Is Film Sharper then Digital or Vice Versa?

[ParkerSmithPhoto]

People also obsess on pixel peeping both digital and film scans at 100%, which is a level of detail you will never get at a normal print viewing size. If you view around 33% in Photoshop, you are seeing approx. the print size if you are working at 300DPI.

 

[Cholentpot]

People also obsess on pixel peeping both digital and film scans at 100%, which is a level of detail you will never get at a normal print viewing size. If you view around 33% in Photoshop, you are seeing approx. the print size if you are working at 300DPI.

I've never gotten a comment on my photos of 'Wow! Look at that resolution and sharpness!'

I have gotten 'Woah, nice photograph. What's the story?' <------------------ THIS IS BIG COMPLIMENT.

It's nice to get feedback on my technical competence but it's really nice to get complimented on my artistic vision.

 

[RalphLambrecht]

Check this out: https://petapixel.com/2019/03/04/review-topaz-sharpen-ai-is-amazing/

Progress being made in AI sharpening tools. I have a theory that AI is going to take over in a way that makes comparing lenses and cameras moot. In about 5 more years throw any crappy image to the AI and it will reconstruct it from the ground up to look however you want it to look. It's hard to imagine how this works. I saw a similar system in 3D rendering. Previously an object had to spend say 1hr rendering reflectivity of light on a 3D vase, but a new system used AI instead, the AI knew what the scene should look like and constructed the scene within seconds based on that knowledge. Thinks are going to get pretty 'out there' soon.

Yeah but ,AI also works on scanned negative files.so, AI cannot be part of the comparison.

 

[Ray Morgenweck]

I’ve been able to make Wet Plate negatives developed with pyrogallic acid developer which I think would exceed any film or digital process in resolution. Also, a Daguerreotype is capable of extreme resolution.

 

[RalphLambrecht]

Adox CMS II 20 has a resolution of 800l/mm.
Scanned at 8000 DPI the resultant resolution is aprox 77Mp.

Developed as SCALA slide:
Example 1 (38mb Jpg)
Example 2 (38mb JPG)

Developed as negative:
Example 3 (42mb Jpg)

resolution is only one part of sharpness.

 

[alanrockwood]

A word of advice. Anytime you see the phrase "Nyquist Theory" on the internet, it's best to disregard the whole thing. 99% of the time it's brought up, it's brought up to prove something it was never intended to provide evidence for. The whole point of the Nyquist theorem is to provide a general guideline for minimal engineering standards for analog to digital conversion. It was never intended to be proof for anything. It is, by it's nature, deeply flawed. For example, the Nyquist Theory assumes a perfectly bandwidth limited system. These do not exist in nature. Therefor, under no circumstance can the Nyquist theory be applied to any system and be relied upon to give accurate results.

About the only time the Nyquist Theorem should be discussed is when you're designing or implementing an ADC system, and you want to know the ballpark for the bare minimum sampling frequency that you could potentially get away with to keep costs, processing, and/or storage space to a minimum. Even then, it should still be tested in a real world scenario to ensure that the results achieved are in line with what was expected.

So it's a handy theory that has it's uses. But more often than not, it's abused on the internet to "prove" some poorly thought out concept concocted by a neophyte with an axe to grind.

You have vastly overstated the case about the Nyquist theorem. You say it is deeply flawed, but actually, it is a rigorous mathematical theorem. No one has ever discovered a flaw in the theorem.

What can be flawed (and often is flawed) is how the Nyquist theorem is used. For example, it provides a limit on what signals can be reconstructed without error. Any signal components above the Nyquist limit will show up at a lower frequency in the reconstructed signal. When I say it provides a limit, this is not to say that in any signalling system that limit is actually achieved. You have pointed out an example of why it might not be achieved because no perfectly bandwidth limited system exists in nature. That is not the fault of the theorem, but it is the fault of the physical systems that handle the signal.

Another widely misunderstood fact is that some people assume that even if the signal were perfectly bandwidth limited, the sampled result is a good representation of the original signal. This assumption fails when frequency components of the signal are close to the Nyquist limit, even for those signals that are below the Nyquist limit. Some people do not understand that all the Nyquist theorem says is that it is possible to reconstruct the original signal without error, not that the sampled result itself is an adequate reconstruction of the signal. I started a thread on this topic a while back, and from the responses it is evident that a lot of people don't understand that a sampled signal, in and of itself, is not necessarily an adequate reconstruction of the original signal.

I don't think we have a fundamental disagreement because you point out some valid points, but some who read your post may get the incorrect idea that there is something faulty about the theorem itself, whereas the fault lies in how many people think about and attempt to use the Nyquist theorem.

 

[guangong]

The real question is not which is sharper, film or digital. The question really is which is more enjoyable: dealing with electronic menus and cards of digital cameras and sitting in front of a computer monitor for hours operating keyboard and pushing mouse, or enjoying the tactile sense of shooting film and camera and whole chemical/physical process of development to finished picture. I use both digital and video for certain projects, but I enjoy shooting film and movies much more.
Which is more satisfying is an individual question of taste. Also, from personal observation, some folks are all thumbs with digital, others are all thumbs trying to load camera and development process.
The good side is the availability of choice.

 

[JWMster]

guangong: +100 Thank you! Choice is good. Truth is that digital is where the innovation $'s have gone, so many of the advances in photography require digital intervention at some point - even if it's in or after the scan of a negative. But the AF in current production tops what any film camera had. Digital serves certain projects better. Film can be more fun - but it can also take more time to do well. I'd warrant for what most snappers want in an image capture, the market has spoken well and the average snap with digital may well be better than the average snap was with film. Different issue. For my part, I also agree with Cholenpot: "Great picture!" tops "Sharp image!" The latter can almost be a fail... if that's the only thing going for the image.

 

[Cholentpot]

guangong: +100 Thank you! Choice is good. Truth is that digital is where the innovation $'s have gone, so many of the advances in photography require digital intervention at some point - even if it's in or after the scan of a negative. But the AF in current production tops what any film camera had. Digital serves certain projects better. Film can be more fun - but it can also take more time to do well. I'd warrant for what most snappers want in an image capture, the market has spoken well and the average snap with digital may well be better than the average snap was with film. Different issue. For my part, I also agree with Cholenpot: "Great picture!" tops "Sharp image!" The latter can almost be a fail... if that's the only thing going for the image.

Boring technically perfect image vs blurry dirty but intriguing?

I know what I'm choosing.

 

[Sirius Glass]

I was reading some previous comments on how film is still sharper then digital on some films, and how the Nyquist theory works into keeping digital from being better. This is for the same size of sensor/film. I was looking at some images of a landscape and it showed more definition in the trees, while the digital image was softer in that area. Same lens. I don't know how the Nyquist theory mixes into all this. Some films have around 100 or more lp/mm, especially B&W. Yet we see some digital sensors challenging medium format. And the digital image is cleaner without grain getting in the way.

So what is the actual truth on all this, or is there any conclusion?

It is real simple, until and unless the all four pixels that make up a complete GRGB pixel become the same size or smaller than grain, digital is and will remain to have less resolution than film, regardless of the number of post or threads stating otherwise. The so called sharper detailed digital images which claim to better than film appear sharper because of the over sharpened software that is showing nonexistent information.

 

[jtk]

Ask yourself "what do I think today's premier photographers use for their most challenging self motivated or paid assignments .?"

 

[alanrockwood]

One way to look at this is in terms of the modulation transfer function (MTF). As a rule of thumb, for digital cameras the modulation transfer function is fairly flat until it reaches a cutoff frequency, then it drops like a stone. Film often has a MTF that starts to drop off at a lower frequency than digital imaging systems, but the fall off is more gradual and still contains some information (i.e. is non-zero) at higher frequencies than many digital imaging systems. The upshot is that we often perceive a digitally acquired image to be sharper, even though a film image may show finer detail, albeit more faintly so the fine detail may be overlooked.

This is actually not a new idea. Even back in the old days, when Modern Photography published lens test results they would give the results in two collumns, one for resolution and one for contrast. The one for contrast more or less took into account that the MTF might be more flat up to a certain point and then drop off rapidly, whereas the resolution measurement indicated the ability to record finer detail if one accepts that the contrast in the fine detail might be rather poor.

 

[kb3lms]

As far as most personal photography is concerned, my opinion is that digital vs film sharpness just doesn't matter that much. You can take a poor photograph with the sharpest of systems and it is still a poor photograph. If whatever you are using works for you, then it's good.

Now there are certainly cases where ultimate sharpness is required. If we're going to the moon or Mars, then we want it all! But in the general case, with today's film or digital gear, the point of sufficient sharpness was passed decades ago.

Besides, if you are really concerned about getting the maximum, sharpness has probably has much more to do with the quality of the lens in use than the "sensor".

Somebody can take just as bad a picture on the moon as on Earth anyway.

 

[Cholentpot]

As far as most personal photography is concerned, my opinion is that digital vs film sharpness just doesn't matter that much. You can take a poor photograph with the sharpest of systems and it is still a poor photograph. If whatever you are using works for you, then it's good.

Now there are certainly cases where ultimate sharpness is required. If we're going to the moon or Mars, then we want it all! But in the general case, with today's film or digital gear, the point of sufficient sharpness was passed decades ago.

Besides, if you are really concerned about getting the maximum, sharpness has probably has much more to do with the quality of the lens in use than the "sensor".

Somebody can take just as bad a picture on the moon as on Earth anyway.

Bingo.

The whole conversation is moot. Look how popular point and shoot cameras have become. People can care less about sharpness after a certain point. The best selling camera of all time had a plastic lens, it got the snapshots without an issue. Never had one fail on me and I used dozens of them. And they still make them.

 

[KenS]

Being 'old and grey' I still prefer larger format film over what I can 'capture' on my digital Coolpix which I do use as
when 'scouting' for 'possibles' to be exposed onto either my 4x5 or my 8x10 film at a later time when the light is 'just right'.
and I have the 'time'. The resulting film negative will be scanned.. perhaps resized and printed onto Pictorico OH film for making prints onto water-colour 'paper' using the 'archaic' print processes using my home built UV light source... a lot more 'work' (and time consuming) but also a lot more "Fun".

Ken