yves d
Well-known member
Are there any current threads that talk about this? Wondering what the current process is like to getting the closest look to 16mm/35mm including post production as well.
Welcome to our community
Be a part of something great, join today!
Emulating analogue film look on dsmc1 epic dragon?
- Thread starter yves d
- Start date
Mike P.
Well-known member
Nothing thorough/step-by-step. Lots of chatter about film LUTs and plugins, but they're typically not nearly as good as their claims/examples. And there's no shooting methodology (which you'd think there should be to get consistent or as accurate as possible results.)
One of the things I liked about Steve Yedlin's resolution discussion was just how well done his (proprietary/secret sauce) film emulation was. Granted it sounded pretty subjective/to taste, but he's had so much experience with film that the results just had the juice... Like when he switched back and forth between Alexa, RED, Sony, 35mm and 65mm, it was very difficult to tell which was which (and if memory serves, I seem to recall him saying other than adjusting brightness, they were blanket applications of his emulation). Pretty in-depth tweaking though.
One of the things I liked about Steve Yedlin's resolution discussion was just how well done his (proprietary/secret sauce) film emulation was. Granted it sounded pretty subjective/to taste, but he's had so much experience with film that the results just had the juice... Like when he switched back and forth between Alexa, RED, Sony, 35mm and 65mm, it was very difficult to tell which was which (and if memory serves, I seem to recall him saying other than adjusting brightness, they were blanket applications of his emulation). Pretty in-depth tweaking though.
Patrick Tresch
Well-known member
Noise texture is interesting in Dragon. You could test some Iso settings to expose your raw negative in the noise field. Exposing precisely will be critical to obtain consistent noise texture throughout the production.
Antony Newman
Well-known member
The following article may give some insights about what film does to color:
https://www.provideocoalition.com/panavisions-new-dxl2-monstro-dialed-11/
LightIron / Panavision have developed an IPP2 pipeline with FilmEmulation LUT and their DXL2 cameras.
As the DXL2 OLPF alters the colors compared with the RED OLPF - the following might help isolate the filtering effect of the DXL OLPF, which could then be couple with the LightIron Film Emulation LUT.
+) Set WB in Dragon : Record a clip of an X-Rite color checker under you expected lighting (Tungsten?).
+) Resolve : Set Project to Unmanaged Color : Timeline to Rec 709 Gam 2.2 (or 2.4)
+) Drag in Clip into Timeline : Apply PANA2 LUT (33 or 64 cube) - IE NON film emulation LUT.
(I think the PANA2 LUT in this forum converted from RWG.3G10 -> REC709 with MedCon MedRollOff and Gam2.2 (or 2.4))
+) Create a sequential Node : In that node - use Resolve color check match tool on frame of the X-rite color checker.
If I have guessed correctly - the WB node is LUT that accounts for the difference of the Panavision OLPF vs the OLPF that Red favours (ie a White Balance that will take a hint of red out of the image if only the PANA2 LUT is used).
+) Next : Replace the PANA2 LUT with the LIGHT IRON Film Emulation LUT.
If this LUT (not used it) also converts into REC709 - I am guessing you can now play with the Light Iron film emulation on Dragon footage (or if not, hopefully someone who knows this subject can chime in!).
If however it leaves the colour space in RWG.3G10 : Download the IPP2: RWG_Log3G10 to REC709_GAMMA22
Test tes test
<REMOVED - BECAUSE REDUSER.NET NOT ACCEPTING FULL MESSAGE>
AJ
https://www.provideocoalition.com/panavisions-new-dxl2-monstro-dialed-11/
LightIron / Panavision have developed an IPP2 pipeline with FilmEmulation LUT and their DXL2 cameras.
As the DXL2 OLPF alters the colors compared with the RED OLPF - the following might help isolate the filtering effect of the DXL OLPF, which could then be couple with the LightIron Film Emulation LUT.
+) Set WB in Dragon : Record a clip of an X-Rite color checker under you expected lighting (Tungsten?).
+) Resolve : Set Project to Unmanaged Color : Timeline to Rec 709 Gam 2.2 (or 2.4)
+) Drag in Clip into Timeline : Apply PANA2 LUT (33 or 64 cube) - IE NON film emulation LUT.
(I think the PANA2 LUT in this forum converted from RWG.3G10 -> REC709 with MedCon MedRollOff and Gam2.2 (or 2.4))
+) Create a sequential Node : In that node - use Resolve color check match tool on frame of the X-rite color checker.
If I have guessed correctly - the WB node is LUT that accounts for the difference of the Panavision OLPF vs the OLPF that Red favours (ie a White Balance that will take a hint of red out of the image if only the PANA2 LUT is used).
+) Next : Replace the PANA2 LUT with the LIGHT IRON Film Emulation LUT.
If this LUT (not used it) also converts into REC709 - I am guessing you can now play with the Light Iron film emulation on Dragon footage (or if not, hopefully someone who knows this subject can chime in!).
If however it leaves the colour space in RWG.3G10 : Download the IPP2: RWG_Log3G10 to REC709_GAMMA22
Test tes test
<REMOVED - BECAUSE REDUSER.NET NOT ACCEPTING FULL MESSAGE>
AJ
Last edited:
- Moderator
- #5
- Joined
- Apr 19, 2007
- Messages
- 13,392
- Reaction score
- 796
- Points
- 113
- Location
- Los Angeles
- Website
- www.phfx.com
For around 20 years now I've been doing film emulation on features mainly. It was actually much harder when the digital cinema cameras (and early other digital options) didn't capture a lot of dynamic range and way more noticeable was the incredible lack of captured color. Things changed in the 2010 era approximately as cameras got much, much, much better.
There's interesting levels to it and what you're end goals are in terms of the aesthetic need to be kept in mind.
I generally tackle it as what level are we going to. Full boat emulation, as if you were watching a film on a film projector in a theater alludes to emulating things like gate weave and potential flicker even.
Halation is an interesting one as I think probably about 1/2 the emulation work I've done has called for it. It's a real divider as to if people truly want it or not. Some love it, some don't.
Similarly grain. I have grain profiles from about 8 decades of film stocks. Which is weird. Early on, as in the earliest days of digital color our main interest was to match digital captured or created elements and seamlessly integrate them with the film stock they were mainly captured on. We would shoot that out to our intermediate stock via the LUX, Solitaire, and eventually Arrilaser. Grain is another one that's really "to taste" these days. Back then though, it was about pin point matching, hence the profiling. The older stocks I profiled mainly were do to 2K and 4K restorations of a few notable films, which was a treat.
Film has a lot of variation due to the host of variables that can occur in processing. Even fairly straightforward processing. But, the somewhat straight shot emulation of modern stocks is a fairly known color target.
The most interesting change in the workflow last decade was some films shot on film weren't printed on film. i.e. We were only dealing with the color and tonal response of the negative stock and not the unique fingerprint of the print stock. More productions are actually doing this now that aren't destined for a mass film theatrical release.
I had a great long chat with Steve and a few others over the years about this topic, but the main question that keeps coming up is:
"What do you want your film to look like?"
And that really is a concept at the core of whatever medium you are capturing on.
That all said, there's a host of decent LUTs and Plugins that get you most of the way there. But interestingly as somebody who's done many Vision 3 and 2 emulations, you do end up with some variety outside of the standard Kodak hero target. Especially if you start taking into account some level of creative aesthetic outside of a straight shot emulation. Sort of highlighting the reason why I just can't make one LUT, because it's more like a jar of Jelly Bellies. They certainly all are jelly beans, but they do all taste differently and even the same flavors sometimes have color variation.
Generally aiming at some sort of Kodak target with a decent print curve seems to be what many are doing these days when I look at the usual post houses. And there's a lot of variety in there too in terms of tone curves. And much like fashion, a lot of color work out there "feels the same" as popular or common looks penetrate the industry.
There's interesting levels to it and what you're end goals are in terms of the aesthetic need to be kept in mind.
I generally tackle it as what level are we going to. Full boat emulation, as if you were watching a film on a film projector in a theater alludes to emulating things like gate weave and potential flicker even.
Halation is an interesting one as I think probably about 1/2 the emulation work I've done has called for it. It's a real divider as to if people truly want it or not. Some love it, some don't.
Similarly grain. I have grain profiles from about 8 decades of film stocks. Which is weird. Early on, as in the earliest days of digital color our main interest was to match digital captured or created elements and seamlessly integrate them with the film stock they were mainly captured on. We would shoot that out to our intermediate stock via the LUX, Solitaire, and eventually Arrilaser. Grain is another one that's really "to taste" these days. Back then though, it was about pin point matching, hence the profiling. The older stocks I profiled mainly were do to 2K and 4K restorations of a few notable films, which was a treat.
Film has a lot of variation due to the host of variables that can occur in processing. Even fairly straightforward processing. But, the somewhat straight shot emulation of modern stocks is a fairly known color target.
The most interesting change in the workflow last decade was some films shot on film weren't printed on film. i.e. We were only dealing with the color and tonal response of the negative stock and not the unique fingerprint of the print stock. More productions are actually doing this now that aren't destined for a mass film theatrical release.
I had a great long chat with Steve and a few others over the years about this topic, but the main question that keeps coming up is:
"What do you want your film to look like?"
And that really is a concept at the core of whatever medium you are capturing on.
That all said, there's a host of decent LUTs and Plugins that get you most of the way there. But interestingly as somebody who's done many Vision 3 and 2 emulations, you do end up with some variety outside of the standard Kodak hero target. Especially if you start taking into account some level of creative aesthetic outside of a straight shot emulation. Sort of highlighting the reason why I just can't make one LUT, because it's more like a jar of Jelly Bellies. They certainly all are jelly beans, but they do all taste differently and even the same flavors sometimes have color variation.
Generally aiming at some sort of Kodak target with a decent print curve seems to be what many are doing these days when I look at the usual post houses. And there's a lot of variety in there too in terms of tone curves. And much like fashion, a lot of color work out there "feels the same" as popular or common looks penetrate the industry.
Michael Tiemann
Well-known member
Much of the "film look" people want to emulate is the result of understanding the strengths and limitations of the medium, and the using good taste and technique to achieve the vision.
This has re-ëmerged in digital Cinema by people understanding and adjusting accordingly.
In neither case are people using the medium as the goalposts. They are using the medium to reach the goal, along with proper costume, set, lighting, make-up, etc.
If it's super-important that your medium be read as film, there are certainly ways to do that. But imitation, the most sincere form of flattery, will only get you so far.
This has re-ëmerged in digital Cinema by people understanding and adjusting accordingly.
In neither case are people using the medium as the goalposts. They are using the medium to reach the goal, along with proper costume, set, lighting, make-up, etc.
If it's super-important that your medium be read as film, there are certainly ways to do that. But imitation, the most sincere form of flattery, will only get you so far.
David Mullen ASC
Moderator
Sometimes it is more practical to decide what you DON'T want the digital image to look like, what characteristics are unique to digital that can be minimized.
Karim D. Ghantous
Well-known member
I recall that Yedlin said that you really need a 3D LUT to do any of this properly. I can't exactly explain what that means, but he explains it very well, starting with a 2D colour matching LUT, then expanding into 3D.
Some say that which still images can look exactly the same, the way the image 'moves' is not so simple to simulate.
Digital cameras still have flaws that you can't just fix with a LUT. You can fix gate weave with Resolve, however.
I suppose you could do a film-out, which saves you from the tyranny of choice. Just print to film, then scan. It's not cheap, though.
- Banned
- #9
Hrvoje Simic
Banned
Depends on which digital cameras.
With today's high quality cameras, that is incorrect.
And it's not "just".
It may seem as "just" when someone simply and effortlessly applies something which takes a long time a lot of work to develop.
That doesn't the fix the issues of digital origin.
Karim D. Ghantous
Well-known member
No, it's correct. Very few CMOS cameras have global shutters, for example. The ones I know about: the A9 (practically global, but not technically), the Venice, the FX9 (like the A9, very good, but not 100%), and the Komodo (which is going to kick serious butt in the market, including the DSLR/DSLM segment). Sony does make a TV oriented 2/3" 3-chip CMOS camera which as a global shutter, the Z750 which hasn't been released yet. But we haven't finished.
No matter how good sensors are, they still fail in extreme situations. I have not yet seen a camera that can capture lights properly. Film has always been good at this, except for CineStill, which is Vision3 with the Remjet removed. And there is the issue of movement, which is somewhat subjective. Super 8 still looks better than any digital camera, as unsuitable as it is for large screens. I'm happy to call a complete win for digital, but not today.
No, it does not. I do have an idea of how a 'virtual' film-out could work, but I am not going to look into it right now.
Nick Gardner
Well-known member
No film camera ever made had a global shutter so.........
Karim D. Ghantous
Well-known member
As good as. I think that the mirror shutter can be improved little bit, but that is a topic for another day.
- Banned
- #13
Hrvoje Simic
Banned
In a first paragraph you have a point regarding rolling shutter artefacts (distortion/half exposed frames in flashes).
But this is related to temporal properties of image capture and has nothing to do with LUTs.
Venice, FX9 and A9 don't have global shutter.
There is no "practically" regarding a type of shutter and something in between. Sensor readout speed and type of readout are two different things.
Regarding failing in extreme situations and what you have or have not seen - that sidetracks from what the statement was. Everything fails in extreme situations, that is not a flaw but a usage limitation. You stated that digital flaws cannot be corrected with a LUT. For tonally related flaws, and unnatural image properties coming from tonal response, yes they can be.
Obviously, motion handling is something else. And DR.