Topics

the elusive film look

Frazer Bradshaw
 

In light of the new 65mm film camera thread, I wanted to re-discuss (for the hundredth time?) the question of what’s so great about film, and inquire with those of you who are highly knowledgeable about imaging technology about a hypothetical:

Certainly, digital acquisition has gotten very good.  I might argue that it is just as “good" as film, but not the same thing, because aesthetics have changed with the advent of digital capture.  The manufacturers have crafted digital cameras that do a surprisingly good job of reproducing the world accurately.  But the kind of accuracy we have now isn’t something that is valued in film — we shoot film for its look and feel, not its perfect fidelity.

Digital cameras that shoot RAW have their electronics, processing and workflows geared to accuracy, and from there many of us try to take them back to a film look in the grade, sometimes through LUTs designed to accurately mimmic the film look.  But what would happen if the electronics, processing and workflow was optimized for a filmmic look rather than for accuracy?  For instance, could Kodak use imaging data from Kodachrome (or any other stock) to design a workflow starting with RAW sensor data that would accurately emulate the look of the original?  

I don’t know the answer to this question, but I wonder how much the love of the film look is really a love of something that is inherently related to film, or how much it’s something that has simply not been a priority for digital camera manufacturers. I hear the voices in my head saying “it’s cheating if you do it digitally”, but is there really some sort of “truth” that can be engineered through analogue processes that is fundamentally “better" than a look that is engineered through digital ones?  Maybe there is, but I don’t know.

When Kodak announced the new Super 8 camera I really wondered, why not build a digital camera that can process in-camera to mimmic any (or all) stocks that Kodak has ever made (assuming they have the records to be able to emulate it)?  A digital Super-8 camera — a retro looking digital camera that shoots ProRes that looks exactly like Ektachrome or Tri-X — seems like it would be a bigger winner in this day and age than what they were/are planning.

Cheers,
Frazer


       Frazer Bradshaw
   Director of Photography
Oakland, CA   415.517.3260

Mako Koiwai
 

Film is elitist … which is fine by me … ie. can’t just work off a monitor.

Fujifilm offers wonderful Fujifilm Simulations in their still cameras. Plus they have a generic Classic Chrome look.

Provia, Velvia, Astia, ProNeg, and Acros … with added B&W filters combinations for the Acros.

https://www.bhphotovideo.com/explora/photography/tips-and-solutions/introduction-fujifilm’s-film-simulation-modes?BI=572

One of the cool things about Mirrorless cameras is that they show you your exposure, filters, simulations, effects, Live.

Of course when shooting RAW the Simulation is only a Look. You can also Post process your RAW IN-CAMERA to see what different looks will look like, plus do things like Pull Process.



They recently introduced the Eterna look, specifically for out of the camera video, when Flog isn’t required.


makofoto, ret. s. pas, ca
Fujifilm Camera Evangelist!

JD Houston
 

On May 3, 2018, at 1:43 PM, Frazer Bradshaw <dp@...> wrote:

is there really some sort of “truth” that can be engineered through analogue processes that is fundamentally “better" than a look that is engineered through digital ones? Maybe there is, but I don’t know.
The exactness of the film look can not be achieved in current digital no matter what LUT you use.

There are two essential physical elements of film, one of which could be simulated at great cost, and the other of which
is just fundamentally different.

Film color is a subtractive color system that uses dyes and transparency to create color.
There is no digital projection system that is subtractive. There are LARGE swaths of colors in film
that digital can not do. (though you can argue there are many the eye could not distinguish).
Color LUTs are at best a close-enough trick that most of the time it will resemble the original.
If you recall any project timed on film and then done on 709, (P3 is just bigger), there are fundamental
differences in the color that may even lead to different choices in grading. But there are parts of
additive colors that are just not ‘nice’ to film transfers. It is like the difference between acrylic (digital) and oils (film)

The other intrinsic part of film is dye clouds suspended in gelatin. The images are captured in diffuse multi-layered
pointillism, compared to the strict either-or boxes of a sensor. You could simulate some of this if you had very
high capture resolution and diffused the recorded scene in a simulation of gelatin and dyes, but it is not worth
the trouble on low-res 4K images. (at least, low res from film’s point of view ;-)

Sometimes it is the impressionistic sloppiness of film that appeals to people. But when I watch
digital, such as the Sony LED wall at 8K, the images are just stunning in a way that film
could never have achieved.

Jim Houston
Consultant
Starwatcher Digital
Pasadena, CA

Mako Koiwai
 


The images are captured in diffuse multi-layered pointillism


***********


But in areas with generous exposure the dye clouds are formed by “Points" of grain that are replaced with dye that can Run Together, forming areas of even density … no longer points.

This can look nice and smooth in areas like bald skies but can also produce splotchiness.


Kind of a cool effect, or not … in extreme situations is where the over exposed areas can be smooth surrounded by very grainy underexposed areas, when using high speed stock, pushed.



makofoto, ret. s. pas, ca

jmann@...
 

“Film is elitist”

Finally we have the definitive resolution between film and digital.

Digital is egalitarian.

I feel so much better

John Mann
Film and Media Studies Program
Johns Hopkins University

Steven Breckon
 

On May 3, 2018, at 1:43 PM, Frazer Bradshaw <dp@...> wrote:

"But what would happen if the electronics, processing and workflow was optimized for a filmmic look rather than for accuracy?  For instance, could Kodak use imaging data from Kodachrome (or any other stock) to design a workflow starting with RAW sensor data that would accurately emulate the look of the original?"

I suggest watching Steve Yedlin's display prep demo. This is basically what he is doing with 5219 & Alexa through extensive sampling of the two mediums and some math in Nuke. 



Steven Breckon
Cinematographer
Los Angeles

Steven Breckon
+1-360-739-9635


On Thu, May 3, 2018 at 2:41 PM, Mako Koiwai <mako1foto@...> wrote:
On May 3, 2018, at 14:08, JD Houston <jdhouston@...> wrote:

The images are captured in diffuse multi-layered pointillism


***********


But in areas with generous exposure the dye clouds are formed by “Points" of grain that are replaced with dye that can Run Together, forming areas of even density … no longer points.

This can look nice and smooth in areas like bald skies but can also produce splotchiness.


Kind of a cool effect, or not … in extreme situations is where the over exposed areas can be smooth surrounded by very grainy underexposed areas, when using high speed stock, pushed.



makofoto, ret. s. pas, ca

Jan Klier
 

I would suggest that modern image processing algorithms are certainly capable of producing very sophisticated results. It seems hard to believe that it’s impossible to process digitally captured images when there are no data limitations (like clipping or roll-off) and achieve the ‘film look’. It’s not a limitation of technology, it’s a limitation of articulating in a technically accurate way what that constitutes that ‘look’.

It may also be an unfair comparison of the old film and the modern digital capture. There were many additional steps taken when developing the film that influences the look. For a recent discussion I looked up ENR processing and came across this article: https://theasc.com/magazine/nov98/soupdujour/pg2.htm. A better comparison would be ‘film + custom processing’ = ‘digital capture + custom grading’.

Which leads to the issue of LUTs. LUTs may be ‘close enough’ tricks, but in some way they are a black box to most people, which means we play with different versions and then like or dislike the result. When we’re chasing a specific look, the ideal situation would be that we can articulate what constitutes that look as attributes of color, saturation, and other measurable attributes which then could be fed into digital imaging processing algorithms to replicate. The crux seems to be that we’re unable to articulate that precisely. So we’re stuck with experimenting and throwing darts at a board with mixed results.

LUTs are mathematical transforms that will yield B for a given A, every single time. But the film look isn’t a single set of exposure conditions. Do we know that a given LUT would really provide the right results for a wide range of scenes? That does not seem certain.

If we’re not using LUTs we play with the controls available in our grading systems. From my experiments with looks such as the popular ‘bleach bypass’ it seems that one common attribute of popular film looks beyond roll off  are attributes like color contrast being decoupled from luminous contrast. The article above refers to CCE (color contrast enhancement). Since this is not one of our primary controls in most contemporary grading applications, it may not occur to us that this may be a critical variable to adjust.

I believe understanding and properly replicating the film look requires to do more extensive study of what the underlying attributes of the look are, not chasing a better LUT.

Jan Klier
DP NYC

On May 3, 2018, at 5:08 PM, JD Houston <jdhouston@...> wrote:

Color LUTs are at best a close-enough trick that most of the time it will resemble the original.

Mark Kenfield
 

But is scrolling through LUTs really any
different to screening tests shot on a variety of different film stocks?

The objective goal remains the same - finding the colour reproduction and general aesthetic bent that you think best suits a particular project?

Cheers,

Mark Kenfield 
Cinematographer 
Melbourne 

0400 044 500

On 4 May 2018, at 1:07 pm, Jan Klier <jan@...> wrote:


Which leads to the issue of LUTs. LUTs may be ‘close enough’ tricks, but in some way they are a black box to most people, which means we play with different versions and then like or dislike the result. When we’re chasing a specific look, the ideal situation would be that we can articulate what constitutes that look as attributes of color, saturation, and other measurable attributes which then could be fed into digital imaging processing algorithms to replicate. The crux seems to be that we’re unable to articulate that precisely. So we’re stuck with experimenting and throwing darts at a board with mixed results.


Jonathon Sendall
 

"But is scrolling through LUTs really any
different to screening tests shot on a variety of different film stocks?"

Not sure that's true. It's more that the film stock, as a choice, guides the
look of the film as a starting point, not a finishing point. The LUT is part of
the finishing point after shooting. I definitely will light to try and achieve
a look but I don't light and shoot to try and achieve a LUT. Maybe some
folks do.

Jonathon Sendall
DP, London

JD Houston
 


On May 3, 2018, at 8:07 PM, Jan Klier <jan@...> wrote:

LUTs are mathematical transforms that will yield B for a given A, every single time. But the film look isn’t a single set of exposure conditions. Do we know that a given LUT would really provide the right results for a wide range of scenes? That does not seem certain.

mostly.


I believe understanding and properly replicating the film look requires to do more extensive study of what the underlying attributes of the look are, not chasing a better LUT

Hi Jan, 

You are at the essence of why LUTs are (and must always be) limited.   They are a lattice sample of a color space
that is usually more complex than a fixed LUT can represent.  (even when you can get pretty close).

You cannot get to the perfect film look in current digital (or LED screen) projection. It is just a different beast.

It is not just scene exposure that varies. Every film LUT you have ever used represents a particular 
negative and print sent to a particular Lab on a certain day.  And then scanned on a certain scanner with
a fixed color response that is unlike either.  All LUTs are made this way as a representation of what a film simulation
might look like.  It is why hundreds of LUTs exist not just for film stocks, but also separate ones, in the day, for Tech, Deluxe,
Fotokem, etc.. for each emulsion batch and some with special processes applied and some without.  Sometimes these were averaged runs over months,
but more often a single snapshot was used for months even years to simulate film color on a projector.

The only people who had a completely accurate model of how the colors responded in film were Kodak and Fuji.
They didn’t let that out for very long (remember the Kodak Look Manager).  But even for them,
they hit the limit of color science that you cannot reproduce all of subtractive film in an additive color system.
This is as basic as it gets.  If people compared film-outs with digital these days, they would be shocked
about how bad certain colors are in digital (hello cyan and purple :-)

The ultimate film simulation LUTs would take a scene photographed by any camera and would be able to 
create a reproduction on a projected screen of exactly the same XYZ colors that a film print (or release print) 
could achieve.   The DI world got as close to this as was possible ( though there were sometimes problems )
because many clients liked to see the film running ‘butterfly-ied” against the film print.  Tough critics.

Now people are using ‘copies’ of film colors that aren’t really as close as the original film was, and few notice.
There isn’t really even a reason to limit Looks to film style colors.

To get better film looks, It is a technology change in projection that is really needed beyond even LED walls, and not just chasing finer sampled LUTs or a more
accurate simulation of the underlying film attributes.  Even if you had both of them to perfection, you still
would not get all the way.

Creative Looks though, the sky’s the limit. (that limit being 300 miles above sea level ;-)

Jim Houston

(so now I’ll have to put up my other title things, just so you have some confidence I am coming from the right place.)
co-chair, ACES Project (2004-2017), AMPAS
member of SMPTE Digital Cinema Standard Ad-Hoc Colorimetry Group (2000-2004)
color scientist (1986-present)
scanner/film recorder specialist  (1985-present)
DI producer (2002-present)



Philip Holland
 

As somebody who's made a fair bit of film emulation LUTs in the last 20 years there's a lot going on, and spent way too much time with a few film scanners, it's also much to do with what's not going on.


Specific stocks have specific characteristics.  The tricky bits are indeed how the medium responds to different lights and of course the anti-halation backing.  However, with a well crafted LUT balanced with a general exposure method and logic, you can get some astonishing results.


I'm more of a Kodak Vision man myself and that's where much of my deeper understanding comes from.  Specifically I'm truly fond of Vision 2 and 3 stocks as their unique differences are interesting and sometimes not exactly clear as to what's always better.  Vision 2 has/had a toughness about it, where Vision 3 sneaks out a bit of grace within all of that.


The key thing, and this is the biggest thing, is film stocks in general are about creating a rather deliberate limited palette.  I'll oversimplify it a bit for the sake of this example.  In terms of "true to life color" for things like blues, reds, greens, secondary, tertiary, and the often difficult for digital to reproduce near neutral pastel palette; it's not about rendering an accurate image, but rather the unique properties of the stock and chemistry that brings out it's unique interpretation of colors in general.  


It's also important to mention the unique tonal response various stocks have and how that relates to exposure practices and the added bonus of it's effect on color throughout the image.  That's where some of the more interesting concepts of filming with daylight or tungsten stocks and how you choose to work with them comes into play.  The additional variation is of course processing method in terms of pushing and pulling, bleach bypass, partial bypass, flashing, etc....


The most interesting thing that's happened for me personally is integrating film-like looks into a newer HDR driven world.  Interesting back at the old studio everything we did was in 16 bit, so on a sheer color depth level that's fairly straight forward, but with the newer higher nit displays and generally higher contrast ratios than we ever got out of a film print it's interesting translating what you think white or even black is.  In the world of film milky shadows were something we avoided like the plague for any feature I ever graded, but that's become rather popular for the lower contrast, often under key look that many find "cinematic".  Which is fine.  The world is your oyster when it comes to the variety of looks you can have.  It's interesting even seeing those look translated to HDR as well.  Peak whites are obviously something we usually avoid as well.  It's famously said a good painter never paints with pure white or pure black and many cinematographers do their best to work within the dynamic range of their capture medium.  That's where some of the over zealous concept of HDR seem counterintuitive to image making, however, with the promise of deeper bit depth and access to more color nuance, HDR certainly is exciting as hell.


Apologize for the wordy reply.  I'm dealing with scans at the moment from my next batch of looks and it's on my mind fresh to the crisp.


Phil


-----------------
Phil Holland - Cinematographer
http://www.phfx.com
http://www.imdb.com/name/nm0390802/
818 470 0623


From: cml-general@... <cml-general@...> on behalf of JD Houston <jdhouston@...>
Sent: Thursday, May 3, 2018 9:43:17 PM
To: cml-general@...
Subject: Re: [cml-general] the elusive film look
 

On May 3, 2018, at 8:07 PM, Jan Klier <jan@...> wrote:

LUTs are mathematical transforms that will yield B for a given A, every single time. But the film look isn’t a single set of exposure conditions. Do we know that a given LUT would really provide the right results for a wide range of scenes? That does not seem certain.

mostly.


I believe understanding and properly replicating the film look requires to do more extensive study of what the underlying attributes of the look are, not chasing a better LUT

Hi Jan, 

You are at the essence of why LUTs are (and must always be) limited.   They are a lattice sample of a color space
that is usually more complex than a fixed LUT can represent.  (even when you can get pretty close).

You cannot get to the perfect film look in current digital (or LED screen) projection. It is just a different beast.

It is not just scene exposure that varies. Every film LUT you have ever used represents a particular 
negative and print sent to a particular Lab on a certain day.  And then scanned on a certain scanner with
a fixed color response that is unlike either.  All LUTs are made this way as a representation of what a film simulation
might look like.  It is why hundreds of LUTs exist not just for film stocks, but also separate ones, in the day, for Tech, Deluxe,
Fotokem, etc.. for each emulsion batch and some with special processes applied and some without.  Sometimes these were averaged runs over months,
but more often a single snapshot was used for months even years to simulate film color on a projector.

The only people who had a completely accurate model of how the colors responded in film were Kodak and Fuji.
They didn’t let that out for very long (remember the Kodak Look Manager).  But even for them,
they hit the limit of color science that you cannot reproduce all of subtractive film in an additive color system.
This is as basic as it gets.  If people compared film-outs with digital these days, they would be shocked
about how bad certain colors are in digital (hello cyan and purple :-)

The ultimate film simulation LUTs would take a scene photographed by any camera and would be able to 
create a reproduction on a projected screen of exactly the same XYZ colors that a film print (or release print) 
could achieve.   The DI world got as close to this as was possible ( though there were sometimes problems )
because many clients liked to see the film running ‘butterfly-ied” against the film print.  Tough critics.

Now people are using ‘copies’ of film colors that aren’t really as close as the original film was, and few notice.
There isn’t really even a reason to limit Looks to film style colors.

To get better film looks, It is a technology change in projection that is really needed beyond even LED walls, and not just chasing finer sampled LUTs or a more
accurate simulation of the underlying film attributes.  Even if you had both of them to perfection, you still
would not get all the way.

Creative Looks though, the sky’s the limit. (that limit being 300 miles above sea level ;-)

Jim Houston

(so now I’ll have to put up my other title things, just so you have some confidence I am coming from the right place.)
co-chair, ACES Project (2004-2017), AMPAS
member of SMPTE Digital Cinema Standard Ad-Hoc Colorimetry Group (2000-2004)
color scientist (1986-present)
scanner/film recorder specialist  (1985-present)
DI producer (2002-present)



Geoff Boyle
 

Thanks Jim,

 

This is the problem, there is no film look.

 

There’s the look of a specific type of film, in a specific lab, in a specific scanner at a particular time.

 

Lets stop looking back and make the best looking pictures we can, however we can.

 

Cheers

 

Geoff Boyle NSC FBKS

Cinematographer

Zoetermeer

www.gboyle.co.uk

+31 (0) 637 155 076

 

From: cml-general@... <cml-general@...> On Behalf Of JD Houston

“Every film LUT you have ever used represents a particular 

negative and print sent to a particular Lab on a certain day.  And then scanned on a certain scanner with

a fixed color response that is unlike either “

 

David Hollander
 

Long time listener here but wanted to chime in with my experience so forgive my ignorance.
 
Please feel free to correct me if I am wrong but my understanding is that when ARRI acquired Cineon technology from Kodak when Kodak went out of business, they used that color science to reverse engineer film curves for the ARRI Laser Recorder.  They then took those algorithms to generate Alexa color science, adding some grain, and voila: the Alexa simulates film better than any other present scheme.  Having shot a lot of film, I am satisfied with what they have achieved as a digital analog to film.
 
That said, I remember one film project in the past five years where it was impossible to find a perfect candy apple red on a digital out to 35mm because according to both Fotokem and Kodak, that color was simply not possible to create on current film stock.  And yet digitally there was no problem.  And Kodachrome was not an option at the time (though it might be in the near future because of the EPA guidelines being rolled back by Trump).
 
So maybe we are doing better than ever before...
 
David Hollander 
DP/Producer
Ex Aaton Des Autres
Ex Panavision Hollywood 
 
Marfa, TX
 
 

Steve Shaw
 

For playing - https://www.lightillusion.com/free_look_luts.html
Film Emulation LUTs using data extracted from the film stocks using a very accurate film spectrophotometer.

Steve

JD Houston
 


On May 3, 2018, at 10:33 PM, David Hollander <fourthdensity@...> wrote:

Please feel free to correct me if I am wrong but my understanding is that when ARRI acquired Cineon technology from Kodak when Kodak went out of business, they used that color science to reverse engineer film curves for the ARRI Laser Recorder.  They then took those algorithms to generate Alexa color science, adding some grain, and voila: the Alexa simulates film better than any other present scheme.  Having shot a lot of film, I am satisfied with what they have achieved as a digital analog to film.

This falls into the category of Urban Legend.

There was none of the trade secrets about film in Cineon. 

Alexa and the ARRI Laser were independently created by knowledgeable color scientists and teams. (I know who they are ;-)


Jim Houston
Consultant, Starwatcher Digital, Pasadena, CA

Jan Klier
 


On May 4, 2018, at 12:43 AM, JD Houston <jdhouston@...> wrote:

To get better film looks, It is a technology change in projection that is really needed beyond even LED walls, and not just chasing finer sampled LUTs or a more
accurate simulation of the underlying film attributes.  Even if you had both of them to perfection, you still would not get all the way.

Creative Looks though, the sky’s the limit. (that limit being 300 miles above sea level ;-)

That makes sense to me from a technological point of view. So we have a color model and reproduction technology that has gaps in the overlap, and a transform (LUTs) that’s too simplistic and rigid to reliably get from A to B, with B being believable in a side-by-side comparison. 

And I also like what Geoff said: 

On May 4, 2018, at 1:40 AM: Lets stop looking back and make the best looking pictures we can, however we can.

There are some that are chasing the film look out of nostalgia, and they wouldn’t be satisfied unless it was a perfect match of their recollection.

But for most it’s about the emotional reaction. They want people to leave the theater and say that was a great film, some of which may say so because it reminded them of qualities they enjoyed in the past without being technical about it. It’s that non-scientific feel we’re aiming for, in a contemporary not a historical perspective.

Which is why I continue to think that we get stuck in technical details, which is of course easy because it’s measurable, or treat is a black box without trying to understand it. So much easier to debate things that have numbers, then things that are harder to quantify. But so many things also have changed on the audience side. We make films very differently in many ways and our audiences expect them in a ‘best looking picture’. Stories are faster paced, more dynamic camera moves, etc.

So if we focus on a technically accurate match on a pure color basis, we may be chasing the wrong end, and may be disappointed that if we only got that cyan in there, the audience would go ‘hoorah, finally a great modern picture that feels like those great old films’. And I say that in a high-level way. We may well find out it just was all about the cyan one day.

It will take the effort to decode the emotional elements of that film look to find a formula that is the ‘best looking picture’ that also uses some of these same emotional elements. And I don’t think there’s a single one. Ultimately it’s just a style, a style we just got very good at by means of the technology of the day.

(so now I’ll have to put up my other title things, just so you have some confidence I am coming from the right place.)

And this is why we all love CML. An impressive group of people with deep subject matter expertise.

I come from a very different world, which gives me a different perspective, but mixing a diverse set of backgrounds is what ultimately unlocks great discussions.

Prior to working with film cameras, I studied and worked for 20 years in computer science giving me insight into bits, algorithms, and what can be done, but also what the limitations are. Along the way I spent many years building diagnostics tools and large scale software and digging in very large databases to find patterns. Ultimately you need to get the data, but the bigger challenge is figuring out what a limited amount data from the field actually tells you what happened and why. You need to ask ‘why’ fives times to find the real answers.

After your field techs replaced 20+ drives in the same data center with similar issues, the immediate root cause was the drive design, but the root cause of the bigger picture and why only that customer had that issue had to do with culture of the executive team at that company and how they made decisions, to name one of many examples I got to work on.

Jan Klier
DP NYC

JD Houston
 


On May 5, 2018, at 8:37 AM, Jan Klier <jan@...> wrote:

But for most it’s about the emotional reaction. They want people to leave the theater and say that was a great film, some of which may say so because it reminded them of qualities they enjoyed in the past without being technical about it. It’s that non-scientific feel we’re aiming for, in a contemporary not a historical perspective.

In some ways, Kodak ran a 70 year experiment on how to make images look good in a theater for a general audience.  
Each iteration was judged by 100’s of DPs and the winning filmstock rose
up in usage to become popular and the foundation of the next generation of film negatives and prints.

In digital -- Software Looks, and LUTs are so fragmented that there can be little learning from one phase to the next.
While everyone can apply their ‘taste’ to the look of a show, it is difficult to carry on the results to a new generation
of filmmakers. So while everyone can ‘scratch their itch’, it can also feel like there is not progress, just variation.

When does a ‘Look’ align with an emotional reaction?  Well, that is the Art, and where none of the tech details matter much.
On the other hand, applying “Lift, Gamma, and Gain” (and sat/desat) leaves a lot on the table and there are entire universes of 
color reproduction outside of the standard toolbox that are not always explored.

I have a set of test images I have accumulated (and the Academy has as well) that have most digital cameras and many film negatives and have
judged the effects of tone rendering on those for years.  For some reason, the various film negatives still to this day
often look better than the results of the digital cameras. (and only a few look worse)   That is a mystery that I have dug
into a lot, and part of my conclusion is that there were very subtle and beneficial reproduction choices in film that have been lost (or straightened out)
in digital reproductions.  They are different things of course, but getting the last 5% of quality out of any system is
a real bear.

In my view, there is not some mystical and elusive quality that can never be known, but that using the searchlight 
of technology can reveal some of the “Secrets of the Masters” even when some of the questions to an Artist are
“now, why did you do that?”, the answer:  “Well it just felt right”.     

There is an illusion of control that LUTs provide because in theory they can take any number in and put out
any number. (the number of course is the color of the pixel)   But in practice, most people create LUTs exactly the same way — take a curve shape, apply it
to RGB primaries, bake into a LUT. [generalization — yes, I know not everyone does that but it is most common].
  It is not surprising then that many results are tweaking in similar ways. 

To answer Geoff,  to know where you are going it helps to know where you have been.  I am not trying
to have a discussion (that was oh so common) bemoaning the loss of film.  That’s just the way it is.
But rather how to capture the elusive quality and extend it into new realms.  Most film LUTs are just
snapshots of a now-becoming distant past. While they can provide nice cinema quality, they often don’t put their best foot 
forward into the (HDR) future, and that next step is what I am looking for.

Jim Houston
Consultant, Starwatcher Digital, Pasadena, CA

Bob Kertesz
 

In some ways, Kodak ran a 70 year experiment on how to make images
look good in a theater for a general audience.
Each iteration was judged by 100’s of DPs and the winning filmstock rose
up in usage to become popular and the foundation of the next
generation of film negatives and prints.
I suggest that had there been, over those 70 years, a dozen or two
manufacturers simultaneously making film negative, and had there been a
new and different stock to come out of many of them every 90 days the
way new cameras are being vomited out now, the Balkanization of film
production would have followed the current model we have in digital cameras.

It was Kodak having a stranglehold on the business (yes, I know Fujinon
and Agfa were there, but in comparatively very minor roles) that allowed
them to control most aspects, from when and what type of negative stock
was released to the final print stock the public saw. People involved in
production could reasonably rely on Kodak's consistency when planning a
year ahead, and 25 year old witless producers who thought the business
started the day they got into it did not go to trade shows and then make
stock selections that overruled the DP.

Kodak, making a not untypical business error, mistakenly thought they
were in the film business when in fact they were in the image delivery
business. They were insanely slow to react to the changes, choosing
marketing snark and hubris over nimbly refocusing their business, and as
a result fell on their face. To this day, they still don't get it.

Just as there are vinyl aficionados who despair at the 'sterile,
digital' quality of CDs and can't stand to listen to compressed MP3
crap, there will be those who pine for the look of film and will try to
emulate the chemical process that made it what it was, probably for one
more generation - when the producers, directors, and DPs who are over 40
now retire or die of old age, the nostalgia and desire for negative will
more than likely go with them.

-Bob

Bob Kertesz
BlueScreen LLC
Hollywood, California

DIT, Video Controller, and live compositor extraordinaire.

High quality images for more than four decades - whether you've wanted
them or not.©

* * * * * * * * * *

Michael Most
 



On May 7, 2018, at 8:54 AM, Bob Kertesz <bob@...> wrote:

 - when the producers, directors, and DPs who are over 40 now retire or die of old age, the nostalgia and desire for negative will more than likely go with them.

As always, Bob has a very uplifting way of expressing himself. The interesting thing this time is that I completely agree with him.

Jim mentioned that "the various film negatives still to this day often look better than the results of the digital cameras.” Define “better.” I think if you’re talking about either civilians or industry people under 40, you’ll generally find that given a choice, even in a blind comparison, between film and digital images of equal “quality” (whatever that means…..) they’ll pick the digital image every single time. What is “texture” to a “film person” is just noise and distracting artifacts to them. What is “rich color” to the film person is just dark and desaturated to them. One can’t use the terms “better” and “worse” without generational specificity, because visual media changes with available technology, and what is “attractive” to one generation is “dated” and flawed to another. No one technology or era has a monopoly on human perception, and while “classic” artistic principles generally still apply, the limitations of previous technology are instantly apparent to those who didn’t grow up steeped in it. In 2018, those who continue to insist that the film medium was “superior” to modern digital images with infinitely more resolution, much wider color gamuts, and no distracting artifacts are going to find themselves a very small minority.

Mike Most
On Location Services Director
Technicolor
Los Angeles, CA.

Mako Koiwai
 


What is “texture” to a “film person” is just noise and distracting artifacts to them


***************

Personally I love digital but I sometimes add grain/noise to my digi stills.

Thank you: https://www.google.com/nikcollection/products/analog-efex-pro/


I guess I cross the line between the two worlds … Hybrids being in …

https://images14.fotki.com/v1665/filedDsF/7ac54/4/43793/4909192/A12T.jpg

makofoto, ret. s. pas, ca

Previous Topic Next Topic