camera owners.

But I think this notion that Arri has a more desirable look is something a lot of people are picking up from what they've read rather than what they themselves have experienced.

I think this is almost certainly true - particularly in the markets that the majority of Red camera owners are in (i.e. NOT feature film or high end TV production). But it is also the point - in that when it comes to rentals to small productions or hiring free lance owner/op shooters with their camera - these (probably uninformed) impressions are the key to these producer's choices. At these levels they usually call and ask for what they want and you have little opportunity to educate or convince them that what you have is as good or better. So again, if disseminating some LUTs/Looks helps to level the playing field it is a good thing.

Finally, If it is true that these LUTs are so specific to a style of shooting as to NOT be generally useful (as some in the thread have suggested, although that doesn't seem to be what was suggested in the Panavision thread) - sorry I brought it up. :wink5:

 

[Hrvoje Simic]

Third, if you're lucky enough with just one type of camera in your timeline, great! Unfortunately, modern timelines are full of different cameras, especially B-Cameras- BMD, GoPro, GH4 etc. What do you do then? More LUTs? With ACES, there is still no need for LUTs, just select an appropriate IDT and you're done. Obviously, mixed source timeline doesn't really apply to DIT work, but wouldn't it be prudent to start thinking about it? Yes, I know, presently LUT workflow IS an accepted workflow, but companies, like FilmLight are slowly changing that.
Once more, Only ACES is an OPEN format, that is available to ANYONE. No need for proprietary secret sauce here...

Sounds cool, then you stick with IDT's.

I'm sticking with analysing each camera by myself through methodology I designed and build and use my own LUTs.

No company is going to change my workflow, Apple tried with FCPsexy, and only helped me save money by ditching buying Apple workstations from now on.

 

[jake blackstone]

Sounds cool, then you stick with IDT's.

I'm sticking with analysing each camera by myself through methodology I designed and build and use my own LUTs.

I'm not a color scientist, as I suspect neither is the overwhelming majority of users, so I'l stick with a solution provided by people in the know- color scientists. If designing LUTs works for you, then stick with that. There is no technical reason why you can't.
It just so happened, that many work with other people in the given project- DITs, editors, multiple VFX places around the world,etc so I can't imagine a colorist or DIT would be creating and supplying LUTs for everyone involved, especially when using some secret sauce...
Do you know for a fact, that your LUT will have no issues, when the grade will be used for P3 for D65 white point DCP, XYZ or for HDR at 2000 nits deliverables? Will you create LUTs for those deliverables too? Or would you rather just use an appropriate ODT for a given deliverable?

 

[Mike Nagel]

So I take it you aren't a fan? Why are you on a RED forum if you despise the very thing that makes a RED a RED, the color science and image? Go troll an ARRI forum.

After u changed ur diapers u may wanna re-read my post.

It's blind senseless buffoons like u why this forum gets such a bad rep sometimes - and really cool peeps are not around here anymore....

I'm not a fan, I'm an owner. I shoot RED, and I know exactly why. And I know how to handle all aspects of the image pipeline.

And like I said, and many others have stated since '07: color is the one Achilles heel, not that it's "not good enough" - clearly it is - we're talking post workflow improvement starting at post debayer.

For you clown, as I know this is hard to understand running on half a brain cell: we want to further improve the cam and ecosystem we invested in.

 

[Björn Benckert]

So question how do I get lut in camera, what format work I though cube 33x33x33 in a look folder would do it, but no dice.

any leading hint how to pour my sause into camera and I upload a really cool on top of Dragon color 2 and RedGamma4 lut.

 

[Phil Holland]

So question how do I get lut in camera, what format work I though cube 33x33x33 in a look folder would do it, but no dice.

any leading hint how to pour my sause into camera and I upload a really cool on top of Dragon color 2 and RedGamma4 lut.

On your RED Media upload your .cube to a /luts folder. Import from there.

 

[Medavoym]

I'm not a color scientist, as I suspect neither is the overwhelming majority of users, so I'l stick with a solution provided by people in the know- color scientists. If designing LUTs works for you, then stick with that. There is no technical reason why you can't.
It just so happened, that many work with other people in the given project- DITs, editors, multiple VFX places around the world,etc so I can't imagine a colorist or DIT would be creating and supplying LUTs for everyone involved, especially when using some secret sauce...
Do you know for a fact, that your LUT will have no issues, when the grade will be used for P3 for D65 white point DCP, XYZ or for HDR at 2000 nits deliverables? Will you create LUTs for those deliverables too? Or would you rather just use an appropriate ODT for a given deliverable?

I agree with you in principle, Jake.

How about the situations where you want to EMULATE a print film stock? It's pretty common.

Only a LUT can do that. A transform cannot.

Is it legitimate then to shoot digital and use a KODAK/FUJI emulation LUT to emulate the film's tonal response and colors? I think it's the only way to do it when you shoot digitally but want to emulate film.

Thanks!

 

[Björn Benckert]

On your RED Media upload your .cube to a /luts folder. Import from there.

Thanks Phil. Actually got on my computer and downloaded the manual. So got it and got it working.

So here is promised lut, it goes ontop of DragonColor2 / RedGamma4 and converts that to a kodak negative then that negative is fluffed into a positive / print....

Not a straight shooter I know. Some one with math skills would say starting from redlog is much better but according to my taste it does wonders if you dial in your Kelvins and iso / exposure underneath. And keep track of goal posts and stoplights as it´s quite easy to expose out of range of camera, and clipping looks excellent with this lut, so better watch out as its always better to toast the highlights in post instead of in camera if you ask me.

So for the curious ones here is a free download and inside the zipped folder you also find a little REDME note with steps how to get it kicking (for all lazy asses like my self that for some reason never download that weapon manual manual

https://app.frame.io/f/Kj5Jn4ZJ


So this is not really proven, built it tonight from what I use in flame and used lattice to combine into one cube 33x33x33 3D lut and then downsized to also make the MG lut.

But for sure I think I will be shooting with a shit load of luts in camera from now on now. When I got the taste for it, I see no reason not to. Be able to scroll them trough in camera is not to shabby. And as long as we are shooting raw it can not really hurt. Love it.

Here are some with and without samples. Straight Dragon color2 / red gamma4 and then same but with lut applied.

PastedGraphic-1 by Björn Benckert, on Flickr

PastedGraphic-2 by Björn Benckert, on Flickr

PastedGraphic-5 by Björn Benckert, on Flickr

PastedGraphic-6 by Björn Benckert, on Flickr

PastedGraphic-8 by Björn Benckert, on Flickr

PastedGraphic-7 by Björn Benckert, on Flickr

 

[Hrvoje Simic]

Do you know for a fact, that your LUT will have no issues, when the grade will be used for P3 for D65 white point DCP, XYZ or for HDR at 2000 nits deliverables?

They are not typical, "colors all around" canned look type of LUTs.
I design and use them to optimize the camera's initial response and as a basis for further grading.

Will you create LUTs for those deliverables too?

All - not likely.
Some of those perhaps.

Or would you rather just use an appropriate ODT for a given deliverable?

ODT likely if going with ACES.

How about the situations where you want to EMULATE a print film stock? It's pretty common.

Only a LUT can do that. A transform cannot.

God is a DJ.
Life is a dance floor.
LUT is a transform.

 

[Björn Benckert]

God is a DJ.
Life is a dance floor.
LUT is a transform.

A 3d lut is more than a transform, or atlast it can be more than a straight transform.

 

[Robino_J]

Looks great Bjorn!

 

[Marc Wielage]

DareDevil - although entertaining show - looked horrible. It makes me kinda angry how they can shoot on Dragon and then use DC/RG4 presets which look like sh*t (and are maybe good enough for dailies) and the dreaded green yellow haze is everywhere... what a colossal waste of potential.

I agree with some of your points, Mike, but just to defend the Daredevil people: they're deliberately going for that yellow look and are actually using hard lights and yellow gels on set in the background on quite a few night scenes. And it's all a custom look done without presets. They actually spend a lot of time doing the color correction on those shows.

So here is promised lut, it goes ontop of DragonColor2 / RedGamma4 and converts that to a kodak negative then that negative is fluffed into a positive / print....

Fluffed? Jesus.

This is not what film looks like. I think there a thousand parameters that film creates that you just can't do digitally, with a LUT, with a correction, or anything else. All you've done is create a look. I don't dispute it's a look -- I just disagree that it looks like it was shot on negative film. I think you're chasing your tail here (and believe me, you're not the first).

My advice is don't worry about this. Just make what you've shot look as good as you can. It's not ever going to look like film. Heck, Kodak doesn't look like Fuji, and different emulsions of either don't look like other emulsions; old Kodak film doesn't look like 1990s film, and 1990s film doesn't look like 2000s film. It's kind of like saying I'm going to take this artificial chemical paste and make it look and taste like steak. It's a version of what you think steak is; it's not steak. And people will always be able to tell the difference.

 

[jake blackstone]

I agree with you in principle, Jake.

How about the situations where you want to EMULATE a print film stock? It's pretty common.

Only a LUT can do that. A transform cannot.

Is it legitimate then to shoot digital and use a KODAK/FUJI emulation LUT to emulate the film's tonal response and colors? I think it's the only way to do it when you shoot digitally but want to emulate film.

Thanks!

If you must use an emulation LUT, then you are free to use it in addition to ACES or RCM. Not a problem.
Said that, if you want film look, then shoot film. No LUT can make electronic camera look like film, despite all those claims made by the LUT makers, sorry...

 

[Phil Holland]

Just want to come in and make a couple of quick points.

Color Workflow

LUTs or no LUTs, ACES, etc... There's many ways to skin a cat. LUTs exist for a very clear reason. As do the ongoing developments in various packages and specific entities.

Myself, LUTs aren't always used, but I've been making them for a long time and do use them as needed. But really what informs my decision about the Color Workflow is the project and occasionally the client.

Jake and Marc know their stuff and we all have strong opinions about "how" to do things. Mostly because we've all found ways to do this professionally when it comes down to it.

LUTs are transforms. They do have their limitations. And they are indeed an industry standard "thing" to use.


Film Emulation

This is a fun one for me because this was and still is a fairly large part of what I do. Back when I worked for a VFX house it was daily practice to create digital elements and/or capture digital assets and match them to motion picture film stock.

What Marc is saying is dead on regarding emulsion variation. Every stock is different. And it's not a difference of just color either. Grain, resolving power, dynamic range, developing method, negative/print variances, etc.... A whole lot of difference.

One concern I've had mainly from the DSLR community is the reaching for a film-like look, which from certain realms sounds like an ethereal smoke creature that nobody can fully describe because they don't know what it looks like.

However, given the parameters of a specific stock, you can make a fairly accurate interpretation and "match" it as necessary.


Canned LUTs

Pros and cons, but mostly a con for me. They are nice to have, nice to use. But everybody has them and popular looks/trends tend to be used by everybody. Both fine, but also not-so-good.

Color should be more than that and it should be something that is tackled on a per project/scene/shot basis as deemed necessary.

Just slapping a quick coat of a canned LUT will give you something, perhaps something pleasing, but at the end of the day you'll have to ask what makes your image unique. What makes your image "yours".

When I use them, most of my LUTs were created for the general LOG Gamma Space and serve often as a good starting point if I decide to go that route. It's rare for me to finish a project with a quick look. It's not rare for me to create a look to be used quickly however.

 

[jake blackstone]

LUTs are transforms. They do have their limitations. And they are indeed an industry standard "thing" to use.

Actually they are not. LUTs are color tables, that are used to determine a resultant color, i.e this color stimuli on the input should be that resultant color. Transforms are formulas for transformations, like 1+2=3. Not the same thing...
Anyone can create a LUT with Resolve. You just create a grade and export it as a LUT. Once that grade is in a LUT form, it is very difficult to decipher what was actually done, but with some effort, it is possible to figure out. And because the grade now essentially is hidden, LUT makers can make any claims they want.
Creating a transform is another thing altogether...

 

[Björn Benckert]

I agree with some of your points, Mike, but just to defend the Daredevil people: they're deliberately going for that yellow look and are actually using hard lights and yellow gels on set in the background on quite a few night scenes. And it's all a custom look done without presets. They actually spend a lot of time doing the color correction on those shows.


Fluffed? Jesus.

This is not what film looks like. I think there a thousand parameters that film creates that you just can't do digitally, with a LUT, with a correction, or anything else. All you've done is create a look. I don't dispute it's a look -- I just disagree that it looks like it was shot on negative film. I think you're chasing your tail here (and believe me, you're not the first).

My advice is don't worry about this. Just make what you've shot look as good as you can. It's not ever going to look like film. Heck, Kodak doesn't look like Fuji, and different emulsions of either don't look like other emulsions; old Kodak film doesn't look like 1990s film, and 1990s film doesn't look like 2000s film. It's kind of like saying I'm going to take this artificial chemical paste and make it look and taste like steak. It's a version of what you think steak is; it's not steak. And people will always be able to tell the difference.

I spent the better part of my life around Domino, Cineon and Genesis and solitare film printers. I bought my first digital film printer when I was 22 years old and I´m far older now so I been doing it for quite some time. So no this is not just a made up look, its partly made by kodak / cinesite them selves. Then i have added a step going from red gamma. I can give you the negative stock number matched and positive print stock number involved but dont have them in my head.

Ofocurse the lab proces is an art form. You need to colortime your positive correctly and so but not far from what you have in your kelvin setting and iso seeing and today we work digital all the way so we dont need to do it blind folded like before. Then we where shooting, developing, scanning, posting, printing and making a positive to watch in the cinema was a week round trip... And normally that was just to see that it actually, looked like shit as one of the steps where ill preformed.

Now with a lut and 16 bit log camera its just a touch screen and slider for iso and kelvi up or down. And the result on the screen is far more telling then an undeveloped film roll in a can. So things is, shit easer these days. But no grain structure is not match etc. But to me it´s quite easy to shot something with weapon at 6k that exceeds anything I ever scanned in 4k from 35mm negative and I scanned a lot in 4k some whole features and I don not know how many commercials and vfx scenes.

Back then we did not have much of a clue but the guys from cinesite and Quantel that came to help us set it all up... well some of those guys know pretty darn well how to color pick a chart, measure the density and build luts that where pretty dead on already 18 years back, as they where the guys inventing it these whole thing with colorwarp cubes.

 

[Phil Holland]

Actually they are not. LUTs are color tables, that are used to determine a resultant color, i.e this color stimuli on the input should be that resultant color. Transforms are formulas for transformations, like 1+2=3. Not the same thing...
Anyone can create a LUT with Resolve. You just create a grade and export it as a LUT. Once that grade is in a LUT form, it is very difficult to decipher what was actually done, but with some effort, it is possible to figure out. And because the grade now essentially is hidden, LUT makers can make any claims they want.
Creating a transform is another thing altogether...

I have no problem with the recent trend of specifically using the term transform (more or less now defining it in a different way) in regards to ACES, however for the last 16-ish years 1D and 3D LUTs specifically used for "transforming" LOG to various other things have been known as transforms. Though the Resolve manual most recently has this definition straightened out it's still referenced occasionally in this way here and there.

It's fine. Times are changing.

I'm merely a dinosaur from the future trying to punch meteors out of the sky in an attempt to seek out my revenge.

 

[Björn Benckert]

Moving footage with the same lut.

4k process 422> https://app.frame.io/f/qMueumVQ

lut> https://app.frame.io/f/Kj5Jn4ZJ

[video]https://player.vimeo.com/video/172556334[/video]



https://vimeo.com/172556334