Megapixel equivalent of film?

[bob_peters]

Howdy

 

Does anyone know of anywhere that has reliable information regarding

the megapixel equivalent of film? I've heard anything from 12 to 40

megapixels mentioned for 35mm Velvia, and 120 for medium format.

 

I don't think there's any debate that film gives much better

resolution, particularly when enlarged, I'm just trying to quantify

how _much_ better out of curiosity.

[mattalofs]

"reliable information" is a relative term

 

<a href="http://www.google.com/search?hl=en&ie=UTF-8&q=megapixel+film">enjoy</a>

[al_kaplan1]

It's hard to make a direct comparison between film and digital. Digital can give an illusion of great sharpness from what might be a fairly soft image. The magic of "sharpening" will do that by putting a hard edge on some things that the lens rendered softly, but softness beyond a certain point remains soft, and detail that might be too fine to be resolved by the lens at a distance, such as hair or the thread texture of fabric, can't be sharpened because there's no information to sharpen. A photo that has no superfine detail is limited only by pixel size, just as a conventional photo is often limited by film grain. The transition areas in a photo, from resolved detail to detail too fine for the lens to resolve, can look unnatural in digital because of the rather sudden jump from smooth tone to "sharpened" details. We need to coin a term for this, like the word bokeh for describing the smoothness of out of focus areas.

[simon_bruxelles]

This question still puzzles me for the following reason: most comparisons between film and digital use a standard of 300dpi to give an indication of image size.

Thus an image printed at 300dpi should have the same detail, resolution and quality regardless of whether it was produced by film or digital technology.

However, if this is correct how is it that I can clearly tell the difference between images in medium format and 35mm, even on a postcard-size enprint? Does this not mean that the 300dpi standard is meaningless ... or are there other factors at play?

Hope I've explained myself adequately. Simon

[jbq]

It depends on so many factors (and especially on the scene itself) that it's totally impossible to really find one number that's a universal answer.

[Rob F.]

Well, as a starting point, how about this. A good film like Velvia ought to be able to resolve, I guess, maybe 80 lines/mm if not more. In inches, that would be 2,032 lines per inch. I'm not sure if that's equivalent to 2032 dots per inch. But I'm sure it's equivalent to better than 300 dpi, and that might validate Simon's point.

 

Another idea: 80 lines/mm would be 2,880 lines across the width of the 135 format; and 1,920 lines for the height. Multiply these two numbers, and you get 5.5 megapixels (or at least 5.5 "square megalines").

 

So If 80 lines/mm is accepted as excellent, would that mean that 5.5 megapixels would be equivalent? Any digital gurus out there?

[bas1]

Simon,

 

You've made a very good point actually, one to consider with the many film to digital comparisons in mind.

 

However the 300dpi limit comes from what the limit that the printers have in the digital world. They simply cannot deal with dpi's over 300 or they are wasted in that you cannot tell a difference between a 300dpi print and a 600dpi print. I'm not sure if that is still valid with today's printer technology. Don't follow that anymore, not digitally interested really.

[neal_shields]

The FBI ran extensive tests and concluded that 35mm 200asa film captured about the same amount of information as 16 meg cameras.

 

http://www.fbi.gov/hq/lab/fsc/backissu/april2002/swgitfield1.htm

 

Other testing shows that you are still getting usefull information off a 35mm negative at 8000 dpi scans:

 

http://www.normankoren.com/Tutorials/Scan8000.html

[jbq]

Neal - beware, Norman Koren's curves are interesting, but don't forget that those are extrapolated simulations, not real-world testing, and that without a noise curve there's little point in having a signal curve.

 

For me, the question sometimes comes the other way - how big a piece of film do I need to have a look comparable to my 10D?

[bobtodrick]

I recall seeing a test in a magazine awhile back where-in they compared Velvia 35mm, resolution-wise to digital. They came to the conclusion that you would need a 22 megapixel sensor to duplicate the film.

[steve george]

Here you go: "real world" tests show that on 100 speed film you can get 40 megapixel equivalent. Very interesting article...particularly interesting how the ability of digital to ever get close to film is limited by physics and not technology so digital is probably approaching it's peak quality already in terms of resolution. (And for as long as a medium format digital back is going to set you back 20,000 GBP for 22 megapixels and severely limited ISO you have to ask why anyone would bother.)

 

http://medfmt.8k.com/mf/filmwins.html

[terry_rory]

Unless every shot is taken with the camera atop a heavy tripod with a cable release and the (top quality) lens at its optimum aperture (usually around f8 give or take a stop) and the film developed and enlarged (to at least equivalent of A3 size) with great care by the best labs then such resolution will always be theoretical.

 

It is doubtful any of us will ever see the difference between a good (6 megapixel+ APS sized or full frame sensor) DSLR photo and one taken with film. The difference will NEVER be visible on a monitor and a good digitally derived 16x20 print and a good 16x20 traditional print will be indistinguishable at normal viewing distance.

 

(I am just talking about resolution as the debate still goes on about colour fidelity and other even more 'esoteric' properties that some people claim to be able to perceive, even on a computer monitor!)

 

Haven't even touched on film flatness. (One variable a decently built DSLR doesnt have.)

 

It really does not matter. If you are a good photographer then you are a good photographer whatever you use (assuming it is reasonable gear) whether digital or film.

 

Resolution in lines per millimetre (just like BHP and 0-60 acceleration stats) is meaningless in itself and will never have any place in determining what makes a great photograph.

[terry_rory]

The magapixel 'race' (especially with super compacts) is pathetic. Cramming 8 megapixels or more onto a tiny sensor where they cannot operate without significantly bad noise/signal ratios above ISO 100 (caused by having to amplify the output from tiny pixels) is stupid when the option is there to have APS sized 6mp sensors that give satisfactory results at ISO 200 - 1600.

 

It is just treating consumers with complete contempt by making them think that the amount of pixels is the sole determinant of quality!

 

We should not sink to that level of thinking with our 'Super Giant' film based sensors with 40mp is 'better' than your puny 6mp arguments. We know better than that stupid 'high st chain store' mentality.

[scott_eaton]

This question is impossible to answer because they only point of reference can be the final print. This of course brings a lot of heated debates since most digital shooters are more concerned about what they are doing with the image while film shooters tend to be obsessed with what film is doing. Another problem with equating film to megapixels is that unlike digital capture, film does not record information in as linear a fashion. When one makes all these resolution claims about film, 99% of them are based on high contrast monochrome test targets. Gee, last time I shot 6x7 slide film, I didn't shoot high contrast test targets, and I'd expect the intelligence of the photographers here along with Neal Shields to take a moment and grasp that concept.

 

Velvia or Kodachrome or Tri-X might have N lines per millimeter of resolution when shooting pairs of high contrast monochrome test targets, but this is the ideal and not the real resolution of the film (or any film for that matter), which falls off dramatically when you start dealing with color and low contrast test pairs. Digital capture isn't prone to the variables above within it's envelope. The non linear response of film is really what gives it it's character and along with comlaints that digital capture looking so sterile at default settings because it's so linear.

 

Low speed color and B/W films can actually hold quite a bit more information than digital capture. The old Agfa holographic films I used to work with had a theoretical resolution of several gigabits per 4x5' sheet, and that was conservative. The problem is extracting all that information from film. Even scanning brings about limits because if you've ever worked with MF or LF films scans, the difference in quality between 2000dpi and 4000dpi is quite dramatic. With digital you just hit a wall, and that's it. Which brings about the controversial examples of high end digital capture being compared with film images when printed large, and most of the conclusions giving the benefit to digital capture much as it pisses of film fans. It's really quite simple and shoulnd't generate that much controversy. When I make a digital print from my Canon 10D Via a LightJet or FujiFrontier, every bit of information gets translated into that digital print in pristine form. When I make a print from my RB67, I either have to scan the film, or make a direct optical print, which regardless of how you slice it puts film through another analog phase and reduces it's efficiency. This is why even the best optical color print I've made by hand from 35mm film don't look nearly as good as my 10D captured digital prints, and images I post that are flatbed scans from my digital prints look better than most film scans from our hardened anti-film crowd. It's not that digital capture has more resolution, but simply that information gets translated into a digital print with several orders of magnitude greater efficiency than making a digital or optical print from a piece of film. Optical prints from film, especially commercial ones, are subject to such an extreme range of quality that I can't begin to list the variables.

 

About the only guys that are getting the most information extracted from film via optical printing are the guys are in their home B/W darkroom and using good technique and fairly low resolution B/W films. I find that ironic, but it doesn't change the fact that the majority of Cibachromes I've seen in my lifetime were printed with such poor technique that the full resolution of the slide film is irrelevant and print quality not even close to the 300-400dpi of popular digital printers.

[neal_shields]

Most people who never visit museums will, in the near future, probably never see anything better than an 8x10 Frontier print at 300 dpi. Given that most consumers don't need anything better than 2400x3000 pixels or 7 meg. Any more than that and the printer is going to throw it in the garbage. I also agree that most people won't be able to tell the difference between an 8x10 Frontier print from a 5 meg camera and an 8x10 Frontier print from a 7 meg camera.

 

Also for most people it is a moot point how much information is on a 35mm negative because again if that run it through a Frontier printer, the printer is going to take a digital picture of their negative at what ever resolution it needs to make what ever sized print is ordered.

 

Most people that arn't aware of what is happening will conclude that the prints that get from film aren't any better than what they get from their digital camera and they will be right.

[scott_eaton]

Most people that are engaged in the actual photographic process are more interested in producing good images vs petty debates about what museums do and how much resolution film has. In regards to museums, I've printed many, many exhibits for local shows via film and am well aware of what the standards are. Those standard are certainly higher than a fuzzy B/W print or Cibachrome with mushy grain in the borders of the print and a matte the size of a picnic table to force your attention to the boring image.

 

In my lifetime I'v printed more 16x20 and larger film images for clients that will sit on their family wall as a cherished heirloom than anybody else in this thread. I no longer print custom, yet when I do consulting work it's astonishing to see how many of my images I personally printed are in corporate board rooms and offices. As it is, if I leave any of my digital prints out in the open they are snatched up and gone within a few minutes, which I doubt I can say for what Neal is producing. Like the other challenges I've made on the topic, I'd be more than happy to put my 8x10 10D images printed on a Frontier up against the best color images from 35mm Neal has, and let an objective crowd be the judge. Most people will swear it's MF, while Neal and the others will still be swearing at digital because they don't have any other excuses.

[bob_peters]

All very interesting and - for once - a rational debate on the difference between film and digital.

 

So if physics is already reaching the limit that digital can cope with it will be interesting to see what the future holds - it's the myth of constant improvement that's driving sales for people buying their 2nd or 3rd digital camera. I get the feeling the current digital surge could be a fad and once the novelty wears off and it'll remain useful just for pros on tight deadlines and hobbyists who don't mind using PC's to get what they want. The middle of the market will be a lot of film users though I think - a lot of people will always want the simplicity of taking pictures, posting off a film and getting the pictures back without the hassle of queuing in a shop to download everything in order to print them.

 

And a lot of us love the sight of a 6x6 Velvia slide projected on the wall :-)

 

Reports of the demise of film are seriously premature (I reckon anyway).

[neal_shields]

"Norman Koren's curves are interesting, but don't forget that those are extrapolated simulations"

 

In several places on that link he says something to the effect that: "The photograph at the right is from an 8000 dpi scan".

 

I assumed that the photographs were actual scans and that some of his data was developed from formulas.

 

I could however be wrong.

 

However, in my own testing which is subject to interperation, I also get results that lead me to believe that 35mm film is equivilent to about 10 to 15 meg when everything is perfect. More importantly however as I go up in format I don't see significant losses. I.E. at 4x5 I can probably get over 60 lp/mm to film near the edge of the frame. Which for me is about a 25% loss from 35mm.

 

P.S. Corvettes will beat Subarus through the 1/4 mile all day long, every day.

 

The point of hard data is not: "do you need the capibility" or "will you use the capibility" but to give you the ability to make decisions for yourself.

 

Most people are probably incapiable of utilizing more information capture ability than is available with a 5 meg digital camera.

[hugo drax]

It all depends but the one real factor that matters is the final print and not all this mumbo jumbo theory. if you want wide exposure lattitude then you need film unless you budget for a Leaf or Phaseone back and a Hassie. I doubt film will dissapear just like Acoustic guitars still exist today even with the advent of electric guitars and synths. Different tools for different jobs I think its silly to ponder this whole MP issue, The final print and client satisfaction is what matters in the end.

 

Now on a side not I did a run of 12x18s on a Frontier (Aerials) that were mistaken for MF output and they were impressed with the color accuracy,saturation and the lack of film grain on the final print. Looks like Digital is in for aerial photos unless you want a print the size of a billboard but how many people ask for that? I think for general work Digital is moving quickly but for fine arts I think film is still the way.

[gerry_szarek]

One is issue not talked about is dynamic range. My canon 10D has about 6 stops nominally (yes I know you can get 8 but they are so noisy to be useless). Descent color negative film has 10, with B&W TriX in at +12 stops.

 

However I will say my 10D is much cheaper to use than the Leica or Hexar just because of film and developing cost, and for most kid shots that end up being 4X6's you can't tell the difference.

[scott_mcloughlin]

Har har. Inevitably someone says something like (1) I can take better pictures than you - let's have a race and (2) 4 out of 5 dentists surveyed say that it's only the results that count, anyway.

 

The first point is completely irrelevant, and the second point is completely WRONG.

 

Tools are tools. Results are results. That fact that good or great results can be produced with a wide variety of tools of differing quality, attributes and tolerances (1) is undisputed in many fields of human endeavor, including photography and (2) still reflects little or none on the tools themselves.

 

Tools need to be compared with metrics.

[mike dixon]

<i>Looks like Digital is in for aerial photos unless you want a print the size of a billboard but how many people ask for that?</i><P>

My ex-girlfriend and her partner have an aerial photography business--most of their print sales are 20 x 24 and above. The extra dynamic range of negative film is also good since they don't have the option of always shooting in ideal light.<P>

(Good thing for me since I'm selling them a backup 645 body and some lenses--it'll help pay for a 20D!)

[morocco]

the megapixel equivalent of film is dependent on the scanner used to digitize the image.

 

obviously there could be a point where a scanners ability to discern detail is greater than the detail in the film, but i'm not sure if thats attainable or has already been attained.

 

the simplest answer to your question that i can give is that my canoscan fs4000us scans a negative or slide with 5700 x 3700 resolution (roughly, this is from memory). 5700 multiplied by 3700 is 21 megapixels. the minolta 5400 should get you more megapixels.

 

do these scanners approach the ability to get all info from film? i don't know. if more is retrieved from film, does it even translate into better prints? not at 8x10 prints size, of course. but how about 12x18?

 

again, a 24x36mm negative gives at least 21 megapixels.

[hugo drax]

20x24 and above film is still still the best bet :) but if you get P25 for that 645 you can go digital at that size level. (16bit/12fstop DR) for a small fee of course :)

[hugo drax]

You really need to go to a Tango/Primescan if you want to get the full range of your film.