OK, how many mega-pixels equals a LF negative?

greenspun.com : LUSENET : Large format photography : One Thread

Although I do own a digital camera, I am tired of friends that know about my addiction to LF photography telling me about the next break-through in the digital world, and how it just makes film more obsolete. Just as a comparison, can someone guesstimate what a quality 8x10 negative contains compared to a digital image file?

-- Bob Krantz (krantz@alaskalife.net), February 12, 2002


Even a non quality 8x10 film image has probably somewhere in the range of 600MB to a gigabyte (1,000Mb) worth of information. But in reality such comparisons are pointless. Each medium has its own unique assets and drawbacks.

-- Ellis Vener Photography (ellis@ellisvener.com), February 12, 2002.

Ellis: with how much color depth? My estimates are closer to a bujillion.

-- Chad Jarvis (cjarvis@nas.edu), February 12, 2002.

I've heard numbers up to 20MP for a 35mm frame, but even if we cosider that wildly optimistic (I don't) we're still stuck with at least 14MP as a bare minimum to reach 60lp/mm (we actually need even more, but this gets us in the ballpark and gives the digital people a chance). Cosidering the square area only gives us (80/1.5)x14MP = 742MP minimum (1GB using the 20MP figure). This gives us the raw capacity of the recording media only, and does not account for diffraction, circle of confusion, film plane accuracy, etc... Even having said this I'm sure that there are people here scanning 8x10 on a drum who can probably state that a 300MB file extracts 90% of the information on an 8x10 chrome - I haven't yet had that experience (unfortunately).

-- Wayne DeWitt (wdewitt@snip.net), February 12, 2002.

What does it matter? Ask them to take their pixelograph recorders out & do a nice 8 your star trail exposure or to take the original negative & make a nice 8x10 platinum print from it. There is a lot the pixel recording machines can not yet do.

-- Dan Smith (shooter@brigham.net), February 12, 2002.

Hi Bob

A fine grained 35mm (24x36mm) slide has about 40-50 millions of grains on it, as I readed somewhere in the past, so now make the mathe by yourself.

But as Ellis already stated, every side has his pros and cons. For example you have not to ubdate your LF eguipment every year but in the digiworld you have to do it!

Good pixel counting!

-- Armin Seeholzer (armin.seeholzer@smile.ch), February 12, 2002.

As Ellis says, there are a lot of variables, not the least of which is color depth. The DCS 460 camera that I use has an array that is approx. 1 sq. in. From that, the system creates a 17 Meg. file, but that is with only 8 bits of color. To match an 8x10 chrome, my best guess would be 2 - 3 Gig.

-- Bruce Wehman (bruce.wehman@hs.utc.com), February 12, 2002.

Between 400 - 500 mega pixels. The newest 4-5 megapixel cams can make an OK 8X10. An 8X10 resolved to 60 lp/mm can generate an acceptable 80X100" print at 10X That's 100 of the little 4-5 megapixel chips.

-- Jim Galli (jimgalli@lnett.com), February 12, 2002.

Does it depends on the resolution (pixels/inch) when scaning the negative? i.e at resolution = 300.. we will have 8*300=2400 and 10*300= 3000 ..now 3000*2400=7200000 which is which is 7.2Mb...seems way too low..because photoshop said it's a 20.6 Mb file....any opinion?

-- dan n. (dan@egmail.com), February 12, 2002.

Based on my estimate of roughly 175 thousand angels on the head of a pin (that is, 175 kilo-angels), 17.3 pinheads/pixel and 400 pixels/inch, it is approximately 17 million bakers dozen pixel- pinheads per picoliter.


-- Nathan Congdon (ncongdon@jhmi.edu), February 12, 2002.

the number of grains contained in a particular piece of film is not actually very relevant in determining the amount of useable information for scanning purposes. the reason is that as you zoom in on an image (with increasing scanning resolution) you reach the limits of the optical photographing process long before you reach the film grain. in other words, no camera/lens combination in the world is sharp enough to resolve lines down to the sharpness of the film. the sharpest line you could ever get on film would be a gradient ten or twenty grains wide; you would never see a line or any other detail that was one grain in dimension. so for scanning purposes, all you ever need to scan to is the sharpness of the lens; scanning all the way to the film grain just gets you a really large fuzzy, file.

the lens/camera systems in 35mm are much sharper than for 4x5, which in turn are sharper than 8x10 (due to lens sharpness, proximity of film to lens, flatness of film, thickness of film, and other factors). so you'd want to scan a 35mm original at higher resolution than a LF original.

for 35mm a 4000 dpi scan (100 MB in 8-bit color format) will take you past the sharpness of the lens and pull out all available information. for grainy films this number will be substantially lower.

for 4x5, 2400 dpi (a 300 MB file in 8-bit format) will pull out all useable information from the sharpest films. i'm not sure for 8x10, but i suspect it'd be in the 600 MB to 1 GB range.

so, digital backs are nowhere NEAR to producing the sharpness you can get from film. but, their far greater tonal range can produce some amazing images that would be difficult to match on film, so as everyone else has said here, each medium has its advantages...

~chris jordan (Seattle)


-- chris jordan (cjordan@yarmuth.com), February 12, 2002.

From a technical standpoint the question could be posed in these terms: if you need to distinguish if a spot on the wall of the nuclear waste tank is a hexagonal bolt head or a round paint spot, how many pixels do you need to reliably determine the shape of the object? Or, what is the optimum lens, format size and film to do the same? This is not an academic exercise. Much of what passes for sharp digital imagery is a mathematical construct that pleases the eye but is filled with empty resolution. Not that it matters too much for landscape photography, but it is an issue if we intend to use digital to accurately document a thing or an event. One measure I have never seen defined is how small a difference between two colors (or two shades of gray) can a system capture. I suspect this figure of merit could be expressed in CIE terms. If film and digital can record differences smaller than human perception, the question is moot. If not, we have a basis for comparison. It also appears that the colors recorded in a digital system (the color space) are for the convenience of the computer, not necessarily allocated optimally for human vision. The color information recorded by film may be better suited for our eyes, especially color transparency film.

-- Andy Eads (aceads@3-cities.com), February 12, 2002.

In answer to the question "how many mega-pixels equal a LF negative"...

Noone really knows because they have not yet produced enough pixels... -Dave

-- Dave Richhart (pritprat@erinet.com), February 12, 2002.

i can pull an 800MB 16bit file or a 400MB 8 bit file when scanning a 6x9 image on my nikon coolscan 8000 at 4000dpi. my printer, an epson 3000 will let me print that image at 16"x24". what inhibits me from moving to sheet film formats from roll film formats is that great sheet film scanners start at $16000.00 and get impressive at around $70000.00. perhaps in 5 or 6 years one will be able to purchase a digital back that captures an image with 11 or 12 stops of latitude in a small fraction of a second, and at a resolution that will look good over the couch. and by then, printer drivers will all be 16bits (like the Cone Quadblack plugin) and all that technology will be judged useful. until then, i wallow in the great good fortune of having so much fun with what i have, with little thought about what's next. vincent

-- vincent bilotta (amplexis@surferz.net), February 12, 2002.

Think in terms of megapixels (mp), or millions of pixels - that's how CCD (used in digital cameras and backs) capturing is measured.

A 600dpi scan of a 4X5 nets an approximate equivalent 6 megapixel file. That same image takes up about 17mb of hard disc room on a computer. 6 mp cameras are available now - two years ago a 2 to 3 mp camera would have been king of the hill.

Within two years a digital camera the size of a cigarette pack will capture 20 million pixels (he said confidently). That 20mp will roughly equal a 1200dpi 4X5 scan.

The CCD industry is going through the same progress as other digital or computing devices have for years - an approximate doubling in capacity every eighteen months. So your friends will be right within five to six years, based on a 2400dpi 8X10 scan - assuming you agree there is no point in scanning an 8X10 at higher than 2400dpi.

No doubt the number crunchers at Fuji and Kodak have done this homework as well - the bulk of both firms R&D $$$ is now spent on digital products and solutions.

-- Michael Mahoney (mike.mahoney@nf.sympatico.ca), February 12, 2002.


A 1st rate 4x5 system with real optics and an excellent film probably can reasonably expect about 25 lpmm across the entire sheet of film. This works out to 50 dots per mm or 1270 dots per inch (dpi). This in turn leads to about 5000 X 6000 pixels. 8x10 will probably be more like 15 lpmm yielding about 6000 X 8000 pixels.

As soon as you try to get the information off the film via a scanner or enlarger the effective resolution will be degraded further.

I for one look forward to digital when the cost and convenience gets real.


-- Jim Bancroft (bancroft@cox.net), February 12, 2002.

Mr. Mahoney (Baloney!)

My 300dpi scans of 4x5 color transparencies produce a 27-29MB file - 600dpi is in the neighborhood of 40-70MB (and more).
Besides, trying to determine what the resolution of a film is, based on a scan is utterly ridiculous!

-- Matt O. (mojo@moscow.com), February 13, 2002.

We had some fun with this one. I certainly appreciate all the conceptual and subjective differences between film and digital, and obviously the contrast of compact camera vs. LF rig. For me, and I suspect for many, even if (when) digital technology surpases LF film resolution and rendering, the craft of exposing and producing traditional negative and prints, and the meaning it has for artistic vision and realization, will keep me in the darkroom for a long time. In any case, these numbers should make for some stunned looks when the next digital geek starts bragging in my direction.

-- Bob Krantz (krantz@alaskalife.net), February 13, 2002.

Chad: Is a 'bujillion' anything like 4 or 5 thousand? Because that's closer to the number of individual colour levels you can create from 3 layers of cyan, magenta and yellow dye clouds.
Each blob of dye in a film is a pretty uniform density, and it's only by overlaying those fixed densities, and dithering them side by side, that you get an ILLUSION of continuous colour tone out of film.
From examining colour film under a microscope, I reckon that each colour layer is capable of only about a 16 to 1 colour variation, at the level of the individual dye clouds.

Please, let's stop fooling ourselves that film is the ultimate imaging medium. If only so that Kodak, Fuji, Agfa and Konica don't rest on their laurels.

-- Pete Andrews (p.l.andrews@bham.ac.uk), February 13, 2002.

Nathan - Could you convert that number out of metric? I have such a hard time visualizing metric units of measure.

-- Joe Lipka (joelipka@earthlink.net), February 13, 2002.

RE: Mr. Mahoney (Baloney!)

(( My 300dpi scans of 4x5 color transparencies produce a 27-29MB file - 600dpi is in the neighborhood of 40-70MB (and more). Besides, trying to determine what the resolution of a film is, based on a scan is utterly ridiculous! ))

Matt O (Turkey!):

I'm happy you have a high enough quality flatbed and scanning skills to draw that amount of information from your 4X5's. The final file size depends on many factors, but that was not the original question.

If you read my post before spouting off you'd realize no attempt to determine the 'resolution' of any film was made.

-- Michael Mahoney (mike.mahoney@nf.sympatico.ca), February 13, 2002.

Jim Bancroft noted above that: "A 1st rate 4x5 system with real optics and an excellent film probably can reasonably expect about 25 lpmm across the entire sheet of film."

I routinely get closer to 50+ lpmm with TMX and Velvia using modern optics on a Tachihara. More in the center. If it's less, it's due to diffraction from stopping down so far to get the required DOF. (beyond f/22) This level of resolution is easy to see with a 10X loupe.

-- Gary Frost (gary.frost@onemain.com), February 13, 2002.

Hi Bob

There may be some useful info here for you on 6x9 and LF digital and digital printing.


Kind regards

-- Peter L Brown (photo_illustration@bigpond.com), February 13, 2002.

Michael - My point is - that to base any assessment of a film when scanned, to a digital image is useless, because the film has been digitized and affected by the limitations of the equipment involved in doing so.
The only real test would be to compare prints of identical images(however large) from each at extreme enlargement.

-- Matt O. (mojo@moscow.com), February 13, 2002.

Moderation questions? read the FAQ