James Cameron Says The Next Revolution in Cinema Is…

[David Mullen ASC]

I get so tired of hearing how digital cameras are advancing at the speed of light but yet it is assumed that lens manufacturers and film stocks stay stagnant. Does anyone really think that 16mm today looks like 16mm back in the 60s, 70s, etc? Vision3 stock on Super8 looks better than 16mm shot 30 years ago. S16 probably can be made to look better than 35mm shot 30-40 years ago. And the same digital technology that is suppose to be a "film killer" is helping the look of film through high detail scans like 4k. David seems to be skeptical of 35mm approaching 4k often, as he worded with thinking 500t stock is more like 3k. This is an opinion and can be nothing other. How can one really say at what point a film scan has ran out of useful information? Unless you log every pixel of resolution and compare the color values to a lower detail scan, it is just perspective and no more scientific than what is being said by the lens manufacturer that Dom mentioned.

 

Matthew, shoot your own tests and come to your own conclusions - if you shoot a line resolution chart, scan the negative at 6K, and manage see more than 4K off of a piece of 35mm movie film, then you may have a point... but over the years, I've seen plenty of resolution tests shot by many people and no one has shown me anything that shows a piece of 35mm color negative movie film resolving higher than 4K. Now maybe you'll say that that's just a lot of people's "wrong" opinions...

 

If there was publishable data showing that 35mm movie film, 24mm wide, could resolve 10K to 12K, don't you think Kodak would be splashing it all over the internet by now??? It's not just movie photography, go online and try to search for resolution figures for still photographers and see what you find out about.

 

Hey, I would love to think my 35mm work is resolving more than 4K, I really would. I want it to have more resolution. I'm shooting modern Vision-3 stocks, using prime lenses, shooting at a decent f-stop, etc. and I'm just not getting there, so what else is there for me to do? What am I doing wrong, Matthew? Tell me.

[John Holland]

Shooting Anamorphic with a say 250 iso must resolve a touch more than 4K.

[Jesse Suarez]

This is a film (cinema) theory debate.

 

Cameron and the like-minded want to move (their) cinema towards realism and immersion- 3D, higher frame rates, surround sound, etc. More expressionistic-minded folk would say that it's the things that makes cinema different from life that make it cinema (the early expressionists were not fans of color or sound).

[Matthew W. Phillips]

Hey, I would love to think my 35mm work is resolving more than 4K, I really would. I want it to have more resolution. I'm shooting modern Vision-3 stocks, using prime lenses, shooting at a decent f-stop, etc. and I'm just not getting there, so what else is there for me to do? What am I doing wrong, Matthew? Tell me.

 

Well, if we use non-scientific evidence, I and some others I have ran across have looked at 2k scans of Super 8 which shows that more detail in being resolved at 2k than 1080p. This isnt all footage of course, but well shot footage definitely shows an improvement in the 2k scan. My reasoning is that if Super 8 could truly resolve 2k detail at 5.68mm wide, then Super 35mm at 24.89mm wide is 4.38x more width and therefore should be capable of >8k resolving power. Maybe in practice this doesnt happen but I would like to know why this is and not just a reply like "oh, youre crazy there is no way S8 resolves 2k!"

[K Borowski]

400 film in 35mm still (the Gold amateur stuff which I'm sure is grainier than modern Vision) was measured to resolve 12 MP (about 4K, although the aspect ratio is skewed 24x36mm)in some magazine article I read back in 2003.

 

 

I highly doubt 500T resolves more than 3.2K (I'm pretty sure it's under, probably closer to 2.5-8K), and probably only the 50D, and now discontinued (Thanks Kodak :-/) 100T go over a true 4K 200T and 250D are hovering probably slightly above and below, respectively 3.2.

 

 

 

 

 

 

 

Sorry to spoil the party, but resolution is IRRELEVANT. You have lower screen resolution in theatres than ever before! 1500 lines on a 35mm print because it comes from a crummy 4th gen. copy of a crummy 2K file.

 

 

 

3D is 1.4K per eye, Technicolor 3D (35mm 3D system) is probably under a thousand (again through no fault of the film). Digital has brought the bar down for everything except S35 blowups and digital originated film.

 

 

 

 

 

What people SHOULD be worried about is dynamic range and color rendition. Film is still king here (even against the Alexa, sorry) but, oh yeah, if you do a crummy scan of your film, you lose all of said advantages and end up with a very expensive mess that looks worse than just shooting digital to begin with.

 

 

 

 

 

At work I spend a lot of time every day throwing out 50% of the resolution making 1:1 optical prints. Does it bother me? Hell no! The results still look beautiful, although I wish I had a 35mm contact printer. . .

[David Mullen ASC]

Well, if we use non-scientific evidence, I and some others I have ran across have looked at 2k scans of Super 8 which shows that more detail in being resolved at 2k than 1080p. This isnt all footage of course, but well shot footage definitely shows an improvement in the 2k scan. My reasoning is that if Super 8 could truly resolve 2k detail at 5.68mm wide, then Super 35mm at 24.89mm wide is 4.38x more width and therefore should be capable of >8k resolving power. Maybe in practice this doesnt happen but I would like to know why this is and not just a reply like "oh, youre crazy there is no way S8 resolves 2k!"

 

If this were true then a TV show like Game of Thrones or a movie on the big screen like Apocalypto, both shot on 1080P cameras, would have looked sharper and more detailed if they had shot in Super-8. Does anyone really believe that? Or look at the BBC Planet Earth series on Blu-ray and tell me that it would have looked more impressive if it had been shot in Super-8.

 

Look just shoot a resolution chart on Super-8 and just show us that is resolving 2K. Then I'll believe you.

 

I've shot plenty of Super-8 in my life, I have a good sense of what sort of detail I can get out of it.

 

This is where you get into the difference between theoretical discussions and real-world results using different cameras and technologies.

[Matthew W. Phillips]

Well David, can you at least concede that Super 8 can resolve 720p? If so, this would still put S35 at a theoretical 5.6k limit or thereabouts. Now I do realize that much of the S8 footage out there is 40t or 64t which is very slow stock. This may account for the differences but I don't think it is fair to say that S8 cannot even resolve 720p. Why not say it is just VHS resolution? :D

[K Borowski]

I wouldn't concede that for any CURRENT S8 stocks, personally, David. K25 and -40A probably did. . .

 

 

 

7201 or '45 would have as well, and that is probably why Kodak deliberately doesn't make them in S8, because they want to sell you 16mm film.

[David Mullen ASC]

Doesn't E6 Ektachrome 100D resolve as much as K40 did?

[K Borowski]

I don't think so David, although I guess I could be wrong.

 

That's probably the only current 8mm stock I'd want to use, now that Plus-X is gone. That or if there is still Fuji Velvia.

 

 

I think Velvia 50 was/is (are they still making it or just 100?) as good as or better than K40A. Kodak's E series lagged behind though,from what I've heard, anecdotally from still phootgraphers. I know that there is a chart up somewhere, and as it is stills, doesn't mention K40A, but K25 was the SHARPEST stock by a long shot of ANY of the reversal emulsions, even the best most current ones.

 

Pretty much all the stocks have only marginal resolving power when you aren't exposing I think it is 2,000:1 contrast ratios? Real world I want to say it's 1.8:1 average. So, again, shooting charts will tell you a great deal about how a stock CAN perform, but in the real world these differences are minimized.

 

I look to flesh rendition, how "clean" the medium is, and dynamic range. Oh, and how much fun I have using said media. Would never have gotten into film had it not been for the magic of the cellulose acetate strip and seeing photographic materials develop from nothing. Better or worse, without that magic, what is so special about it? I know exactly how the photographic process works, acids bases, plastics, silver halides, reducers, restrainers, preservatives, accelerators, bleaches, fixers, stabilizers, prebaths, stops, how all of it works, except maybe the more complicated dye couplers, but I haven't gotten jaded and burned out enough (I will admit to days where I get sick at looking at giant piles of film) to lose site of the magic I am so privileged to work with five days a week.

 

I wrote my first computer program probably before I developed my first roll of film, but I just can't get as into the new way. I think you will find it hard to find someone who can argue as to the beauty of the craft in the electronic realm. It's dull and unremarkable number-crunching, checking just as every other facet of the realm of digital electrical engineering and computer science.

 

 

Obviously cost comes into it here, I need to private message you about some scary things that are coming soon for film. And film, at least when you completely ignore archival issues, is always going to be more expensive (unless you are making contact prints optically for a 35mm-only run). I still think this expense is worth it, for the present and for future posterity.

 

If anything, with our move away from the theatres (where the quality has PLUMMETED in a decade as well) onto YouTube, FB, into the era of 2MP camera phone pictures being the only documentation of day-to-day events by the common man, 35-, 16- and even 8mm film are even MORE acceptable now for dramatic production.

 

 

I think I said this in another thread but watching some Star Trek episodes on my phone the other day, along with Twilight Zone, the photography, envisioning a B&W viewer in a mid '60s home, is completely appropriate again. Movies on such a small screen, or a compressed YouTube image, are really really hard to even see what is going on.

 

 

I think someone had told me we live in the new age of Commodus again: Anything goes. Quality is no longer king.

[Matthew W. Phillips]

Agreed Karl to a lot of what you said. I think there is a magic to film that is hard to explain. I used to have these same types of arguments with guitar players (of which I am one) about the sound of Vacuum tube amplifiers vs solid state amps. I always felt that tube technology sounded much better, warmer, fatter, and just was pleasing to the ear. I cannot prove it from a scientific or data standpoint but I hear it. I have no bias toward tubes other than they sound good. Same with film...why would I have a bias for an expensive format that requires patience on my part to send off to a lab and get telecined and sent back before I can edit it? Film is expensive and a hassle in this age of instant gratification. Im terribly impatient to begin with so I wouldn't bother with something like film if I didn't feel it looked good. Maybe digital has or will exceed the resolution...so what? Does that make it look better? I saw a comparison the other day of Alexa to Red to film. The Alexa looks, IMO, way better than the Red. It definitely held its own with film in the technical categories but for some reason it didn't evoke that wonder in me that film does. I honestly don't know what it is about film that makes me feel that way and I wish I did. I just know it is not there with any digital camera I've seen thus far.

[John Sprung]

This is where you get into the difference between theoretical discussions and real-world results ....

 

That reminds me of a favorite quote:

 

 

"In theory, there is no difference between theory and practice. In practice, there is."

 

 

-- Hmmm, instead of giving everybody the name, let's guess who said it. ;-)

 

 

 

 

-- J.S.

[K Borowski]

If no one guesses John, I will happily ake that one for my own.

[Marcus Joseph]

Has everyone seen the Zacuuto shoot out? I saw it in HD last night and I was thoroughly disappointed with how they must had processed the film, there was no way the results were as they turned out in regards to the over exposure test.

 

David, I actually thought Game of Thrones was shot on film, I was thoroughly impressed, only until later in the seasons did I think there were hints of digital and thought maybe they shot a bit of both. I was incredibly shocked when I found out it was Alexa.

[K Borowski]

Marcus, how they processed the film?

 

I had this discussion with a rep. from Arri about how "chemical film" is a misnomer. Why? Because processing is a very MINOR effect on the look, unless the chemistry is WAY OUT OF CONTROL. Even here, you can't cause major shifts in color, density, even with heavy pushing or pullling.

 

The big effects would be due to bad printing, SCANNING, or TELECINE. So much of what I see of "grainy film" is actually poor scan sharpening, aliasing artifacts called "grain aliasing" which hyperexaggerates grain structure because the scanner is having trouble cleanly reproducing a nonquantized, linear, even structure, just like it still has trouble with fine patterns in fabric, anything that isn't exactly aligned with the x- or y-axis of the sensor.

[Marcus Joseph]

There was a significant amount of grain that looked like it shouldn't have been there on the film stock they were using, I am not sure if they pushed the film stock or even how they scanned it. But it could have also been my perception cause the high-end digital cameras looked so clean? It was specifically in the underexposure test rather than the overexposure one, it just didn't look like film should have looked.

 

I am eager to see the next installments where it starts to get very technical, but personally I prefer the kind of test where they just shoot the same content repetitively (almost a short film length) as John Brawley put up on this forum. It makes you think of seeing all the formats in a much more realistic and applicable way where you can say that does work and what doesn't, rather than just a single locked off shot technical demonstration looking at dynamic range and no camera or object motion.

[Ben Syverson]

Digital will always look great... until you have highlights. Negative film can easily roll off 10 stops of highlights, but digital must clip to ugly pure white eventually, because the sensor is linear rather than log, like film.

[K Borowski]

It looked so clean? Well, sure there are no grain artifacts in unfocused highlight areas.

 

But what does a digital sensor do to achieve this? I saw an example where basically, the chip, in camera will sense where the edges are, then basically soften everything else where detail isn't important. My problem is, I want ALL of the detail sharp; it's all important.

 

 

Even the ARRI ALEXA does this. So you are getting clean imagery basically through selective softening of the image. The Alexa doesn't use software, some sort of filter instead, but I don't look for cleanliness, I look for sharpness, detail (not that I like excessive grain).

 

Just saw Black Swan in 1080i. You know what, it looked pretty good, even at 1.9K!

 

It's pretty amazing how a 6.4K scan downsampled to 3.2K or 4K, not on 500T pushed a stop looks so remarkably better than a 2K scan (maybe downsampled from 3.2 or 4K but still just a final 2), or better yet a contact print, looks so much cleaner. Where did the grain go, if you're supposedly going past the resolution of the film (and you are going past the resolution at 6.4K, definitely at 4 as well).

 

It's all the limitations of the CCD in the scanner that you're overcoming by making each scan site 1/4 as important to the overall look of the final image.

 

 

 

 

 

BTW, this is still photography, but I remember some article comparing high-speed film to low-speed flim. It said that a 6x7cm negative (I think it was 58mmx65?) shot on 400 film had roughly the same resolution as a 645 negative shot on 160 film (58mmx45mm). (3770 vs. 2610mm sq.) So just 1-1/3 stops of speed causes a 30% reduction in resolution due to grain, or at least a 30% increase in graininess, think the larger image area would still have a resolving power advantage along with reduced DOF. IDK if that progression continues every 1-1/3 stops, but, shooting at 500T going to 200T if the former is 2.5K, the latter should be somewhere in the realm of 3.6K. Going again to 50D (which, as daylight balance uses the more natively-sensitive & theerefore smaller blue layer is even finer than an equivalent tungsten speed) the equation holding true would bring it up to 6.2K. I find that number almost impossible, as it would mean 34 MP on a scope frame or 25 for a 13-perf. frame. Anyway, it is certainly over 4K when you get down to slower filmstocks.

 

The real question is: Why are the fastest, grainiest filmstocks always the ones compared against digital sensors shot at their prime ISO and lit to that speed? Why are they always tested at their cleanest color balance? Tungsten film is the grainier type of film when compared to daylight-balanced stocks. It is a flexibility compromise because the filter factor correcting daylight film to tungsten is 2-1/3 st.

 

 

I am a film man through-and-through, but howabout comparisons that are fair, balanced and informative? The digital camera manufacturers love to tout the flexibility of "not having to carry different film stocks", but I think having one chip, at its best at a single speed, color balance, and compromised in some way or another at all others, is a real weakness, just as long as you're comparing it to a production that doesn't have to juggle back and forth to avoid running out of one type of film ;-)

 

In a couple of months, I think I am going to commission my own test to honestly, objectively test all of the major stocks and all of the major digital cameras against one another scientifically with charts and controlled lighting and in real world lighting conditions. As long as I can make my money back at least, selling it at a nominal cost, I think it would help settle once and for all a lot of this BS and speculation and supposition that probably makes 30% of the posts on here silly;, pointless, noisy religious battles basically; based little on fact.

 

I'll even bring in an advisor from the digital camp to make sure that the digital cameras are shot at their bests as well. Anyone up for the job in the NYC area? :-)

 

 

Marcus, looking back at your post I see you say film looked noisy underexposed. It certainly gets very grainy very quickly anything more than 1 stop underexposed. But there is the additional problem of a scanner trying to gain boost this lack of information. You certainly do get ugly-looking grain on film with anything more than slight underexposure; digital wins here hands-down, just like film wins with overexposure hands-down.

[John Mastrogiacomo]

That reminds me of a favorite quote:

 

 

"In theory, there is no difference between theory and practice. In practice, there is."

 

 

-- Hmmm, instead of giving everybody the name, let's guess who said it. ;-)

 

-- J.S.

 

 

Yogi Berra. I'm a Yankee fan!

[Marcus Joseph]

Yeah I believe it was about 4 stops underexposed, but I completely understand your point there, the film was high speed, the film was tungsten. I guess they demand the film be as close to the digital sensitivity and native characteristics to digital as possible, but when in a real world scenario would you use a high speed stock for a scene involving forced overexposure? I suppose consistency would be one, but all of those elements would be taken into account in preproduction nonetheless.

 

I'd love to assist a large scale camera test, if only I was in NYC.

[K Borowski]

Yeah, sorry, you're just half a league, half a league, half a league too far away Marcus (probably, what almost 10K km?)

 

4 stops? I'm surprised there were any usable images salvageable, that there were is a testament to digital printing tools. Using a straight contact printer, it'd be impossible to get even close to normal values without extensive flashing, pushing, of print stock.

 

 

Aren't most digital sensors still about 200 ISO sensitivity? I think it'd be a very fair comparison to compare against V3 200T or Eterna 250 overexposed 2/3 st. each. :-)

[Marcus Joseph]

Yeah, sorry, you're just half a league, half a league, half a league too far away Marcus (probably, what almost 10K km?)

 

4 stops? I'm surprised there were any usable images salvageable, that there were is a testament to digital printing tools. Using a straight contact printer, it'd be impossible to get even close to normal values without extensive flashing, pushing, of print stock.

 

 

Aren't most digital sensors still about 200 ISO sensitivity? I think it'd be a very fair comparison to compare against V3 200T or Eterna 250 overexposed 2/3 st. each. :-)

Haha at the least, I think I have to move out to somewhere like New York or LA eventually.

 

I think the sensitivity is moving up (faster than the highlight information), I think the Alexa is rated at about 800.

[John Sprung]

Yogi Berra. I'm a Yankee fan!

 

Correct! I've never been much into sports, but I am a Berra fan.

 

http://www.yogiberra.com/yogi-isms.html

 

 

 

-- J.S.

[Ben Syverson]

I heard that Yogi-isms were mostly written by the sports writers of the day... Hence the classic Yogi-ism "I really didn't say everything I said."

[Alan Hill]

Doesn't E6 Ektachrome 100D resolve as much as K40 did?

 

 

Having just seen an HDNet broadcast of a Charlie's Angels episode, one thing regarding resolution comes to mind. Had ABC chosen to shoot that show digital in the 1970s they would have locked in those shows with 1970s technologies, namely NTSC and video tape. Thanks to some money spent by ABC, Farrah, Kate, Jaclyn & Cheryl in HD have never looked better. Man, those women were beautiful.

 

Sadly any discussion of resolution seems odd, considering that the primary demographic that advertisers want to buy (18-34, or is it 13-18) are mostly viewing content on their phones.

 

I used to calibrate my brother's large screen TV so that it accurately represented reality, as best as possible across the thousand or so satellite channels he had the option to watch. Only to come over a few weeks later and find that one of this young daughters had put the TV back into Vivid mode again. Claiming that the picture was looking dull and washed out.

 

So much for all that hard work that went into filming those programs. Not to mention what Satellite compression does to resolution...