10
\$\begingroup\$

Possible Duplicate:
What megapixel value is equivalent to which ISO film?
Is it true that '80s 35mm photofilm had quality corresponding to 24 megapixels?

I know this is a little bit like comparing apples with oranges but I am trying to find the point where digital photography caught up with conventional film photography.

Today's high-end full-frame digital cameras have 35mm sensors which is the size of a conventional 35mm film. These two technologies (digital and analog) are fundamentally different but at the end of the day, the images produced are consumed alike.

Now, consider an 35mm analog image recorded on top-quality film with top quality equipment. What would be the resolution (megapixels) and color-depth of an equivalent digital image.

\$\endgroup\$
2
  • \$\begingroup\$ This is such an apples-to-oranges that I can't tell that the OP is really asking. When we enlarged 35mm film, we often got grain. It was considered artistic. It was, in fact, just noise, or the limitation of recording ability of the film. \$\endgroup\$ Dec 24, 2012 at 4:43
  • 1
    \$\begingroup\$ It depends on the film \$\endgroup\$
    – Fake Name
    Dec 24, 2012 at 8:20

2 Answers 2

16
\$\begingroup\$

Instead of picking arbitrary numbers out of the air, do the math to get some comparisons.

A "35mm" frame is 36x24mm in size. Look at the resolution spec for some films and lenses. Some films were rated at nearly 200 lines/mm, but some much less. There was a tradeoff between sensitivity and grain size. That added noise and lowered spacial resolution of more sensitive films. Lenses also cover a range. Let's say roughly 50 lines/mm would be "good", and 100 lines/mm astonishingly superb. Of course that's only at the optimum f-stop and camera mounted and held very still.

You said top equipment, so let's see what 75 lines/mm comes out to as a starting point. A "line" is actually one complete light-dark cycle, so you have to allow for at least 2 pixels per line width. So the 75 lines/mm becomes 150 pixels/mm, which means a full 35mm frame would have 5400 x 3600 pixels = 19.4 Mpix.

However before you run off and declare that the answer, look at all the judgement calls that went into ultimately getting that number, and that film resolution and digital resolution are in some ways a apples to puppy dog comparison. Film not only has some maximum resolution, but also grain. This is effectively noise added to the image. Pixels have some noise on them too, but this noise is random and occurs on the pixel grid pattern.

Film and lens resolution is a "soft" thing. Neither reproduces frequencies up to the limit perfectly fine, then suddenly mushes everything to the average after that. The contrast falls off with frequency, so the resolution spec is some arbitrary point along that curve. Usually the -3dB point is used. In contrast, what one pixel does is pretty much independent of its neighbors. The resolution is fixed and finite due to the pixel pitch, but that also introduces aliasing which is something completely foreign to the analog film process.

Before I switched to a digital camera, I used to scan negatives at about 9 Mpix. That wasn't a deliberate choice, just happened to be the limit of my scanner. However, at that resolution the grain was clearly evident. I now get effectively 12.1 Mpix from the same image area with my digital camera. I can tell you that subjectively the pictures from the digital camera look better than the scanned negatives. The pixel-level noise is much lower, mostly because the grain noise is gone. The digital sensor is also significantly more sensitive such that 9 Mpix would be a joke with "high speed" film to match the sensor.

The digital sensor also has more dynamic range. The camera has a 14 bit A/D internally. Of course you don't get 16k useable levels from every pixel, but you do get a lot more than with film. This opens up a lot more options when post processing. Not only could many of those things not be done with optical processes, but the original dynamic range captured on film simply wasn't there. I usually used color negative film because it had better dynamic range than slide film, but that was still well below what a good sensor can do today.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ judgement: Use a modulation transfer function for a more quantitative measure of "sharpness" and "resolution". This makes "lines" a measurable concept at a specified MFT. \$\endgroup\$
    – JDługosz
    Oct 19, 2014 at 9:35
  • \$\begingroup\$ When discussing resolution regarding color images from digital sensors, shouldn't the effects of a Bayer mask and demosaicing be taken into account? A Beyar masked sensor can't reach the same resolution in terms of the final image as a monochrome sensor with an equal number of pixels with no debayering applied to its output. \$\endgroup\$
    – Michael C
    Jan 30, 2017 at 13:56
  • 1
    \$\begingroup\$ @Michael: True. The Bayer pattern throws yet another wrench into the works. You get twice the special resolution in green than you do in red or blue. With good de-mosaic software, a picture can appear to have nearly the resolution of the sensel pitch, but then be worse in pathological cases, like nearly vertical or horizontal sharp edges. \$\endgroup\$ Jan 30, 2017 at 15:21
  • \$\begingroup\$ I'm pretty sure a while back I read somewhere that when shooting standard test charts with alternating B&W lines the resolution with a Bayer masked sensor comes out to about 1/√2 of the native resolution of the sensor. I'm fairly certain it was in an article comparing a 15MP Foveon sensor to a Bayer masked sensor that returned the same result with the test charts, so that has been a while back. I can't find it with a cursory search now, though. \$\endgroup\$
    – Michael C
    Jan 30, 2017 at 15:30
  • \$\begingroup\$ Well written! Here's what brought me to this SO and question: my grandparents are/were a bit on the hoarder side. I want to scan old photos they took from the 50s-70s. In many cases, I'm fairly certain they probably kept negatives, but I would have to tear apart their house to find them. So the question becomes: given the era and the likelihood that they probably used consumer grade film/cameras most of the time, what is the fidelity/resolution payoff of hunting down those negatives? If the "analog megapixel" difference between a scanned photo and scanned negative is less than say…15%, I won't \$\endgroup\$
    – Anthony
    Jan 18, 2020 at 2:49
7
\$\begingroup\$

That depends entierly on what you mean by the question.

If you mean the point where there is about the same amount of detail recorded, perhaps 5 MP, 8 bpp.

If you mean the point where you can make an enlargement the same size without feeling that the quality is lacking, perhaps 20 MP, 12 bpp.

If you mean the resolution that you can scan a negative and get out all the information, perhaps 50 MP, 16 bbp.

Those are of course just ballpark figures, and varies depending on what film you are using, how you measure detail, what you find acceptable, et.c. It should however give you an idea of what you could expect.

As you say, it's like comparing apples and oranges. We can blow up a negative to a size where you clearly see every grain, without seeing that as a lack of quality. When you blow up a digital image so that you see the pixels, most will feel that the quality is lacking. Generally a digital image needs a higher quality to be perceived as equivalent.

\$\endgroup\$

Not the answer you're looking for? Browse other questions tagged or ask your own question.