Digital Resolution and Film using Leicas...

I found a rather controversial article on Luminous Landscape at

http://www.luminous-

landscape.com/reviews/cameras/d30/d30_vs_film.shtml which compares

shooting Provia 100 using Canon vs. the D30. I was rather surprised

and yet convinced by the results using the Canon equilibrium.

 

The question is: How scientific is the test? What do people make of

this conclusion? Is it based on supposed eyeball resolution or

measured resolution? I wonder what Erwin Puts would make of this

article or dismiss it entirely.

 

One of the main reasons is that I've been shooting film a lot and

hope to add digital shooting to my outfit for clients--especially

engagement photos for the church I attend because of the fast

turnover (I'm personally biased towards Canon digital which is

superior to Nikon in many respects... D30 + Leica is a possibility).

 

I wonder whether this Luminous Landscape would be negated if we shot

a Leica R8 with APO-Telyt against the D30 with a Leica lens and R-

adapter. Anyone up to the task?

 

(And no film vs. digital bashing again... pleez).

Don't throw away your Leica's yet - all we see here is the fact that current film scanners still can't quite get all the detail off 35mm film. I have a Nikon 4000ED and an Imacon 3 neither can differentiate detail on my Provia F trannys that I can see with a 15x loupe. BTW - the Nikon can get more detail (in the center of the frame) than the Imacon.

Point of that article is - digital produces better looking prints from lower resolution images (technically speaking). It is like classical TV and satelite TV - you can see the difference.<br>

Bottom line: film has better resolution and digital has better signal/noise ratio. Depends on subject and film used. I'm quite sure digital will have loooong way to beat TMX-100 behind good lens for resolution, but to extract this resolution and get it onto paper, this is another thing..

I have still to locate a clear concise and convincing test comparison between 35mm digital and film. I understand that to achieve 50 lines per millimeter from a scanned negative - i.e. an average lens performance -requires a scanner which has a minimum of 13,000 dpi. To achieve an LPMM of over 50 requires an even more powerful scanner. If you have a lens that can resolve say 100 LPMM on film then it seems that you will not be able to match the performance on digital for many a year.

Tony, you understood totally wrong. 4000DPI scanner has theoretical resolution of 4000/2.54/10/2=78 lp/mm, practically cca 5-10% less.

If you will not suffer from health problems or enviousness, you can have look what canon 85/1.8 does with TMAX-100 from f/1.8 to f/16 scanned at 4000DPI. With 40x microscope you can see WAY more. This is of course tiny part of negative.

This is very interesting - 4000dpi is the resolution in file it produces i.e it will produce a file with 4000pixels per linear inch. However the actual resolution is much less as it's compromised by optics and other mechanical/electronic errors. I have yet to devise a test as to the actual resolving power of my Nikon 4000 but I would be very suprised if it came anywhere near resolving 4000dpi!

Petr. I not a mathematician but I'm sure you're right. The info was given me by an experienced engineer from a computer manufacturer who said that they were researching three difficult areas to improve digital output, pixel size, print speed, and scanner dpi. He said there was no point in using high quality lenses at present because no current scanners could get near the margin using 35mm.(July 2001)

I have no reason to doubt his information.

i've seen similar film vs. digital tests, i'm not impressed as i

don't think the test method is valid

 

i don't know (or care about) the answer to the question about

which is 'better', just that this sort of test is not how to find out

 

the test is showing whether a direct to digital image (that was

then processed to sharpen it, using a method designed for

sharpening digital images) is better than a *scanned* copy of a

direct to film image

 

it is not comparing digital to film

 

leaving aside a lot of other issues...

 

the scan was done at a resolution considerably less than would

be needed to resolve the smallest dye particle present, so data

were discarded, this was then processed using the same

sharpening method, i.e. one intended for images made of arrays

of pixels, not random splodges of dyes

 

sound fair?

 

how about we flip things around, we'll use a film writer to image

the *unprocessed* digital image onto 6x4.5cm film (to

approximate the higher pixel density of the scanner used in the

first test vs. the d30), then compare under a microscope to see if

we think the 35mm film image is better/worse than the 6x4.5

digitally imaged film version

 

i'm not bored enough to try this, but i think i can guess the

answer

 

film is analogue, digital isn't. tests that convert one to the other

need to be very carefully designed, and performed with great

care, to eliminate error and avoid introducing bias; i don't think

the test concerned is even remotely adequate to answer the

question (and i still don't care anyway :-)

As someone who started with rangefinders at 12-13, then moved to SLRs, then (about 2 years ago) purchased Olympus E-10, and now use old rangefinders almost exclusively, I can tell you this.

 

1. Resolution is NOT the only or the most important criterium for comparison. All digital cameras currently (with one very pricey exception?) have small sensors, i.e. while your lenses create images that look as if they were taken with 35-50-135 35mm equivalents, the image is created with an actual wide angle, with extremely large DOF. Therefore digital images are flat and sharp all through, while 35mm images are soft and 3-dimentional because of the selective focus.

 

The unpleasantly drawn digital images are the reason I turned back to film (and rangefinders in particular, because I mostly shoot in the street). THAT is the biggest problem of digital, not the resolution

 

2. The resolution of a, say, 4 Megapixel image does not consist of 4 million red values, then 4 green, and 4 blue. It consists of roughly 1 million red, 1 mil. blue, and 2 mil green, which are combined using INTERPOLATION to get the result: the sensor is MONOCHROMATIC, and so to create colour, small filters are used in front of the pixels, to colour them into RGB. The matrix is usually built in a R-G-G-B manner, that's where the rough calculation comes from.

 

The "true" sensor has been announced (foveon), but is not available yet to the consumer.

 

3. The biggest advantage of digital, one that cannot be even understood by the film-only people is the ability to see your result immediately on the small screen. In about 1 year a photographer (if he is capable of critically thinking at all, that is) progressed probably as much as a 3 to 5 years of experience of "blind" shooting with film would require.

<I>I have a Nikon 4000ED and an Imacon 3 neither can differentiate detail on my Provia F trannys that I can see with a 15x loupe</i><P>I tend to agree that the limits of electronic film scanning don't always approach their stated cliams and the Nikon vs the Imacon comparison is a good example. However, I don't find this relevant at all to the discussions on Luminous landscape. A piece of film is only as good as it's next step of reproduction, including A/D conversion, and given an analog/conventional route that next step will more than degrade any inherent/theoretical advantage that film has over digital sampling, be it from the original piece of film or direct digital capture. That's the problem. <P>I've come to call this the "Heisenburg film effect", where film proponents argue the superior resolving power of film over current digicams, and in absolute sense they are right. The problem is that you have no way to extract all this information into another media, especially a conventional optical print of R-type method, without losing a good deal of this theoretical resolution and dynamic information. I, for one, could care less how pretty a slide looks on a light table unless I can reproduce it in a more practical form. This is what Reichmann is getting at, and good for him for offering a pragmatic comparison vs theoretical LPM chart comparisons and other mush the often times I can't tell these debates from stereo equipment reviews or CPU benchmarks. If a photographer makes 8x10's on a frequent basis, then evaluating the final quality of 8x10's as a point of reference makes sense, and I can't find much quibble with Reichmann on his results.<P><I>I wonder whether this Luminous Landscape would be negated if we shot a Leica R8 with APO-Telyt against the D30</i><P>Leica owners often claim 'MF' quality with their 35mm optics, but as somebody who's made a LOT of enlargements from virtually every type and camea made I offer them a serious recommendation for a reality check. Also, I don't quite understand why the type of lens used on a camera would effect it's scan quality. Handholding an SLR mounted with a 28-200 Sigma wide open at 1/30sec while driving an ATV on a bumpy road vs a tripod mounted APO lens would offer some subjective discrepency, but lets get real.

First of all id like to say the provia shots look like shit, i cant imagine provia 100F giving these poor results. Also scanning an image(an anologue/digital conversion) will reduce image quality(fair comparison?).

 

Ok lets analyse what this guy did,

 

Slidefilm and b&w film have high contrast, auto-leveling is not only nonsense if the film provides enough contrast, but will increase noise in black parts. Also if you have a neutral color rendition, auto-leveling is also nonsense. The really nasty thing is that he used autoleveling while scanning the slidefilm(a linear setting wile scanning skips any compensations including auto-leveling). The D30 was used at RAW mode, no auto-leveling, white balance or something like that. Then he says:

 

"Again, to remove subjectivity as much as possible the only other adjustment made to either image was to use PhotoShop's Levels in Auto mode."

 

He wants to remove subjectivity(HELLOOO!! *$%**^%%*%$##^%%@) by doing another unnessecary auto-leveling on the scanned Provia because he needed the auto-leveling on the D30 image! The contrast 'overkill' and the washed out colors are a result of this double autoleveling, try it if you have a filmscanner and some slides at home.

 

The 34 mb scanned file smells like a 8bit/channel to me, the canon D30 is 12bit/channel, difference in color depth are quite big between those two.

 

If i would know more about the lens and those cameras he used i would probably found more 'tricky' things.

 

When i reread the 'tone' in which he presents his findings i cant get over the feeling he is quite pleased with his results, not scientific or objective AT ALL. Also normally all information is provided, im missing alot of technical details.

 

Greetings,

I'll continue to shoot film, for both personal and commercial

work, until such time as digital capture offers the potential that

film does, relative to my needs. Current digital capture and

interpolation doesn't come close to the potential of well crafted

film (with the exception of state of the art multi- pass/stitching

units) , but the scan quality of current desktops evens things out

so that a 4000ppi scan to 8x10 epson print is about equal to an

8x10 from a good 6meg capture. As desktop scanning

technology evolves, the well crafted film I have on file can offer

better and better output quality, whereas the 6 meg. capture is

still going to be 6 meg. Assuming that interpolation software

gets better and better, as it will, and that capture technology gets

bigger and better, as it will, a time will come when there will be

no argument about quality difference, only aesthetics. When that

happens, which I estimate at about 5 years from now, perhaps

less, I'll happily go to digital capture for practicality, and decide

then what the aesthetics offer. For me, that time is still in the

future and I'd rather have a great piece of film in the can now.

I think it is clear that slow film is superior at capturing detail than current "35mm" digital cameras - the real point is whether it makes a difference in the final print of up to 11 x 14. His opinion is that they are pretty well neck and neck, but he does have a few proviso's to it. It is not so much that he is saying that digital is better, but that it is pretty well as good and offers other advantages. Personally, I remain spectacularly uninterested in digital cameras, because they are very expensive, require lots of batteries, have a mag factor on all lenses and quite honestly they just do not fit my workflow - I have a system and I like what I have been doing for the last 20 years too much to change it so I can sit infront of a d....d computer all night. Reichmann does make the point that the D60s files need more careful and extensive tweaking than scanned film. Exactly. I take slides and print only a few once in a while - there is nothing like this in digital yet so I see no purpose in changing my system. Digital cameras are also large brutes that are as ugly as sin and aggressive.

 

I like the LL site, but new products are often very favorably reviewed -- only to find that when the next model comes out the old one is condemned as no longer much use. Still film remains strong and in some ways above it all. But ultimately, as Scott says, if the digital camera produces great photos at the size you like then why not use it, if it fits the way you work and you want to obsess about constant upgrades of storage media, software and so on. Film just carries on...and on...It is rather nice.

If as true leicofyles (or leicoids?) you want to feel better by proving to yourself that you have the BEST, change the record.

 

Film versus digital argument fills tens of thousands of internet pages with monotonous uninventive drone. The major note in that collective hum is "resolution". But you are on a shaky ground here, as Luminous Landscape illustrates. And who needs that extra resolution, that is lost upon producing a hard copy anyway?

 

Change the record: digital = very wide, short focal length - therefore digital looks like sharp, harsh shit, while our Leicas (and old film in general) make nice, three-dimentional, beautiful selective focus with great boke (do not put an "h" at the end, thank you).

 

You see, one has to attend to one's real needs and select tactics accordingly. Prove your superiority with right means.

I certainly agree that Leica has very, very fine glass, but what Leicaphiles often forget is that the weak link is the film, unless shoot exclusively TechPan from 20pound tripods, not to mention all the quality reducing steps (traditional or digital) on the way from a piece of film to a viewable image as pointed out before.

 

In that sense, a Leica glass - 100MegaPixel digital back combo could be a MF killer and I am sure that this is gonna happen in a couple of years from now.. The fact that now the same chip that is used in most MF digital backs is now available in a 35mm body illustrates what I mean.

When shooting with a handlend camera, neither film, nor optics resolution is the limiting factor. Even second-best lenses - as far as resolution is concerned, not other characteristics - are going to produce UNDISTINGUISHABLE images (from the best ones)

 

The limiting factor is the camera shake, and you will go to the lowly 30-40 lines per millimiter, very,very far from the (hardly) achievable 100 lines per mm.

 

Get off the beaten track.Remember the old IBM logo - THINK!

Provia is my main slide film at the moment, I know how it scans because I do it personally, and I didn't find anything radically wrong with Reichmann's scanning technique. I'm not thrilled with auto-leveling either and prefer to manually adjust my B-W points to stay within the dynamics of that individual piece of film, but it's a trivial complaint.<P> Next time, he should just stick to drum scans in such a comparison just to shut up the peanut gallery. Be damned if the results defeat expectations...must be the procedure then.

Regarding Michael Bender's correction of the common usage word "bokeh" to the correct Japanese Romanji spelling of "boke", my Japanese dictionary defines "boke" as follows:

 

boke noun:idiot; fool; touched in the head from; out of it from; space case; Alzheimer's (impol);

 

 

Hai, domo arigato gozaimashita.

Rick:

from a 1997 posting by a real Japanese (to an Olympus forum):

<p>

<pre>

"I know it is a little off-topic from Olympus, but I could not resist;

 

In photographic or any imaging medium such as either in digital or

analog format, or in even graphics arts, the term "boke" is widely

used. However, "bokeh"(noun) or "bokeru" (present verb) also

describe person's mental palsy state such as see

in Alzheimer's Disease or in brain disorder (Hydrocephalus).

 

To be accurate in the translation the terms "boke", "bokeru",

"boketa", "boke-me" should have preceeded by "pinto" or "pinto ga",

which means focus. I am not sure how "Brandpunt" became "pinto" in

Japanese. Incidentally, "pinto" is not same as pinto in English and

it is pronounced differently - even stresses in 'i' and 'o'.

 

Regards,

Jim Terazawa

</pre>

<p>

His spelling reflects the fact there is no "h" sound at the end of the word - actually, it was introduced in English spelling to indicate to the unenlightened that it is not "bou-kei" (as in French bouquet, which in turn must be "buke", no diftongs there at all). And it follows a long, long tradition of those misspellings to forestall misprononciation - and the hilarious consequesnces of morons misreading the simplified moronic in even more moronic way:

"van gou" instead of "van hoh", hard "h" twice, is another example.

<p>

However what you dug up in your dictionary is not completely inappropriate. It can and should be turned in a motto of this leica forum.

(actually. to be precise, Van Gogh is not a misspelling in another language, it's just a misreading of the original spelling, as far as I know)

Scott,

 

As i pointed out auto-leveling on slidefilm isnt very desireable, it seems you at least agree on that. Doing it twice isnt really going to help 'a normal' comparison because noone would treat slides that way. And this is what comparing 2 DIFFERENT media is about IMO, both potentials should be shown at their best, not seemingly processed in the same way. Well you can call me everything you like but i think this approach to comparing media is 'pulling' results. This was the question in the original post which i tried to comment. I dont expect any result which would interest me, as far as i am concerned anyone should use what he/she likes best.

 

Greetings,

I work in the visual effects business and deal with digital images on a daily

basis. There is no doubt in my mind that 10 years from now you will be able to

buy a digital camera for a reasonable amount of money which will produce

stunning images that technically match or exceed anything we can do in

35mm today. But I will continue to shoot film as long as it is available. Why?

Because I like the mechanical experience of using my Leicas. I enjoy the

different finger prints that each of my lenses imprints on the final image. I'm

fascinated by the black art of how much shaking my tank during developing is

going to affect the contrast and grain in the final image. To me photography is

an art and a craft. In my business we have a saying: "You can train almost

anyone to run a piece of software, but you can't teach them to be an artist."

Almost anyone can take a picture, dump it in Photoshop, fix their mistakes,

print it out and get reasonable or even good results. Anyone can shoot 10

versions and check them on the little LCD screen until by process of

elimination they get what they want, without ever really having to know what

they are doing. But how many people do you know who can shoot like lets

say Ansel Adams? A shot that is perfectly framed, exposed and printed,

without any kind of "crutch" or second chances, using rather nebulous

techniques that can't be controlled at the touch of a button? To me THAT is the

challenge, the fun and the reason why I shoot. In years of shooting I have

taken less that 10 shots that I consider perfect, real keepers, but there are very

few things that are equally satisfying. It's not about wether film is better than

digital. To me it's about the art, craft and the difference between something

that is hand made or "off the rack".

 

 

 

Cheers,

 

feli

I think Petr has hit the nail on the head. Comparing R8 APO-Telyt vs another high quality lens from Canon or Contax or Nikon is a moot point for most purposes. Can film resolve more than the CCD on a D30 or D60... yes. Can a current $2k scanner pick this up.. no. Will you ever be able to bring this resolution to paper.. unlikely. I love film. But the bottom line seems to be that if all you do are 8x10 prints or even 11x14, a D60 will be better and more convinient.

Ravi,

 

I think you forget film is made to be printed in a darkroom NOT on a printer. They are 2 different media....

 

Greetings,