Film and digital resolution compared

[ColSebastianMoran]

Just discovered this great exposition by Rich from a couple years ago.

@giganova, thanks for the fabulous entry today. I love real world comparisons.

I then converted the 24 Mpix digital image to monochrome for comparison with the FP4 b/w film. Otherwise, the picture is completely untouched and I made no changes in Photoshop.

Just a small nit: The camera algorithms and/or/ the RAW conversion have done some manipulation of the image. So it's not completely unprocessed.

A more substantive comment:
- Digital responds to increasing spatial frequency quite flat until it falls apart completely. That is, digital maintains higher contrast until it dies.
- Film gradually loses contrast as spatial frequency increases. Thereby, film can maintain some contrast in very detailed areas, where digital lost it completely.

My own experiments:
- Shooting digital vs. Fuji 200 C-41 film. Scanned Coolscan V, 4000 ppi.
- 6 MPx digital was comparable to film, cleaner image, not quite same resolution
- Today, 24MPx digital is way better than I can get in 35mm film

Here's my comparison, 100% actual pixels:

 

[giganova]

Good points, Colonel!

Film gradually loses contrast as spatial frequency increases. Thereby, film can maintain some contrast in very detailed areas, where digital lost it completely.

Is this what you are referring to?

 

[Kent]

I just have seen this thread today. Thanks for reviving it. 😉

It always interesting to read those comparisons and yours is surely one of the more scientific ones. But the problem with them is that they are merely theoretical, because film and sensors work in different ways. Thus they cannot really be compared objectively. What we can do is to use our equipment and see which one offers us a "better" (= "higher") resolution effectively. And as you have pointed out there is a long chain of influences. Who of us has a drum scanner?

And to be honest, I don't even care if film or sensors show a higher resolution. The only question that everybody has to answer on a very personal level is: Is that good enough for me, my needs and for what I expect?

For me it goes like this:

ISO 200 color neg film 135, exposed in a good cam and decent lens, flatbed neg scan = good enough!

ISO 400 b/w neg film 135, exposed in a good cam and decent lens, flatbed neg scan = good enough!

ISO 800 b/w neg film 135, exposed in a good cam and decent lens, flatbed neg scan = good enough!

ISO 160 colour neg film 120, exposed in a good cam and decent lens, flatbed neg scan = definitely good enough!

ISO 400 b/w neg film 120, exposed in good cam and decent lens, flatbed neg scan = definitely good enough!

Fuji X-T1 "APS" (up to ISO 6400) with a decent lens = definitely good enough!

Nikon Df "fullframe" (up to ISO 6400) with a decent lens = definitely good enough!

Leica M8 "APS-H" (up to ISO 640, b/w higher) with a decent lens = definitely good enough!

Nikon D7000 "APS" (up to ISO 3200) with a decent lens = definitely good enough!

Pentax K100D "APS" (only 6 MPix, up to ISO 800) with a decent lens = good enough!

Panasonic GF3 "µ4/3" (up to ISO 800) with a decent lens = good enough!

Nikon D1 "APS" (only 2,7 MPix, up to ISO 400) with a decent lens = just good enough (in most cases)!


Modern cams with good lenses and good old cams with good lenses and good film will always be "good enough" for me!

My personal photographic abilities and the draw-backs of real-life-shooting will be much more restrictive to the quality of the outcome than any of my gear. 😉

 

[giganova]

But the problem with them is that they are merely theoretical, because film and sensors work in different ways.

That's why I posted analog/digital image comparisons today (see post #71).

 

[RichC]

Only just noticed the replies! I'll give my thoughts soon - got to finish my tax return first, to avoid the £100 missed deadline fine!

What I will say quickly though is that it my comparison was for a good, practical reason. I shoot film and digital, and wanted to know how comparable they were in terms of megapixels so I had a guide as to how large I can print film images. Megapixels are a useful measurement for me as I'm very used to making C type and inkjet prints from digital camera images, and know how sharp such a print will be given its resolution (in dpi) and size (width/height in cm or inches).

Prints using my Mamiya M1000 and Nikon D800E seemed to show that 645 film and 36 MP full-frame digital were comparable, with same-size prints having similar detail.

But the eye can deceive and the mind is good at convincing you to believe what you like to rather than truth - hence my comparison. Which did actually confirm that my two cameras are comparable!

Film images do have a different aesthetic to digital images, and this is important. My test was simply about how much detail can be seen in a real print with real equipment, and to accurately estimate this detail. To reiterate: real detail that a person can actually see.

Lastly, all this becomes important only if (a) you're interested in seeing detail and (b), given (a), you make large prints that push the boundaries of your camera.Both points apply to me!

 

[Merlijn53]

I just have seen this thread today. Thanks for reviving it. 😉

It always interesting to read those comparisons and yours is surely one of the more scientific ones. But the problem with them is that they are merely theoretical, because film and sensors work in different ways. Thus they cannot really be compared objectively. What we can do is to use our equipment and see which one offers us a "better" (= "higher") resolution effectively. And as you have pointed out there is a long chain of influences. Who of us has a drum scanner?

And to be honest, I don't even care if film or sensors show a higher resolution. The only question that everybody has to answer on a very personal level is: Is that good enough for me, my needs and for what I expect?

For me it goes like this:

ISO 200 color neg film 135, exposed in a good cam and decent lens, flatbed neg scan = good enough!

ISO 400 b/w neg film 135, exposed in a good cam and decent lens, flatbed neg scan = good enough!

ISO 800 b/w neg film 135, exposed in a good cam and decent lens, flatbed neg scan = good enough!

ISO 160 colour neg film 120, exposed in a good cam and decent lens, flatbed neg scan = definitely good enough!

ISO 400 b/w neg film 120, exposed in good cam and decent lens, flatbed neg scan = definitely good enough!

Fuji X-T1 "APS" (up to ISO 6400) with a decent lens = definitely good enough!

Nikon Df "fullframe" (up to ISO 6400) with a decent lens = definitely good enough!

Leica M8 "APS-H" (up to ISO 640, b/w higher) with a decent lens = definitely good enough!

Nikon D7000 "APS" (up to ISO 3200) with a decent lens = definitely good enough!

Pentax K100D "APS" (only 6 MPix, up to ISO 800) with a decent lens = good enough!

Panasonic GF3 "µ4/3" (up to ISO 800) with a decent lens = good enough!

Nikon D1 "APS" (only 2,7 MPix, up to ISO 400) with a decent lens = just good enough (in most cases)!


Modern cams with good lenses and good old cams with good lenses and good film will always be "good enough" for me!

My personal photographic abilities and the draw-backs of real-life-shooting will be much more restrictive to the quality of the outcome than any of my gear. 😉

I was reading this thread from the beginning, well, the first and the last page to be honest, and was asking myself what's the use of all this nonsense, what are we trying to prove?
Then I read this post and thought, you're absolutely right, that's the only way to look at it (imho)
Frank

 

[pvdhaar]

First of all, kudos for RichC and Giganova for their posts; thoroughly enjoyable reads! And it's interesting to see that you both come to similar conclusions from a theoretical and practical point of view. And, by the way, it's very like what I find using the layman's technique of just comparing prints and checking whether they 'look nice'..

There is however a fly in the ointment, and that's the actual P&P of film. If you develop and scan yourself, then yes, film is pretty good and can match what you found. But if you bring your film to a lab for develop & scan/print.. it's such a hit and miss. Sometimes I get 120 back so rotten that it can't even match 35mm shot with a 3 element lens on a Mju-I. And mind you, that's not 120 shot on a Holga, but carefully metered with a 6x6 SQB and multicoated 80mm on a tripod using mirror lock up..

From a practical standpoint, that means that for some people, the crossover of resolution may seem to come at a lower MegaPixel count depending on how/where they (have) developed and scanned their film. Oh, and film P&P seems to be getting still worse lately..

 

[giganova]

If you develop and scan yourself, then yes, film is pretty good and can match what you found. But if you bring your film to a lab for develop & scan/print.. it's such a hit and miss.

Right on! Even if you shell out serious money on professional lab scans, they can not match in terms of resolution and dynamic range what you can do at home with a decent scanner and a bit of experience. All scans that I received from professional labs had cranked up the contrast so much that most of the details in the mid range of the images were lost.

 

[ColSebastianMoran]

Good points, Colonel!

Is this what you are referring to?

Yes, exactly what I'm talking about.

Digital has high contrast until it falls apart. Film degrades more gradually with increasing spatial frequency. If you are willing to look at very low contrast, film can resolve a lot of detail, even in presence of grain.

Back to my comparison shot in #78 (above) at the famous Cheers pub in Boston: The film has more resolution, the 6MPx digital is a better looking (i.e. sharper looking) image. That was only a D100!

 

[Hatchetman]

Giganova posted some examples above of FP4+ scanned vs. digital. They both look good but a couple things...FP4+ is not a minimal grain film. I would use Acros or TMAX 100. I would also be curious to compare an analog print to a digital print. A scan is always going to degrade the quality of the image. If you are just posting images on a computer screen, using 35mm film is going to be an uphill battle.

 

[:: Mark]

My rather non-scientific testing with Delta 100 suggests that at best there is about 16MP worth or useful data in a file. An E-M5 with 25mm f1.4 comfortably out-resolves an M7 with a 50mm C-Sonnar (admittedly not the sharpest lens, even stopped down to f8). I also ran some crude tests to compare the Epson flat bed scanner vs a 5DIII with 100mm macro lens, and concluded that both scanner and DSLR were out-resolving the film and that the film image was limited by the grain size and that probably the best 35mm output would not be around 16MP with Delta 100 (significantly less for faster films, and maybe slightly more for B&W scans made from Ektar 100).

However, this pretty much ignores the main reason why you might choose to shoot film rather than digital today. The film grain is quite effective at giving the perception of more detail than there really is. HP5+ only seems to give at most 10MP of detail, yet large prints can be vastly better than a 10MP camera precisely because of the effect of the grain and the visual effect that it imparts.

I made the comparisons mainly because I was trying to figure out my scanning. Film and digital have very distinct characters and shooting experience, and unless extremely large images are needed (A3 and up) resolution is usually the last criteria I would use when deciding how to approach a project...

 

[edge100]

I also ran some crude tests to compare the Epson flat bed scanner vs a 5DIII with 100mm macro lens, and concluded that both scanner and DSLR were out-resolving the film and that the film image was limited by the grain size and that probably the best 35mm output would not be around 16MP with Delta 100 (significantly less for faster films, and maybe slightly more for B&W scans made from Ektar 100).

f/14 = diffraction limiting sharpness. On FF, f/11 would be better, f/8 better still.
iPad as light source = limiting by iPad pixel size. This is a major issue for resolution (you've effectively limited it to the size of the iPad's pixels).

Other issues: how did you ensure film and camera were level and vibration-free (mirror is a good start but not super accurate in my experience...a bubble level is far more analytical)? How did you ensure film flatness (neg carrier also a good idea, although quality varies)? How did you release the shutter on the Canon? Did you autofocus or MF? If the latter, how did you assess focus?

I have scanned 1000s of film images (35mm to LF) with a D800 and a 100mm macro and the resolution is VASTLY better than a flatbed. It's not even close. The D800/macro combination outresolves a 9000ED and isn't far off from a drum scan.

Suffice it to say, you are not getting the most out of your film with your DSLR method.

 

[:: Mark]

f/14 = diffraction limiting sharpness. On FF, f/11 would be better, f/8 better still.
iPad as light source = limiting by iPad pixel size. This is a major issue for resolution (you've effectively limited it to the size of the iPad's pixels).

The M7 and E-M5 were shot on a tripod, at f8 and f4 respectively (which gives pretty much the best aperture in both cases).

That said, I don't usually worry about diffraction limiting when shooting film - any difference between f8 and f16 on the C-Sonnar is largely irrelevant with film, whereas on digital there is a clear difference.

Scanning with an iPad backlight is not ideal. You can avoid the pixelation simply by moving the screen further from the film -I do this routinely with backlit macro images, but struggle with the (not very continuous) light spectrum that the LEDs emit.

A drum scanner will obviously produce a higher resolution scan. But it is a mistake to assume that this is a higher resolution view of the scene photographed - effectively all you are resolving is more detail on the grain. This might be the aim - but it is not fundamentally different to adding simulated grain to a lower resolution digital image (although whether or not you like this aesthetically is another issue entirely!).

One thing I would really be curious to understand is why I usually get sharper B&W images made from (good) colour film vs low ISO black and white. Maybe it is just the grain size, but maybe also the digitisation works better with a colour input and subsequent B&W conversion.

 

[Ronald M]

I rather expected this when I started researching an upgrade from digital and realized the advantages of film. But now I have to wonder if I still want to make optical paper prints, or go K7 with a good printer. Not sure I am ready to even think about laser-scanning photo paper...

So much for sleeping tonight, dammit! :bang:

Optical printing set up is still open to me.

Inkjet prints are financially not competitive unless I make many more prints than I do. Dried ink, short shelf ink, fading unless you use pigmented ink and then I am back to 6 month shelf life. Then you get into printer problems.

I will never buy another ink printer.

My prints go to a pro lab that uses laser printers on Kodak photo paper. I can not tell them from what I produced in my darkroom, i.e. perfect.

Scanning put an extra resolution losing step in the process. There are no scanners made that the average amateur can afford that produce better than average or low quality. Think of the KM5400 or Nikon scanners. Noise reduction does not well on film grain. The programs that did are gone, think grain surgery.

Film and color chemicals are increasing difficult to obtain and have been for three or four decades.

AiPro Lab does my printing for me. I can burn/dodge and manipulate as much as I want and they furnish a pro level print at a reasonable price. As good as I could do in my darkroom. Their prices are less than cost of ink to me and I need not maintain a printer.

If you wish to do film for nostalgia, be my guest. My film Leicas are all sold except for 111C & 111F and a very mint M6. Digital RF Leicas and Pro Nikons are my tools of choice now.

If you need more res from small format, I have seen super photos stitched that are 40x60. They pass the magnifying glass & smell test easily.

 

[edge100]

The M7 and E-M5 were shot on a tripod, at f8 and f4 respectively (which gives pretty much the best aperture in both cases).

I meant that your Canon was set to f/14 (as per your post) when "scanning" the film. THAT is diffraction limited.

 

[Ronald M]

I rather expected this when I started researching an upgrade from digital and realized the advantages of film. But now I have to wonder if I still want to make optical paper prints, or go K7 with a good printer. Not sure I am ready to even think about laser-scanning photo paper...

So much for sleeping tonight, dammit! :bang:

Optical printing set up is still open to me.

Inkjet prints are financially not competitive unless I make many more prints than I do. Dried ink, short shelf ink, fading unless you use pigmented ink and then I am back to 6 month shelf life. Then you get into printer problems.

I will never buy another ink printer.

My prints go to a pro lab that uses laser printers on Kodak photo paper. I can not tell them from what I produced in my darkroom, i.e. perfect.

Scanning put an extra resolution losing step in the process. There are no scanners made that the average amateur can afford that produce better than average or low quality. Think of the KM5400 or Nikon scanners. Noise reduction does not well on film grain. The programs that did are gone, think grain surgery.

Film and color chemicals are increasing difficult to obtain and have been for three or four decades.

AiPro Lab does my printing for me. I can burn/dodge and manipulate as much as I want and they furnish a pro level print at a reasonable price. As good as I could do in my darkroom. Their prices are less than cost of ink to me and I need not maintain a printer.

If you wish to do film for nostalgia, be my guest. My film Leicas are all sold except for 111C & 111F and a very mint M6. Digital RF Leicas and Pro Nikons are my tools of choice now.

If you need more res from small format, I have seen super photos stitched that are 40x60. They pass the magnifying glass & smell test easily.

 

[Deleted member 82967]

Post of the year. Thanks for tackling a hard subject so objectively.

10 years later ... post of the decade !!

 

[ColSebastianMoran]

10 years later ... post of the decade !!

Yes, this thread has weathered well.
My point in 2016: 6MPx digital looked pretty good compared to good 35mm shot, 24Mpx will be much higher res.

My point today: We now have 50+ Mpx in high end bodies. And, I think 24MPx will be pretty good for camera scanning of 35mm.

 

[hap]

Optical printing set up is still open to me.

Inkjet prints are financially not competitive unless I make many more prints than I do. Dried ink, short shelf ink, fading unless you use pigmented ink and then I am back to 6 month shelf life. Then you get into printer problems.

I will never buy another ink printer.

My prints go to a pro lab that uses laser printers on Kodak photo paper. I can not tell them from what I produced in my darkroom, i.e. perfect.

Scanning put an extra resolution losing step in the process. There are no scanners made that the average amateur can afford that produce better than average or low quality. Think of the KM5400 or Nikon scanners. Noise reduction does not well on film grain. The programs that did are gone, think grain surgery.

Film and color chemicals are increasing difficult to obtain and have been for three or four decades.

AiPro Lab does my printing for me. I can burn/dodge and manipulate as much as I want and they furnish a pro level print at a reasonable price. As good as I could do in my darkroom. Their prices are less than cost of ink to me and I need not maintain a printer.

If you wish to do film for nostalgia, be my guest. My film Leicas are all sold except for 111C & 111F and a very mint M6. Digital RF Leicas and Pro Nikons are my tools of choice now.

If you need more res from small format, I have seen super photos stitched that are 40x60. They pass the magnifying glass & smell test easily.

Do you have alink to AirPro Lab?

 

[RichC]

Yes, this thread has weathered well.

Thanks - OP here... Still alive and taking photographs! 😉

The maths theory could do with a bit of tweaking: the results are still 100% correct, but there are some iffy bits in the working out I really can't be bothered to fix!