Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

THE DYNAMIC RANGE OF 35MM MOTION PICTURE FILM......

R

Rohan B

Well-known member
Joined
Feb 18, 2011
Messages
371
Reaction score
0
Points
0
Hey guys,i really wanted to know about this,i searched about it on many forums and websites,but could,nt find a satisfying and straight answer.
Red one had a DR of 11.3 stops,RED MX has a DR of 13 stops,RED EPIC has DR of 13.5 stops natively and 18 stops with HDRx, ALEXA has a DR of 14 stops, but
what is the DYNAMIC RANGE of the latest offerings from kodak and fuji,what is the DR of 35mm motion picture film??????

I really want to know about this...........
 
Rohan B said:
Hey guys,i really wanted to know about this,i searched about it on many forums and websites,but could,nt find a satisfying and straight answer.
Red one had a DR of 11.3 stops,RED MX has a DR of 13 stops,RED EPIC has DR of 13.5 stops natively and 18 stops with HDRx, ALEXA has a DR of 14 stops, but
what is the DYNAMIC RANGE of the latest offerings from kodak and fuji,what is the DR of 35mm motion picture film??????

I really want to know about this...........
Click to expand...

The thread Kalani posted the link to is very informative.
In the last "big" test, the SCCE presented at NAB 2011, Kodak 5219 was said to reach 14.5 stops of latitude, but other film stocks were not tested, so there could be more, and it'd be interesting to see how or if the new Scanity system improves those numbers. The digital camera that got closer to film was Alexa (14.3 stops), while Red MX "reached" 11.9 stops and Sony F35 11.3. Epic was not tested.
If I remember correctly, the documentary about the tests will be available on Zacuto's website in a month or so.
 
Unfortunately the SCCE is not what I'd call a reliable test. For instance they've been quoting resolution figures for many cameras that go beyond what the sensor of those cameras are capable of - for instance 1080p cameras with more than 1080p of resolution. If you're going to test DR, do it yourself.

Graeme
 
Graeme Nattress said:
If you're going to test DR, do it yourself.
Click to expand...

Totally agree, Graeme, and I happen to believe that's something very true for every aspects of a camera or medium. I never doubt "official" specs or tests, nor I do take them as "The Truth", but they mean very little to me until I put my hands (and eyes) on the camera. That's the way I was taught and the way I learned, and that's the best way, at least for me :-)
 
I also feel there's a difference between "testing" and "measuring". I have to test and measure. In testing I'm saying "which do I prefer for this use", and in measuring I'm attaching a number to an attribute. Not always do numbers reflect "real-world" use. That can often mean you just need a different measurement for instance with resolution - limiting resolution is interesting, but it's a single number. Plotting an MTF curve is another kind of measurement that gives you much more practical information. For testing though, you'd use the precise lens that you're intending to shoot with, at the aperture you want to shoot with, and you'll get a much more telling result than looking at an MTF plot I did with my lens and my choice of aperture.

Graeme
 
Exactly Steve - they published resolution numbers that are utterly wrong. In LP/PH 540lp/ph is 1080lines/ph, which is 1920 vertical lines = 1.9k. So for any of the "HD" cameras, numbers greater than 540lp/ph are false resolution from aliasing: F35, F3. Good resolution tests from Alan Roberts show AF100 is a lot less than 500lp/ph. It's stated that "The lp/ph numbers are the average of horizontal and vertical measurements (although these were very close for all cameras measured)." but we know F35 and the vDSLRs all have very different responses horizontally than vertically and that's plainly obvious. Alexa raw at 848 equates to 3015, which is more than the 2880 the sensor provides.

Graeme
 
Rohan B said:
Hey guys,i really wanted to know about this,i searched about it on many forums and websites,but could,nt find a satisfying and straight answer.
Red one had a DR of 11.3 stops,RED MX has a DR of 13 stops,RED EPIC has DR of 13.5 stops natively and 18 stops with HDRx, ALEXA has a DR of 14 stops, but
what is the DYNAMIC RANGE of the latest offerings from kodak and fuji,what is the DR of 35mm motion picture film??????

I really want to know about this...........
Click to expand...

Unless you personally test everything you are comparing yourself using the same methodology so you can see comparative differences, the numbers are pretty meaningless. You can't take a figure from one person's testing and put it up against a figure from another testing and think these are absolute figures and draw any meaningful conclusions.

Besides, DR is a bit more complex than absolute numbers of stops, there is the way it falls off that matters, and how it does it in each color layer or channel, etc. Even when shooting a DR test, some people count grey steps visible in the noise (or grain) floor where it is likely that most of the time, you'd crush this area to black in color timing, which is another variable. It's one reason why the published DR figures seem so far off from real-world shooting DR ranges, which are generally more narrow.

I try to never takes published numbers very seriously, in my mind, it's all ballpark figures. The way I see it, film negative is between 14 and 15 stops and the best digital cinema cameras are close to that figure now, only about a stop short.
 
David Mullen ASC said:
Unless you personally test everything you are comparing yourself using the same methodology so you can see comparative differences, the numbers are pretty meaningless. You can't take a figure from one person's testing and put it up against a figure from another testing and think these are absolute figures and draw any meaningful conclusions.

Besides, DR is a bit more complex than absolute numbers of stops, there is the way it falls off that matters, and how it does it in each color layer or channel, etc. Even when shooting a DR test, some people count grey steps visible in the noise (or grain) floor where it is likely that most of the time, you'd crush this area to black in color timing, which is another variable. It's one reason why the published DR figures seem so far off from real-world shooting DR ranges, which are generally more narrow.

I try to never takes published numbers very seriously, in my mind, it's all ballpark figures. The way I see it, film negative is between 14 and 15 stops and the best digital cinema cameras are close to that figure now, only about a stop short.
Click to expand...

Hi david,you have been shooting on the red epic camera for quite some time now and have tested the camera in verry practical real life scenarios during ur shooting BIG SUR, if u have to rate the native dynamic range of red epic(without HDRx) in f-stops,how many f-stops would you give it........

waiting for ur replies......
 
David Mullen ASC said:
Besides, DR is a bit more complex than absolute numbers of stops, there is the way it falls off that matters, and how it does it in each color layer or channel, etc. Even when shooting a DR test, some people count grey steps visible in the noise (or grain) floor where it is likely that most of the time, you'd crush this area to black in color timing, which is another variable. It's one reason why the published DR figures seem so far off from real-world shooting DR ranges, which are generally more narrow.
Click to expand...
That's very true. We ran into these questions all the time at Technicolor just comparing Fuji negative and Kodak negative; the Fuji was technically faster, but the toe of the signal just fell off the edge of the world much faster, going right into the mud. The black fall-off slope in the Kodak negative was a lot more gentle and more forgiving. In fact, I'd say in general that comparable 35mm negative stocks were always more forgiving to exposure problems than digital.

I see similar things with different digital pickups. For example, when we were working on Supernatural, we had scenes shot with both the Arri D21 and an early Red camera, and the differences were fairly apparent. We were able to match them to the point where no one at home would be able to tell the difference, but it was a struggle.

The other thing I see with dynamic range in digital cameras is that the range isn't exactly linear. Highlights blow out and react differently than the mids, and the mids react differently than the blacks. Complicating this is the various LUTs (which I've often called the "LUT-of-the-month club"), different manufacturers' ideas of what "log" is, and inconsistencies in exposure. Bottom line, I don't think the numbers tell the whole story.

I am looking forward to the Zacuto documentary. I'd like to see the ASC jump into this and do a new shoot-out with all the major players, particularly the Epic.
 
Red mx only has 11.9 stops of dynamic range?????? That is strange,most of the tests i have seen show above 12.6 stops......
 
Rohan - I'd question the measurements of anyone who can't even measure the simple things, like resolution, correctly. To say that the Alexa in RAW has a measured resolution of 3015, which is greater than the number of recorded pixels on the sensor, horizontally 2880 is just wrong.

Graeme
 
Rohan B said:
Hi david,you have been shooting on the red epic camera for quite some time now and have tested the camera in verry practical real life scenarios during ur shooting BIG SUR, if u have to rate the native dynamic range of red epic(without HDRx) in f-stops,how many f-stops would you give it........

waiting for ur replies......
Click to expand...

I don't think in those terms. I don't show up on the set and think "well, I have 12.6 stops to play with today, how do I want to use them?" It doesn't work that way. No one measures the amount of stops for a shot, looks at their meter, and says "that area has 12.7 stops, I need to lose .1 stops out of it..." It just doesn't work that way, and it didn't work that way in the days of film, not if you are a creative cinematographer, not an engineer.

With digital cameras, you can see what you are getting live, so you basically work with what the camera can deliver. Most good cinematography has a pretty strong contrast range, but at the same time, it's not like we put important information at 5-stops over or 5-stops under, so the reason we want a wide dynamic range is so that things fall-off into white or black naturally, gracefully. A wide dynamic range is also useful in color-correction, because we all make exposure mistakes.

When working on film where you didn't see the results live, it was more important to know where the ends fell off, but even then, you weren't thinking in terms of one-tenths of a stop... no human can really meter a scene that accurately anyway, not if they are honest. We shot film for printing, so we worked within the range that the print stock could handle... we'd know that a face that was five-stops under would more or less be black, same goes for five-stops over, so generally we "played" in a 10-stop range... but we knew that the negative had more DR than 10-stops. Same goes for video, when someone would say that they couldn't really overexpose a face by more than three stops or underexpose it by more than five stops, that didn't mean that the DR was 8-stops.

This is why I don't pay attention to numbers like "12.6" versus "11.3", etc. Even if I did all the testing myself, I can't be sure if one day I'd get "11.3" and the next day, I'd get "11.4". And I'm certainly not going to take seriously if one manufacturer says "11.3" and another manufacturer says "11.4".

What matters more are comparative differences, like "do I have one more stop of overexposure detail with X camera compared to Y camera IF I set the midtones to the same exposure?" Anything that isn't practical is just an intellectual discussion. Talk to any great DP about shooting on film versus some digital camera and they aren't going to give you figures like "11.3" versus "14.5".
 
Rohan B said:
Red mx only has 11.9 stops of dynamic range?????? That is strange,most of the tests i have seen show above 12.6 stops......
Click to expand...

See, you are falling into that trap of believing in numbers. Someone shoots a test one day and sees 11.9 and someone else shoots a test on another day and sees 12.6... you could assume of course that the highest number ever seen by anyone is the max DR, but on the other hand, you could assume that on any given day you may not be able to achieve max DR so maybe you should believe the most average number counted, or you can assume that different people are counting the steps differently, or using a different testing methodology altogether.

More important than numbers are comparative differences -- if a number of people consistently find that X camera has one more stop of DR than Y camera, whether or not they all agree on the numbers, then odds are high that X camera has more DR than Y camera. On the other hand, that could mean that the majority of people are shooting the Y camera improperly, for whatever reason, though that's another thing to factor.

Even bad tests tell you something, if only what not to do. But if everyone makes the same mistake with a particular camera, then that's something for the manufacturer to think about.
 
David Mullen ASC said:
When working on film where you didn't see the results live, it was more important to know where the ends fell off, but even then, you weren't thinking in terms of one-tenths of a stop... no human can really meter a scene that accurately anyway, not if they are honest. We shot film for printing, so we worked within the range that the print stock could handle... we'd know that a face that was five-stops under would more or less be black, same goes for five-stops over, so generally we "played" in a 10-stop range... but we knew that the negative had more DR than 10-stops. Same goes for video, when someone would say that they couldn't really overexpose a face by more than three stops or underexpose it by more than five stops, that didn't mean that the DR was 8-stops.
Click to expand...

Exactly David. When shooting film I always took what I read in American Cinematographer with a heavy dose of salt. Could be useful, but not unless it was backed up by my own testing and experience. My DP friends and I would ask Kodak for a 400' load every time they came out with a new stock and then shoot under and over exposure tests with flesh tones, gray scale, and Macbeth in the frame, spot metering everything and recording it. Then print two passes: a one light and a time to match normal exposure (courtesy of FotoKem). If we liked a stock and began shooting with it, the tests were only a starting point. Real world experience was needed to complete our understanding of how the negative responded to light, and beyond that, how it translated into print or video delivery. It's one thing to look at grey patches in a test; it's quite another thing to point your spot meter at the folds in a sheer curtain and know (without digital WYSIWYG) whether you are going to see detail in the curtain. Even more important - where are the faces going to end up?

The RED MX sensor (and any other digital sensor) is our stock. I try to bring the same level of preparation to understanding how it responds to light that I would to film. It's tragic that there is so much hype about the various digital cameras. Good on you and Graeme for encouraging others to do their own tests and setting the bar high for the craft of cinematography. Your collective wisdom and technical expertise is welcome in a sea of confusion.

Tim Eaton
 
The variables in most testing methodologies make comparisons problematic. Other posters have noted several key issues. David M. has cogently weighed in with real world perspective. Does that mean that every serious DP needs to run the same tests their peers have already run in order to "know" what's what? Perhaps.
FWIW, any camera system that uses compression, even Graeme's lovely RedCode wavelet version, needs to be evaluated based on more than static charts in order to be fully understood.
Due to fundamental differences in design and output format comparing disparate camera systems is highly dependent on processing, post and even display choices.
Determining signal to noise ratio, which to me is the salient metric for "true" DR, is also hard to quantify due to the inherent subjectivity of noise tolerance.
IMO the really great news is that the top cameras on the market today all have tremendous imaging characteristics. Digital can stand up to film in 2011, and that's darn impressive.
This is a very exciting time to be a cameraman.

Cheers - #19
 
You must log in or register to reply here.
Back
Top