Jump to content

"The cinematographer of Knives Out wants to end the film-vs.-digital debate”


Alex Anstey

Recommended Posts

12 minutes ago, David Mullen ASC said:

There's no right or wrong way to be an artist, and not every artist values technical consistency.  Some are a bit more "rock and roll", seat-of-the-pants types who take risks and fall-off the cliff now and then. Chris Doyle is like that, for example. Of course, they are all working with directors who encourage risk-taking.

I was thinking of Chris Doyle, too, but I'm less familiar with his work.

  • Like 1
Link to comment
Share on other sites

  • Premium Member
6 hours ago, M Joel W said:

But I was always fond of DPs who pushed things a little more aggressively and experimentally. I love Kaminski's photography from the turn of the millennium. Either he's calculating everything and riding a razor's edge stacking so many diffusion filters and ENR and whatnot together–one thing to add contrast, another to remove it–or it's more intuitive and he's working from accumulated experience. I wonder what his sets look like, I suspect his lighting is higher key than it appears on screen.

Regardless, Kaminski's approach I think invites pushing things just a bit too far and hoping for happy accidents, whereas Yedlin's doesn't as much. And that's what I meant about ontology vs phenomenology–whether you hear the autotune in the voice of the singer, I'd like to think that its presence in the producer's arsenal dictates something in the end result regardless. I can't appreciate or distill the technical specificities of Steve Albini's exclusively analogue approach to production, but I'd like to think that I can still hear that his production sounds different texturally from someone else's, to extend the music metaphor.

I feel the same way. That’s probably a result of becoming interested in cinematography in the late 90s/early 2000s. The craft was a bit more punk rock back then. Nobody’s baking camera negative in the oven anymore...

  • Like 1
Link to comment
Share on other sites

13 hours ago, Matthew J. Walker said:

There's also people who think the Earth is flat.

So anyone who disagrees with your subjective opinion is in denial of basic facts? That’s a rather arrogant and close minded stance for a student to take.

Link to comment
Share on other sites

  • Premium Member
10 hours ago, David Mullen ASC said:

Film projection today can be all over the map -- I saw a screening of "The Love Witch" in 35mm and was disappointed to see the end credits fluttering / ghosting.

Out of time shutter. I have the same issue with my DeVry. It's easy to fix, but it's a pain. 

I saw "The Love Witch" at the Silent Movie Theater and the projection was stellar. 

 

Link to comment
Share on other sites

  • Premium Member
10 hours ago, Satsuki Murashige said:

I suppose he thinks the risks of getting subpar footage isn’t worth taking a chance on happy accidents. And he might be right - there are so many uncontrollable variables on set that it must give him peace of mind to know he can at least control the image down to the pixel level after it enters the lens. But I do think the level of control he’s reaching for is a bit like using autotune and click tracks. Sure everything is on-beat and in-tune, but it doesn’t swing as hard, at least to me. 

I agree wholeheartedly. He's only interested in the technical aspects of the look. I gotta say, his stuff does look very good tho. 

Link to comment
Share on other sites

12 hours ago, Stuart Brereton said:

So anyone who disagrees with your subjective opinion is in denial of basic facts? 

No. Let's not be silly.

12 hours ago, Stuart Brereton said:

That’s a rather arrogant and close minded stance for a student to take.

This one is a declarative statement I assume meant to be unanswered, but I'll respond to it anyway.

Yes it is. 

Link to comment
Share on other sites

On 7/17/2020 at 8:55 PM, M Joel W said:

It's an interesting discussion. 

It's almost a matter of ontology vs phenomenology.

Except I don't think you can separate the two in practice. What you're watching is the collective result of the process. And while two complex processes could theoretically return the same result, more often than not they don't. The Alexa shots in Jedi fooled me (they didn't fool everyone), but they were also shot and lit similarly to the film shots and presumably graded against an exact reference taken from other coverage. Many of the qualities Yedlin is emulating are dialed in by eye–if you watch the full demo, certain colors don't correspond surjectively between film and Alexa and halation, for instance, has to be dialed in based on the presumed brightness of the blown out source if highlights are truly clipped.

But these are almost irrelevant considerations relative to how differently a set is run when shooting one medium or the other.

I had a strange experience–I watched The General and Transformers in theaters with a similar audience and the train sequence elicited more of a response than any of the action in Transformers. There's a lot of expertly shot practical stuff in Transformers, but it's almost diluted by the addition of CGI and motion control calling into question what's real. The excellent CGI and compositing in a way detracts from what's real, but of course what's real contributes to the CGI and compositing... Transformers held up imo. Just an odd experience.

I remember being disappointed that the new Twin Peaks series moved to Alexa rather than film, but ultimately it's largely a show about the failures of nostalgia and, ironically, I think shooting on 35mm would have been a nostalgic failure. I agree Kubrick would have moved to digital (what do I know, though). And there's nothing wrong with that.

The biggest disappointment for me is that digital, especially Alexa, is primarily just different levels of a "film look" filter. A lot of earlier digital didn't try to look like film, and is more interesting for it. (From von Trier to Speed Racer to Miami Vice.)

Just because two processes can result in the same result doesn't mean they will. Usually the monkeys won't type Shakespeare. I think you have to go with what feels right. But for me, I'm surprised to find digital feels better than it used to. S16 is still particularly interesting to me. Makes me wish I had more money.

I still need to watch Irishman, but I'm not sold on de-aging. Then again I'm also not sold on the bald cap in Taxi Driver.

Larry Smith said that Kubrick would test all new digital cameras and at the end he would say "nothing tops film quality"

Edited by panagiotis agapitou
Link to comment
Share on other sites

On 7/17/2020 at 3:43 PM, Stuart Brereton said:

The year is 2249, and cinematographers gather for the annual Film vs. Digital debate, now in its 241st year...

Don't worry by 2065 the film fanbase will have died off.

  • Upvote 1
Link to comment
Share on other sites

On 7/18/2020 at 3:53 AM, Heikki Repo said:

For some reason I doubt that art lovers and auction houses are going to be filled in the future with such works. Why? Is it day dream like "nostalgia"? Not seeing the easiness of new digital way of making things? Perhaps. Perhaps the same is also true for seeing theater -- how tedious, those actors have to act every night, again and again! Why not just video it and show that same, perfected show every time? And why do people care whether or not something is happening "live" on television? Isn't the end product the same? Is it just the misguided feeling of belonging, being part of something that is happening -- even if you are in your own living room, separated by hundreds of kilometers of what is happening.

1. Well, digital art has to have its own display medium. And that would be based on blockchain technology. Have a look at https://superrare.co/.

2. Live media is important because it focuses our collective consciousness. We feel it, even if we don't even know how to spell 'consciousness'. ? That's a debate probably not suited to this site, though...

On 7/18/2020 at 10:39 AM, Tyler Purcell said:

From what I understand, nobody else has used Yedlin's "look" however. Until he sells it to other people as a plugin, there is really not much to talk about. Yes one man cracked the code, but he's the only one with the key. 

It doesn't matter - he shows us how he got his LUT etc. Experienced DITs and DPs and colorists should know what to do with his instructions.

On 7/18/2020 at 11:13 AM, Tyler Purcell said:

I've been traveling the country shooting a documentary on 16mm and 35mm. I can't tell you how many people have walked up to me and engaged in lengthy conversations about my project, just because it was shot on film.

Apparently, the key to attracting attention with digital cameras is a matte box. If you have one, people assume it's a serious production. If you don't have one, people just think, 'video camera'.

On 7/19/2020 at 3:37 AM, Simon Wyss said:

A very false image seems to have settled in some heads of what film technics were a hundred years ago.

I still don't know how the three-strip Technicolor process managed to keep all three strips aligned so well.

2 hours ago, panagiotis agapitou said:

Larry Smith said that Kubrick would test all new digital cameras and at the end he would say "nothing tops film quality"

If he were alive today he'd shoot either Monstro 8K or the new BM Ursa 12K. Just my guess, don't read too much into it. ?

BTW a good part of Yedlin's philosophy is correct. It's just data, all other things being equal. But he does not seem to acknowledge that digital, even the Monstro or the Alexa 65, cannot handle light sources. I have never seen a digital camera get light sources right. And this is odd because Monstro technically has the same DR as film.

So film is the superior medium to shoot at night, because that's where digital fails. Maybe the answer is to underexpose the poop out of digital in these situations. And when digital gets that part right, we might see a complete shift to 100% digital.

Of course you can get two images to look the same, as long as no data is missing between them. Duh!  This should be obvious. That's what LUTs do. But, there are other factors.

  • Upvote 1
Link to comment
Share on other sites

  • Premium Member
3 hours ago, Karim D. Ghantous said:

I still don't know how the three-strip Technicolor process managed to keep all three strips aligned so well.

The cameras have register pin movements. Interpositives are made from the original camera negatives on pilot pin movements printers. The imbibition takes place on endless steel belts full of pins that the print and the dye strips are placed on. Nothing can budge.

https://www.eastman.org/technicolor/technology/dye-transfer-printing

Link to comment
Share on other sites

  • Premium Member

I don't buy that the Alexa sensor (it's the same ALEV-III sensor in all their cameras) cannot handle bright light sources -- the sensor captures 14+ stops of total dynamic range and if you rate it at ISO 800, then you get 7-stops of highlight information, very similar to color negative film. I've been shooting on the Alexa for ten years now and haven't found it to be inferior to film negative in terms of overexposure information.

Link to comment
Share on other sites

  • Premium Member
4 hours ago, Karim D. Ghantous said:

So film is the superior medium to shoot at night, because that's where digital fails.

Not sure this logic tracks...

If anything, night exteriors and low light sensitivity are where digital cameras shine. Film negative needs a fair amount of exposure in the shadows to render detail without excessive grain, digital sensors need much less. It’s never been easier to get a usable exposure in existing light, something that would have required huge lights on condors on 500ASA film.

It’s been a huge benefit, not just for bigger budget films, but also for low budget projects and even Youtube bloggers, some of whom have been doing great night exterior work. You just can’t make those kinds of images on motion picture film, even wide open at T1.3 and pushed 1 stop. Film stills photography is a different story, since you can shoot long exposures.

Edited by Satsuki Murashige
*typo
  • Like 1
Link to comment
Share on other sites

On 7/19/2020 at 2:56 AM, Stuart Brereton said:

So anyone who disagrees with your subjective opinion is in denial of basic facts? That’s a rather arrogant and close minded stance for a student to take.

Rather arrogant to assume the earth is round, globehead!

Link to comment
Share on other sites

  • Premium Member
On 7/19/2020 at 5:58 AM, Satsuki Murashige said:

I feel the same way. That’s probably a result of becoming interested in cinematography in the late 90s/early 2000s. The craft was a bit more punk rock back then. Nobody’s baking camera negative in the oven anymore...

I do! ? 

  • Like 2
  • Upvote 1
Link to comment
Share on other sites

  • Premium Member
On 7/19/2020 at 12:09 AM, Satsuki Murashige said:

My understanding of Mr. Yedlin’s method only comes from watching his demos. But from what I could gather, the purpose of his footage prep method (which he does regardless of format) is to convert everything into a common starting point without any subjective grading. That’s his secret sauce. He processes the Arriraw footage to his preferred standard base look, and the 35mm DPX film scans to the same look, using his pre-built LUTs and effects nodes. He can only do this because he has based these presets on his sensor data measurements for the specific cameras and film scanners. Once the footage has all been brought to a common point, then he grades them the same way. 

If this is correct, then that’s a rigorously scientific way of working that’s very specific to him. Almost nobody else works this way (though I guess you could argue that labs and post houses should work this way). 

If you watch "color grading tutorials" on YouTube most of them are instructing that this is the only "right" method for grading ANY footage. 

To me it is one proof more that the whole point of the Internet is not making your own work or thinking with your own brain. But it is to endlessly copy other person's work and opinions to the point that there is no facts or opinions anymore and everything "fact" based is really just based on a single original opinion which is copied from a single person and then repeated and varied so many times that it has become a fact just because it is heard so many times now. 

So it seems that the fanboys have watched some Yedlin videos and then made their own theory that the Yedlin method is how color grading really works and everything else is false information and ignorance. When this is repeated enough it becomes a Internet fact and cannot be easily changed anymore. End up to the Wikipedia and so on. There is lots of other examples too from different fields

 

  • Like 1
  • Upvote 1
Link to comment
Share on other sites

Anyone can do what Yedlin did on their own, even in a punk rock way. The only difference is that Yedlin really dived deep into the code and developed his own personal plugin to do so.

Here's a DIY Yedlin method at high level (in regard to the concept for  and diving deeper:

  • Shoot color charts on various formats with the same lens, lighting, exposure, etc; the only thing that should change is the format.

  • Adjust each format to match either one format, proper values on the charts, or to an incredibly specific look

  • Export LUTs for each format

A deeper step between 1 & 2:

  • Bring all of those formats into a wide color gamut and wide gamma (hey, this sounds like ACES!)

An even deeper step:

An even deeper step:

  • Create your own debayering algorithm for the RAW sensor data

And so on and so on and so on...

What Yedlin does is that he keeps diving deeper and deeper into the levels of abstraction to more finely control what's happening under the hood. The above steps are my best guesses of what each level of abstraction is, obviously he'll know more (and so will a lot more colorists).

Film, digital, it doesn't matter. If you follow the method above, or whatever method to match cameras, you can ultimately fool the audience to think a digital camera is film and vice versa.

Link to comment
Share on other sites

20 hours ago, Simon Wyss said:

The cameras have register pin movements. Interpositives are made from the original camera negatives on pilot pin movements printers. The imbibition takes place on endless steel belts full of pins that the print and the dye strips are placed on. Nothing can budge.

https://www.eastman.org/technicolor/technology/dye-transfer-printing

Ah, okay, I think I get it. But the cameras still had pull-down claws, right? So there would have been some tiny bit of misalignment in the camera - correct? Anyway, now I can see how double-printed IPs can work. I assume it's a similar set-up.

20 hours ago, David Mullen ASC said:

I don't buy that the Alexa sensor (it's the same ALEV-III sensor in all their cameras) cannot handle bright light sources -- the sensor captures 14+ stops of total dynamic range and if you rate it at ISO 800, then you get 7-stops of highlight information, very similar to color negative film. I've been shooting on the Alexa for ten years now and haven't found it to be inferior to film negative in terms of overexposure information.

I've seen a few films that were shot on the Alexa, from Dallas Buyers' Club to Twin Peaks Season 3 (which was supposed to be shot on 35mm). They all fail the light source test. I believe the DR numbers - in fact I pay a lot of attention to numbers. I know, for example, that the Red One had 12 stops where the Alexa has two more. Dragon has one more again, while Monstro, I think, has one more than Dragon.

But, results trump numbers (either that, or film is measured incorrectly by everyone). Here is a quote from RedUser member Brian Boyer in a recent thread about the new BM 12K Ursa Mini Pro:

Quote

I've been going on about this for years but almost no one else seemed to care all that much. To be fair, I believe Graeme did try to address this in the color science when Dragon first came out but people complained that the color red appeared too saturated. I remarked I'd rather have a red blob where a tail light should be versus a white blob with a pink halo. Again, no one really cared and white blobs with colored halos where traffic, brake, tail, and neon lights are present has ruled the day with little protest from you lot. 

 

19 hours ago, Satsuki Murashige said:

Not sure this logic tracks...

If anything, night exteriors and low light sensitivity are where digital cameras shine. Film negative needs a fair amount of exposure in the shadows to render detail without excessive grain, digital sensors need much less. It’s never been easier to get a usable exposure in existing light, something that would have required huge lights on condors on 500ASA film.

It’s been a huge benefit, not just for bigger budget films, but also for low budget projects and even Youtube bloggers, some of whom have been doing great night exterior work. You just can’t make those kinds of images on motion picture film, even wide open at T1.3 and pushed 1 stop. Film stills photography is a different story, since you can shoot long exposures.

Not necessarily, especially outside, with the huge contrast between light sources and unlit bitumen.

I present an example of 5219, rated at 1250, pushed to 2000. Stop was between 1.4 and 2.  It's not stabilised and it's also shot in an urban environment, so there is a lot of light thrown around from all kinds of lights. This shows not just how good 2-perf can look, but how good cinema scanners are. Photo scanners are awful in comparison.

 

  • Upvote 1
Link to comment
Share on other sites

I'm hoping this whole debate will just calm down and people will relax about using either digital or film for the actual shoot. You want to do either, fine. There are costs to be considered for both - perhaps digital costs more in post if you are seeking a film look for instance. What is starting to interest me more and more is the idea of shooting on digital but making film prints. Or, let's say, shooting on 2 perf or 16mm possibly, doing a DI (is that what it's called?), and printing to 35mm anamorphic film for exhibition in cinemas with film projectors. I think that would be such a good look -- it would get around some of the photochemical/optical limitations of shooting on smaller formats and blowing up to 35mm or even 70mm. Okay, so not too many film projectors in current use but I'm sure they're stashed away somewhere, lovingly greased and protected from dust and awaiting their time to shine again. As I wrote a while back here, I found two myself in a tiny garden shed along a very rough track out in the wilds. 

  • Like 1
Link to comment
Share on other sites

  • Premium Member
1 hour ago, Karim D. Ghantous said:

I present an example of 5219, rated at 1250, pushed to 2000. Stop was between 1.4 and 2.  It's not stabilised and it's also shot in an urban environment, so there is a lot of light thrown around from all kinds of lights. This shows not just how good 2-perf can look, but how good cinema scanners are. Photo scanners are awful in comparison.

 

Yes, I've seen that sample before, and I think it looks great. Personally, I would be very happy with those images. But objectively, the shadow detail in available light night exteriors can't compete with even these consumer digital cameras. Regardless of your personal aesthetic preference, I think we can agree that you can't get a usable exposure under moonlight with 5219 at 24fps.

 

Link to comment
Share on other sites

  • Premium Member

I would say these sample frames from Evan Richard's collection of movie frame grabs  (https://www.evanerichards.com) really highlight the difference in formats. Both gorgeous, but very different in terms of shadow detail and texture. It should be obvious which is film and which is digital.

'Her' DP. Hoyte Van Hoytema

Her_175.jpg

Her_176.jpg

Her_369.jpg

Her_372.jpg

Her_512.jpg

Her_526.jpg

Her_535.jpg

 

'La La Land' DP. Linus Sandgren

LaLaLand_078.jpg

LaLaLand_079.jpg

LaLaLand_080.jpg

LaLaLand_178.jpg

LaLaLand_224.jpg

LaLaLand_273.jpg

LaLaLand_397.jpg

LaLaLand_442.jpg

  • Upvote 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...