Film emulation software ...ColorVision? Just work throug Davinci?

[Andi Obarski]

Hey all!

Im shooting a feature with my Epic dragon and wanted to try out some film emulation software. Specifically my director and I are looking for Kodak Vision2 50D, vision3 250D and V3 500T.

im a coloring novice and would like some advice from folks who work with Davinci to see if these specs are possible in Resolve Lite or if it is worth it to purchase another software (I've been specifically looking at ColorVision Impulz)

EPIC 4K
23.98
RC 6:1
LL OLPF

thanks!

 

[Christoffer Glans]

Use the FilmConvert plugin for Resolve. It would be better with MX cause it matches the color space exactly to that of Kodak and other film stocks, but you might tweak it into getting the same look on Dragon.

 

[Marc Wielage]

My opinion is that there is no such plug-in that actually emulates shooting on film negative. I agree that these so-called LUTs do change the picture, but whether they actually match the look of the negative is very subjective.

My advice is to just hire an experienced colorist and have them come up with some looks based on good lighting and good set design choices. And use color charts on set so you know what was shot and have a known reference as a starting point. The LUT doesn't necessarily make the look (or even a good look at all) -- it's about the DP, the lighting, the lenses, the exposure, and the colorist.

Consider this: there are 29 companies out there, each offering cheap LUTs that claim to recreate the look of film... and yet they all look different. Are any of them correct? To me, they're just offering a wide variety looks and subjective ideas, without a lot of real hard science. There's a lot of smoke & mirrors here.

Print LUTs you can kind of excuse, since those can exist to a point. But recreating the look of film negative is a very nebulous area prone to a lot of subjective judgments.

 

[Christoffer Glans]

FilmConvert have done extensive research to match certain color spaces to different filmstocks. They haven't just put something "that looks like film" in there. Together with that you have full control over the parameters of that LUT.
It's recommended by top cinematographers and it's the best conversion I ever seen. As you can see when they compare Vision3 to the old Red One MX with Filmconvert it looks pretty close.
https://vimeo.com/27316178

There are times where you definetely need to try and match film. Whenever you shoot both digital and film you need precision for it. A colorist that can match perfectly to film would cost a whole lot more than using the LUT as a starting point for the grade. This isn't rocket science and it's not true that you can't match digital to film, you just have to shoot it without clipping and then tinker a bit with the settings of the LUT.

 

[Marc Wielage]

FilmConvert have done extensive research to match certain color spaces to different filmstocks.

It's all smoke & mirrors and very dubious science. I would have zero problem if they said, "hey, kids! Here's some cool looks!"

I again say: there are dozens of companies out there that purport to having LUTs that emulate film stocks:

http://antlerpost.com/Store.html
http://www.captainhook.co.nz/blackmagic-cinema-camera-lut/
http://www.cinelogdcp.com/cinelog-looks-acr/
http://www.cineplus.ch/vistalut.html
http://www.colorgradingcentral.com/osiris
http://deluts.businesscatalyst.com/
http://filmconvert.com/
http://www.groundcontrolcolor.com/
http://helloluxx.com/product/prolut-cinematic-colour-grading-presets/
http://www.kojicolor.com/
http://www.looklabs.net/
http://www.mytherapy.tv/lookstore/
http://store.pixelfilmstudios.com/plugin/fcpx-lut-cinematic
http://www.rocketrooster.ninja/
http://www.technicolor.com/en/solutions-services/cinestyle
http://vision-color.com/osiris/


The problem is: you have all these companies selling (or in some cases giving away) so-called "film-look LUTs," but none of them agree with each other even on specific film emulsions. And they don't disclose how they create the LUTs in the first place, nor do they reveal any of their measurements, tests, and color science (or pseudo-science) required to actually create a LUT. With the exception of Nick Shaw's Antler Post, very few of these companies know anything about anything. They're marketing sizzle without the steak. 90% of it is total bullshit, just guesswork at best. And I've used 3D LUTs from Kodak, Technicolor, and Filmlight, and those guys are about as expert as anybody can be on color science. Yet even those were not 100% reliable and did not agree with each other.

I have to say, I admire what Cinegrain is doing right now: they're marketing what they simply call "looks" and not attaching it to a specific film type, and they provide a LUT and the actual Resolve grade so you can dig inside it and see what they're doing. This makes a lot of sense, and I don't mind this since it's nothing more than an interesting look, one which may or may not work for the style of work you're doing. But they don't claim it's Kodak or Fuji or anything else.

http://www.cinegrain.com/

Note there are free LUTs offered inside Baselight and Resolve, and those don't agree with each other, either. The bottom line is that you can get a great starting point just by using a great colorist and coming up with a new look. Or just by spending time and turning the knobs.

There are times where you definetely need to try and match film. Whenever you shoot both digital and film you need precision for it. A colorist that can match perfectly to film would cost a whole lot more than using the LUT as a starting point for the grade. This isn't rocket science and it's not true that you can't match digital to film, you just have to shoot it without clipping and then tinker a bit with the settings of the LUT.

No. You just have to know what you're doing. The LUT has no way of compensating for how well or how badly the material is shot. I have personally matched digital to film (and vice-versa), and it's not that big a deal. It just takes time and experience. A button won't do it, and there is no "one-size-fits-all" adjustment that will work, either.

 

[PatrickFaith]

I think the foundry does a good job at this, but it's takes a few years of experience to do it right. For most A level vfx work that is done using film(which there is a lot of), they de-grain in nukex and seamlessly regrain the vfx part into the frame.

 

[Andi Obarski]

Thanks everyone! You've been super helpful.

 

[Marc Wielage]

I think the foundry does a good job at this, but it's takes a few years of experience to do it right. For most A level vfx work that is done using film(which there is a lot of), they de-grain in nukex and seamlessly regrain the vfx part into the frame.

Absolutely. Grain Management is still a huge part of filmmaking, and some mild noise does take the "edge" off of VFX. But it's more than just a gamma curve, a color curve, and grain.

 

[NateWeaver]

Marc,

VisionColor and Impulz, I asked exactly how they made their LUTs, and the creator Jonathan Ochmann got back to me.

Suffice to say he made a lit box with a ~30,000 color chip chart, shot the chart under controlled strict conditions in the chip chart box with the films in question, got the film scanned, and then read out the patch values with custom software. He went into many, many more techie details in the email and asked me not to pass them on. Stuff about how out of gamut, DR beyond target gamma, etc, many things I hadn't considered.

Now, he had lots of choices as to how the data was scaled, adapted the resulting LUTs to the limitations of the various imput color spaces and gammas, but the core data set from the scans is solid, in my opinion. He was upfront about the nature of these subjective choices he had to make.

He set out to as faithfully model various films' color response, and while not probably perfect, unless he literally made up the whole scenario and is lying then I think there is real value in what he did. Using the LUTs for last year or so, my eyes like it. And I know for sure I could not replicate what the Impulz LUTs are doing by hand. I could replicate the broad strokes, but using and playing with them for 18 months...they are much more complex and have many more subtleties than they let on in first blush.

The other manus...in my opinion they are just looks based on coarse low sample color charts. If that.

It IS possible to model such things in math. Since film is about to become very difficult to get ahold of and shoot, I support any and all efforts to model film, in LUTs or otherwise.

Besides, this is exactly what Arri did with the color science in the Amira, and what I imagine Graeme did with RedColor. I'd bet my arm Arri started with film characterizations created with the Arriscan units and tweaked to make work.

 

[Marc Wielage]

Besides, this is exactly what Arri did with the color science in the Amira, and what I imagine Graeme did with RedColor. I'd bet my arm Arri started with film characterizations created with the Arriscan units and tweaked to make work.

Actually, Glenn Kennel at Arri is one of the main people who came up with the color science on the D20 and D21 cameras, leading up to the Alexa. Glenn is also the guy who was at Kodak for many years, part of the team that created the Cineon format and system back in 1992. There's nobody on the planet who knows more about color science than he does.

Again, my observation is that all these so-called LUTs are different, so they're inherently inaccurate -- both with each other and with actual film. I also believe you can replicate any look just with any top-grade color-correction system and an experienced colorist. In fact, you can actually do a lot more with a correction, since windows, keys, and blurs aren't possible in a 3D LUT. But you can't sell that for $49.95.

One other big variable: how the film was developed and what the specific chemistry was. I've seen two pieces of negative come out of the lab that were 20% different from each other, shot 10 minutes apart from the same camera and lens. Was it Kodak's fault? Different emulsion batches? Different exposures? Different chemicals? Different temperatures? It's all of the above. Film itself is not always that consistent.

I find a lot of people trying to chase "the film look" have never actually shot film, so they don't quite grasp what it is they're looking for. Juan Salvo from The Color Space/NY has some interesting thoughts that I think apply here:

We’re often asked what cameras we prefer to work with as colourists, or whether we prefer to grade with film or digital sources. The answer is usually the same: the best image is the one captured by a talented cinematographer, under careful lighting, with good art direction, properly exposed and white balanced. That’s true of many professional cameras, even some not costing very much (BMD makes a few). In an ideal world, from our perspective as colourists, the camera “burns-in” as little of the look as possible, and just captures true to life information. The truth is, if the camera has enough information, the film “look” can easily and very effectively be added by a talented colourist. In other words “the Colourist is the new film stock.”

That to me is the right point of view. It's not about the LUT or a plug-in -- it's about the DP and the specific choices made with the colorist. The LUT (or no LUT) is trivial.

https://www.facebook.com/theColourSpace/posts/1618603668420131
http://liftgammagain.com/forum/index.php?threads/film-vs-digital.5232/#post-46560

 

[luigivaltulini]

In other words “the Colourist is the new film stock.”

https://www.facebook.com/theColourSpace/posts/1618603668420131
http://liftgammagain.com/forum/index.php?threads/film-vs-digital.5232/#post-46560

Yep MARc,
this be a good point,I completely agree.
Today the colorist be crucial, also easy to find people who think they are colorists, because they use Davinci, and sometimes do more damage than normal.

 

[Michael Tiemann]

[...]

You just have to know what you're doing. The LUT has no way of compensating for how well or how badly the material is shot. I have personally matched digital to film (and vice-versa), and it's not that big a deal. It just takes time and experience. A button won't do it, and there is no "one-size-fits-all" adjustment that will work, either.

Marc's excellent point ties together with a point referenced in another thread, namely that everything (good) comes from the (right) starting point. Citing the reference from that post,

[In 2004,] Viggiano found that white balancing in the camera's native RGB tended to produce less color inconstancy (i.e., less distortion of the colors) than in monitor RGB for over 4000 hypothetical sets of camera sensitivities. This difference typically amounted to a factor of more than two in favor of camera RGB. This means that it is advantageous to get color balance right at the time an image is captured, rather than edit later on a monitor. If one must color balance later, balancing the raw image data will tend to produce less distortion of chromatic colors than balancing in monitor RGB. [See https://en.wikipedia.org/wiki/Color_balance#Color_balance_and_chromatic_colors]

What the heck does White Balance have to do with Film Looks? It has to do with starting from the right place. I suspect the reason that there are so many film looks out there is because everybody has their own concept of how they work with film, and that concept dictates the subsequent behavior of the stock. Those looks are right for the people who start with the same approach and intention. But there is no universal answer, because the math falls apart exponentially the further one gets from the explicit (or, if not defined, de facto) starting point and notional direction. The better you can get it right in camera, the better everything else is from that point forward. But doesn't that contradict the value of RAW?

RAW is not a panacea, neither for White Balance nor for Film Looks (nor for exposure, contrast, etc). It does provide unprecedented latitude for correct or creative decision-making, but it does not provide unlimited latitude. Pick the right starting point and the right direction you want to take it, and RAW (and the right OLPF) will give you a more-or-less comfortable margin of error. Which is huge! But it does not completely liberate you from picking the starting point and direction, as Marc said.

 

[Marc Wielage]

I agree with Michael -- getting the right exposure and white balance/color temperature up front does help with providing a reasonable starting point. Without that, no faux LUT will give you the look you want.

 

[Andi Obarski]

Michael,

That is precisely what my director and I have been doing. A couple of updates since I started this thread... we've been looking at the stocks in Filmconvert (a program that my director and I decided upon by the timing and budget of our production) and have decided precisely what exposure value we want to assign to characters in differing scenes to give us our desired look. Again, shooting in DragonColor and REDlogfilm

While this is a somewhat bastardized version of what higher budget productions do, we're then emulating a similar look based off of camera tests with color charts in REDcineXpro and are setting a look preset to the camera based what we've built.

Will keep you updated on other practices we've been employing.

 

[Marc Wielage]

That is precisely what my director and I have been doing. A couple of updates since I started this thread... we've been looking at the stocks in Filmconvert (a program that my director and I decided upon by the timing and budget of our production) and have decided precisely what exposure value we want to assign to characters in differing scenes to give us our desired look.

My opinion is that you'd be better off having a colorist dial in the look you want instead of going with an off-shelf LUT or look. Filmconvert won't give you nearly the range or choices that another creative person can provide.

You're eventually going to have to get the project color-timed anyway. Have an experienced colorist show you some looks for an hour, and I bet it'd be very much worth your time. The looks may or may not be appropriate for the final, but at leat they'll give you some possibilities at what could be done on down the line.

 

[Ulf Krentz]

Marc, yes I absolutely agree. Out of curiosity I bought the bundle from vista-color and I have to admit I would have better spent the amount on some food and drinks...

Cheers

Ulf

 

[David Kellermann]

Has anyone tried CineGrain's LUTs?

 

[Björn Benckert]

Koji have a nice selection of luts and also camera conversion luts. Basically so you can have a mix of arri, sony, red, BMC and so on and apply log2lin luts with a selection of film stocks. Does not cost much and they, as I understand, work in 16 bit float which is good. They come with a lut conversion thingy so easy enough to convert to what ever format needed cube, 3DL or what ever is preferred.

 

[fred gasc]

..., as I understand, work in 16 bit float which is good. They come with a lut conversion thingy so easy enough to convert to what ever format needed cube, 3DL or what ever is preferred.

Bjorn,
I don't understand this (sorry for my ignorance but I'm on my learning curve).
Does a Lut have to be 16 bit float compliant?
I thought that once we promoted to a 16 bit float, all manipulations were
done within it.
This is a point I don't get.

 

[Marc Wielage]

Has anyone tried CineGrain's LUTs?

I believe Cinegrain is in the process of adding some new grades to go with the LUTs, and I think these will be useful in Resolve. To me, a multi-node correction is a better deal because then you can change separate parameters at different stages and have more control over the effects. Too often, a LUT slams the image too hard, and I think you need something much more subtle most of the time.

In truth, for that, you need a good colorist.