Dynamic Range of Film still better than Digital Cameras?

[f64_or_bust]

You're talking about read noise. And you're correct that we've removed it in modern cameras -- moot issue. DR is limited by shot noise not read noise. Shot noise is in the signal (light). Here's a good explanation: What's that noise? Part one: Shedding some light on the sources of noise Also look at this concerning the size of the photosites: The effect of pixel size on noise

No, I'm talking about "noise", both shot and both types of read noise (pre and post amplificatino). Shot noise comprises most of the noise in an image, but when ISO amplification is applied (high ISO) you amplify the shot AND the read noise that occurs before amplification. And that happens in the shadows. What seems to be more under control in current cameras is the read noise contributed post amplification, which should be pretty low these days. But read noise before amplification is going to follow the same principles in all electronic noise, including electrical spectral density (noise levels rise with frequency) and temperature (noise rises with temperature). I haven't looked into exactly how low pre-amplification read noise is these days, but there is a theoretical limit dicteated by physics that can't be surmounted (except in post). The same limits exist in all electronics.

In the end, noise is noise, and regardless of the source, it all pretty much looks the same.

You have paralell issues in film, though. The areas with lower exposure, like shadows, will start to reveal grain, and if you push-process, you might increase the aparent sensitivy of the film, raising the virtual ISO, but you'll increase the grain too. The same kind of thing happens if you try to pull up shadows during printing. Highlights will eventually hit the gamma toe and compress, and there's no recovering that. The real problem here is that in the physical darkroom you don't have even a fraction of the tools to pull that off that you do in software of 20 years ago. And now, we can reduce noise in post. To reduce grain in film, you have to shoot bigger negatives, dictating different cameras and lenses. From personal experience, I'll tell you that you can't actually print the full DR of film. The paper doesn't match in DR in it's "brightness" or "blackness". In the art of printing from film, you have to use whatever tools you have to fake DR. And suffice to say, you have absolutely NO selection tools available, and have enough trouble bending the gamma curve!

You also can't project the image at anything close to the original film's DR, assuming you even found a slide film that could do better than 10 stops (Ektachrome E100 is a practical 5-7 stops, and Kodachrome is long gone). The reason is ambient light on the screen sets your black level, and unless you have a projector with zero spill, a perfectly dark room and require your viewers to wear black head to toe including masks, one guy in the front row wearing a white shirt will reflect enough light from the screen back onto the screen to hold your image to well under 8 stops. Projecting digital images has the same problem, of course, except that at least you could start with an image with better than 7 stops DR.

I just can't find a DR advantage for film. It's just not really there, and mostly overblown.

 

[Ysarex]

No, I'm talking about "noise", both shot and both types of read noise (pre and post amplificatino). Shot noise comprises most of the noise in an image,

I'm talking about shot noise which as you note is the noise we see now -- read noise is under control. So DR in our digital images is basically shot noise limited.

but when ISO amplification is applied (high ISO) you amplify the shot AND the read noise that occurs before amplification. And that happens in the shadows. What seems to be more under control in current cameras is the read noise contributed post amplification, which should be pretty low these days. But read noise before amplification is going to follow the same principles in all electronic noise, including electrical spectral density (noise levels rise with frequency) and temperature (noise rises with temperature). I haven't looked into exactly how low pre-amplification read noise is these days, but there is a theoretical limit dicteated by physics that can't be surmounted (except in post). The same limits exist in all electronics.

In the end, noise is noise, and regardless of the source, it all pretty much looks the same.

The source of the noise matters. Read noise is in the hardware and shot noise is in the signal. I brought up noise relative to sensor size saying that we tend to get more DR from larger sensors. From the DPreview article I linked above: "...at the same f-number (both cameras set to F2.8), the full frame camera will see four times as much light as a camera with a Four Thirds sensor, since it is exposed to the same light-per-unit-area but has a sensor with four times the area.

As a result, when you shoot two different sized sensors with the same shutter speed, f-number and ISO, the camera with the smaller sensor has to produce the same final image brightness (which the ISO standard demands) from less total light. And, since we've established that capturing more light improves your signal-to-noise ratio, this means every output tone from the larger sensor will have a better signal-to-noise ratio, so will look cleaner."

So there's a noise advantage and therefore a DR advantage that derives specifically from the sensor size apart from the size or density of the photosites.

You have paralell issues in film, though. The areas with lower exposure, like shadows, will start to reveal grain, and if you push-process, you might increase the aparent sensitivy of the film, raising the virtual ISO, but you'll increase the grain too. The same kind of thing happens if you try to pull up shadows during printing. Highlights will eventually hit the gamma toe and compress, and there's no recovering that. The real problem here is that in the physical darkroom you don't have even a fraction of the tools to pull that off that you do in software of 20 years ago. And now, we can reduce noise in post. To reduce grain in film, you have to shoot bigger negatives, dictating different cameras and lenses. From personal experience, I'll tell you that you can't actually print the full DR of film. The paper doesn't match in DR in it's "brightness" or "blackness". In the art of printing from film, you have to use whatever tools you have to fake DR. And suffice to say, you have absolutely NO selection tools available, and have enough trouble bending the gamma curve!

You also can't project the image at anything close to the original film's DR, assuming you even found a slide film that could do better than 10 stops (Ektachrome E100 is a practical 5-7 stops, and Kodachrome is long gone). The reason is ambient light on the screen sets your black level, and unless you have a projector with zero spill, a perfectly dark room and require your viewers to wear black head to toe including masks, one guy in the front row wearing a white shirt will reflect enough light from the screen back onto the screen to hold your image to well under 8 stops. Projecting digital images has the same problem, of course, except that at least you could start with an image with better than 7 stops DR.

I just can't find a DR advantage for film. It's just not really there, and mostly overblown.

 

[webestang64]

Deleted post.

 

[f64_or_bust]

I'm talking about shot noise which as you note is the noise we see now -- read noise is under control. So DR in our digital images is basically shot noise limited.

Only one type of read noise is "under control", and that's read noise post gain. Read noise pre gain is a matter of thermal noise, essentially shot noise happening in the electrical domain. And it's generally not possible to equal the theoretical thermal noise floor. At high ISO (which is achieved by amplification/gain) both shot and pre-gain read noise are amplified.

The source of the noise matters. Read noise is in the hardware and shot noise is in the signal.

And both exist when you read a photosite. One is a function of the size and lensing of the photosite, one is a function of thermal noise, and both are combined and indistinguishable from eachother.

I brought up noise relative to sensor size saying that we tend to get more DR from larger sensors. From the DPreview article I linked above: "...at the same f-number (both cameras set to F2.8), the full frame camera will see four times as much light as a camera with a Four Thirds sensor, since it is exposed to the same light-per-unit-area but has a sensor with four times the area.

Yup. Shot noise.

As a result, when you shoot two different sized sensors with the same shutter speed, f-number and ISO, the camera with the smaller sensor has to produce the same final image brightness (which the ISO standard demands) from less total light. And, since we've established that capturing more light improves your signal-to-noise ratio, this means every output tone from the larger sensor will have a better signal-to-noise ratio, so will look cleaner."

Yup. Again, shot noise.

So there's a noise advantage and therefore a DR advantage that derives specifically from the sensor size apart from the size or density of the photosites.

Yup, no argument. But I'm also including pre-gain read noise, which is not shot noise, and is definitely a factor in high ISO shadow noise.

 

[Ysarex]

Only one type of read noise is "under control", and that's read noise post gain. Read noise pre gain is a matter of thermal noise,

But not a major issue because according to you; "Shot noise comprises most of the noise in an image," So it's shot noise that is primarily responsible for limiting DR. Therefore DR tends to correlate with sensor size since shot noise correlates with sensor size.

essentially shot noise happening in the electrical domain. And it's generally not possible to equal the theoretical thermal noise floor. At high ISO (which is achieved by amplification/gain) both shot and pre-gain read noise are amplified.

And both exist when you read a photosite. One is a function of the size and lensing of the photosite, one is a function of thermal noise, and both are combined and indistinguishable from eachother.

Yup. Shot noise.

Yup. Again, shot noise.

Yup, no argument. But I'm also including pre-gain read noise, which is not shot noise, and is definitely a factor in high ISO shadow noise.

 

[f64_or_bust]

But not a major issue because according to you; "Shot noise comprises most of the noise in an image," So it's shot noise that is primarily responsible for limiting DR. Therefore DR tends to correlate with sensor size since shot noise correlates with sensor size.

And then I qualified that statement by noting under what conditions read noise becomes an issue (high gain/high ISO). Also, while shot noise is a function of photosite size and design, read noise may not corellate with that. As shot noise is lowered, pre-amplification read noise won't necessarily follow (I would guess, it just won't at all). Read noise will be a function of temperature and the specific photosite electrical impedance. I don't have information handy that relates to the electrical impedance vs photosite size (vs read noise), but I suspect it could be more complex than just that.

We are REALLY nit-picking here now. The long and the short of it is, film does not have more DR than what's possible in digital image capture. Noise limits both, but there are options in a digital sensor that don't exist with film, and the big jumps in film noise (grain) can only be had (at the same ISO) by using bigger film, which dictates larger cameras and lenses, which may only be practical in a very few applications.

High speed motion pictures of rocket launches is not one of them.

 

[Ysarex]

And then I qualified that statement by noting under what conditions read noise becomes an issue (high gain/high ISO). Also, while shot noise is a function of photosite size and design, read noise may not corellate with that. As shot noise is lowered, pre-amplification read noise won't necessarily follow (I would guess, it just won't at all). Read noise will be a function of temperature and the specific photosite electrical impedance. I don't have information handy that relates to the electrical impedance vs photosite size (vs read noise), but I suspect it could be more complex than just that.

We are REALLY nit-picking here now.

Yes, I remember the last camera I had that presented any kind of visible read noise under strained conditions and that was a Canon 5dmkII and that was 2008. All of my Fuji and Nikon (and even Canon) cameras since then equipped with Sony sensors exhibit no discernible read noise until past the point that shot noise makes the image unusable.

The long and the short of it is, film does not have more DR than what's possible in digital image capture. Noise limits both, but there are options in a digital sensor that don't exist with film, and the big jumps in film noise (grain) can only be had (at the same ISO) by using bigger film, which dictates larger cameras and lenses, which may only be practical in a very few applications.

High speed motion pictures of rocket launches is not one of them.

 

[Red Wolf]

Forgetting outside analysis for a moment... My personal experience with recent testing remains that film holds highlights much better but that digital saw into the shadows deeper and with less noise when the shot was very dark. Whatever the measured reality, my digital cameras produces really great images that are usually very acceptable for my purposes, which are mostly creative, not documenting scientific data of an event like a launch.

If I were recording a rocket launch for such a reason, I'd probably use film to hold the details in those flames and the areas around them, which could be important in analyzing things. I have shot launches on video (NTSC) on digital (motion and stills) and on film. The digital camera doesn't hold color or detail in that extreme top end of exposure. Then again... neither did the film, but it did a lot better and rolled off much more gracefully and at a more extreme level of exposure.

In the end, for me, they are just two different mediums for imaging. I work with both. I like both. Digital is just so much easier, cheaper and more convenient to use. That really is the deciding factor for most of what I shoot. But I have a love for film that will always be there, as well as a respect for what it does well.

 

[Ysarex]

Forgetting outside analysis for a moment... My personal experience with recent testing remains that film holds highlights much better but that digital saw into the shadows deeper and with less noise when the shot was very dark. Whatever the measured reality, my digital cameras produces really great images that are usually very acceptable for my purposes, which are mostly creative, not documenting scientific data of an event like a launch.

If I were recording a rocket launch for such a reason, I'd probably use film to hold the details in those flames and the areas around them, which could be important in analyzing things. I have shot launches on video (NTSC) on digital (motion and stills) and on film. The digital camera doesn't hold color or detail in that extreme top end of exposure.

My digital cameras always do hold color and detail in the extreme top end of exposure if that's what I want them to do. That usually is what I want and so my typical exposure practice is to place the brightest diffuse highlight at the sensor saturation threshold where I will record color and detail. Furthermore since digital sensors record data linearly I have better highlight tone separation than you typically get from film which you have to expose while accounting for the film's curve.

Then again... neither did the film, but it did a lot better and rolled off much more gracefully and at a more extreme level of exposure.

In the end, for me, they are just two different mediums for imaging. I work with both. I like both. Digital is just so much easier, cheaper and more convenient to use. That really is the deciding factor for most of what I shoot. But I have a love for film that will always be there, as well as a respect for what it does well.

 

[smoke665]

My digital cameras always do hold color and detail in the extreme top end of exposure if that's what I want them to do. That usually is what I want and so my typical exposure practice is to place the brightest diffuse highlight at the sensor saturation threshold where I will record color and detail. Furthermore since digital sensors record data linearly I have better highlight tone separation than you typically get from film which you have to expose while accounting for the film's curve.

You know one of the comments I hear from the film side is, digital is to sterile, to sharp, to much detail, etc., so is increasing DR to such high levels really that important??? I've dabbled in everything from oil, to watercolor, to pen and pencil. In none of these is the detail or DR even close to what film can portray, so as an artist i have to ask if we might chasing something unneeded? Maybe we need to concentrate more on content and composition, to see the beauty of the forest, not worry so much about the trees. If you take "creativity" out of photography why do it just let AI paint something for you.

 

[Overread]

You know one of the comments I hear from the film side is, digital is to sterile, to sharp, to much detail, etc., so is increasing DR to such high levels really that important??? I've dabbled in everything from oil, to watercolor, to pen and pencil. In none of these is the detail or DR even close to what film can portray, so as an artist i have to ask if we might chasing something unneeded? Maybe we need to concentrate more on content and composition, to see the beauty of the forest, not worry so much about the trees. If you take "creativity" out of photography why do it just let AI paint something for you.

I would argue that its never a bad thing to chase improvements, otherwise we'd still have the first generation of film cameras.
Technical improvement of the machinery and the user-operation of the machinery increases the range of creative options one has. Increasing Dynamic Range means being able to take artistic shots of a wider variety of subjects in more varied lighting conditions.
Heck just seeing ISO's going from when I started with a 400D, where ISO 800 was about the limit; to what we have today where we can consider 100,000 and greater. That doesn't make one a better artist, but means one can shoot in weaker and weaker lighting and still get great shots where the composition and content is not marred by ugly grainy results.

 

[Ysarex]

You know one of the comments I hear from the film side is, digital is to sterile, to sharp, to much detail, etc., so is increasing DR to such high levels really that important???

No, but it's not harmful and it's a nice option to have. Back on the first page of this thread I said: "So for most of our normal day in day out photography 6 to 7 stops of DR is plenty and both film and digital provide that."

It's not real common but we do encounter very high contrast lighting and might like to capture the scene. Over the decades film would frustrate me under such conditions -- eg. the first photo I presented in this thread. I often shot color transparency film which completely crashes and burns in very high contrast light. I carry a camera with me everywhere I go. Film would stop me from taking photographs that I easily take now with digital. I really like to be able to photograph whatever I see. Digital opened that door for me and no way am I going back through and closing it. Again another example: The scene below presents 12+ stops of DR. Not a single ray of sunlight is reaching the foreground while in a blue sky puffy white clouds are sunlit. I was just walking in the park -- nothing special. I liked the scene and I saw it the way I photographed it -- click. The camera JPEG is there for reference.

With color transparency film in a camera the scene is impossible. With color negative film good luck and you'll have hell to pay to get anything close to what I did with virtually no effort. I really enjoy the freedom I have now to photograph whatever want, whatever I see. Film never permitted that.

I've dabbled in everything from oil, to watercolor, to pen and pencil. In none of these is the detail or DR even close to what film can portray, so as an artist i have to ask if we might chasing something unneeded? Maybe we need to concentrate more on content and composition, to see the beauty of the forest, not worry so much about the trees.

This is a false dichotomy. Getting the expressive content of the image right never stands in opposition to doing a good job with the tech side. Doing the tech side well doesn't take creativity away. The tech side supports the aesthetic expressive side and we do the tech side well because we care about the expressive side. I'm sitting here right now listening to my recording of Gidon Kremer playing Weinberg's Three Pieces for Violin and Piano. Kremer didn't go shopping at Walmart to get the violin he plays. He loves and respects Weinberg's music and does the best job he can preforming it.

If you take "creativity" out of photography why do it just let AI paint something for you.

 

[smoke665]

It's not real common but we do encounter very high contrast lighting and might like to capture the scene. Over the decades film would frustrate me under such conditions

I don't remember a lot of problems, with capturing the intended shot. It isn't all that hard to reduce the DR with supplemental light, move to a location with better light, or wait till the right time of the day. I dealt with the problem.

This is a false dichotomy. Getting the expressive content of the image right never stands in opposition to doing a good job with the tech side. Doing the tech side well doesn't take creativity away. The tech side supports the aesthetic expressive side and we do the tech side well because we care about the expressive side.

No it isn't. Look at the old masters, you won't count the eye lashes on any portrait, but it doesn't take away from the beauty of the piece. Composition/lighting/color etc. is far more important. I.m just saying that we might be reaching the limit for practical application on DR. Ive read the average untrained human eye is only capable of registering 10-14 stops.....we're there now.

 

[Overread]

I don't remember a lot of problems, with capturing the intended shot. It isn't all that hard to reduce the DR with supplemental light, move to a location with better light, or wait till the right time of the day. I dealt with the problem.


No it isn't. Look at the old masters, you won't count the eye lashes on any portrait, but it doesn't take away from the beauty of the piece. Composition/lighting/color etc. is far more important. I.m just saying that we might be reaching the limit for practical application on DR. Ive read the average untrained human eye is only capable of registering 10-14 stops.....we're there now.

Thing is I bet there were times you were limited by the light and you either adapted or didn't shoot. Increasing the performance of the machine meant that those times keep reducing. There are also many situations where you can't control things; where you can't wait or add more light or move to a different location. Anyone doing event/reporting or such photography. Heck even landscaping if you're on holiday you might not have the option to revisit the same spot 50 times to wait for the perfect light. So having a greater range of tolerances and performance for the camera means you can get that better shot at the time.

It means the realisation of ones artwork and creativity is less limited/restricted.
It's the same as how a better quality brush enhances what a painter can do. Yes they can still paint well with a rubbishy brush and there will be instances where cheap or top end won't make a huge difference; but there will be times when they just need a better quality.




When it comes to performance of the human eye its a bit of a false element I feel. The eye moves around all the time so even if the photo has a greater dynamic range than the eye can take in all at once; the eye can move around the scene itself. It's similar to how the eye can only see at something like 24fps; but having greater fps actually helps produce smoother motion in animation (which is why video games often aim for 30min and 60 ideal with some even wanting faster at 120 and such).

Also for a camera recording a scene if it can record a greater dynamic range in clear detail and with good exposure; the photographer can then pull that into editing to work with

 

[Ysarex]

I don't remember a lot of problems, with capturing the intended shot. It isn't all that hard to reduce the DR with supplemental light,

It's impossible if you're photographing the Grand Tetons or the pond in my park.

move to a location with better light,

That's a different photo.

or wait till the right time of the day.

That's a different photo.

I dealt with the problem.

I no longer have the problem.

No it isn't. Look at the old masters, you won't count the eye lashes on any portrait, but it doesn't take away from the beauty of the piece. Composition/lighting/color etc. is far more important.

The expressive content of any art work is most important. If you believe that and you're the creator then you're happy to short change the craft/tech side because meh, who cares since it's not the most important thing?!! The two aren't opposed to each other. If you care about what you're trying to express then you'll do the best job you can expressing it.

I.m just saying that we might be reaching the limit for practical application on DR. Ive read the average untrained human eye is only capable of registering 10-14 stops.....we're there now.