26 Megapixel 40D Compared to 35mm Film

[rashedahmed]

When you compare a 35mm film camera,compare it with a FF sensor,at least Canon 5D.

[mauro_franic]

Start from the top Rashed.

[jpo3136b]

I read this whole thing; I understand the "26 megapixel" bit. I noticed some logical flaws here. I mention them in hopes

that they might support some future solutions.

 

1. The DSLR sensor is 61% of the size of the 35mm film. This is a big point of contention. The characteristics of the

DSLR were named so that they were "upgraded"; that is, the characteristics were projected into What Would Be if the

sensor was full frame. Meanwhile, it is 61% of the size of the 35mm film. This was confusing in the labeling of what

was going on. Somehow, it lent an inflationary tone to the claims.

 

2. Same lens for two different formats? For example, I use 645 lenses on my 35mm film and 28mm DSLR

(23.5X15.7mm CCD). These frame equivalencies are not the same. When you switch formats, either focal length will

hold, or angle of view will hold.

 

To hold angle of view (important here, because it implies images in the test are crops of either 35mm film or

enlargements of the 28mm digital): 50mm lens on a 35mm camera at a factor of 1.5: 50/1.5=33mm equivalent focal

length on the 28mm DSLR.

 

So, if those comparison images were made under the same optical conditions relative to the capture surfaces involved,

the same lens should not be used on both cameras. One would need a 50mm on the 35mm film camera and a ~30mm

lens on the 28mm DSLR to hold angle of view, and compare the test images without reducing or enlarging one or the

other in the side by side pictures.

 

Were those side by side images originally of the same size? I don't think they would be. They are the foundation for our

understanding of "which is sharper"-type questions. They are presented in the web page illustration as being the same size.

 

I know, from using 645 lenses on 645, 35mm and 28mm that while focal length is focal length, the effect that same lens

will have on different format cameras is not the same. There are distinct changes in what I see in the viewfinder, and

distinct changes in what I see on the films and digital sensor. Moving a lens from one format to another is not the same.

It is not equivalent. It is observably different.

 

3. DSLR sensors are probably built with a structure that's a staggered matrix. Film grain distribution will be governed by

randomness. This implies that the two different kinds of structures will respond differently to different kinds of lines and

surfaces. Orientation and degree of curvature of the recorded surface will probably be illustrated by the two different

kinds of matrices in two different ways, when examined closely.

 

4. 35mm film for a clarity test? Now, if I submit anything less than 645/120mm film for a magazine cover, will they just

throw it in the trash, or throw it in the trash and send me a rejection letter?

 

5. That little 24mm sensor is pushing a level of clarity that is mighty close to matching a recording system that is

almost 40% bigger, if size matters.

 

6. Apples and oranges. I'll fire up the proverbial flamethrower with the best of 'em, but I think the whole film/digital

decision is a local judgement call.

 

I love film, and use it a lot more than this new DSLR. But this equivalency test, like every other equivalency argument

I've seen, simply does not match up. I'm just gonna crawl back over to my baths and print some more.

[jpo3136b]

The vignette comparisons don't work because of the three dimensional curve of the lens glass.

 

And before someone says it, I'll answer. Even though the 35mm image would be cropped down to an equivalent size for

the image comparisons, it's still not equal.

 

A lens is a curved, three dimensional piece of glass. Using a section of glass is not the same as considering the effect

of the curve over the whole glass. Does a section of an arc have the same effect as a full arc? No. If it did, Non-

Euclidean geometry would be invalid. The full arc, the effect created within the lens barrel has bearing on these tests.

 

And, that video clip about the Nikon digital camera making a billboard sized ad alongside a 35mm film image doesn't

work as an equivalency for the converse reason; in the pro-digital arguments, the effect of optics is always minimized

during an equivalency claim.

 

I'm no scientist, but I don't think this stuff, either pro-film or pro-digital, jibes with Newtonian Optics or basic mechanics.

I could be wrong, but I consistently see equivalency claimed in tests between structures that are inherently different.

 

I'm glad to see the OP tried to put some work into this, though. Kudos there. J.

[StuartMoxham]

Nice to see you redid the comparison with the 40d regulary framed. I would also like to see some color comparisons but with more usual subjects. A portrait and a landscape would be a good example. I would also like to see some other films not just the usual fine grained velvia or tmax 100. Portra 400 against the 40d at ISO400 would be most interesting for me. Maybe some Fuji Press 800 against the 40d at ISO800. These are the sort of films that I used in the past.

[nickc1]

The points about optics are interesting - as film will capture an image where the angle subtended by the rays of light to it are far greater than is possible today using a digital sensor, and, if you are using the same lens the digital will only use the 'sweet spot' in the centre of the lens, it is clear that optically the two situations are probably going to be very different - one of the reasons why I think this is an 'apples and oranges' question like the poster above suggested.

 

If I just wanted to capture an image I would not expect either the process or the results from sketching with charcoal to look the same as those from an oil painting - so why should I expect digital capture be the same as using silver halide?

[peterq]

"

John O'Keefe-Odom [subscriber] [Frequent poster] , Nov 16, 2008; 02:51 a.m.

Respectfully, the optics and mechanics don't jibe in this review.

 

 

3. DSLR sensors are probably built with a structure that's a staggered matrix. Film grain distribution will be governed by randomness. This implies that the two different kinds of structures will respond differently to different kinds of lines and surfaces. Orientation and degree of curvature of the recorded surface will probably be illustrated by the two different kinds of matrices in two different ways, when examined closely.

 

"

 

Thank you John,

this is where it becomes apples to oranges in the comparison of the poster.

 

What is the comparison about ?

Grain / smallest detail that can be captured? Obviously not "quality" (whatever that is).

 

The poster doesn't put a question, but only some observations without a clear goal.

 

Conclusions are ours?

My conclusions are simple:

Film is fundamentally different from digital, so comparing them is only comparing them like apples and oranges,

even grain size compared to pixel size doesn't say that much;

A scanner is a bad tool to prove the details and quality of film, because it transforms one domain to the other;

If one wants to shoot film or digital, please do.

[StuartMoxham]

Personaly I don't mind film to digital comparisons. While they may be fundamentally different we usually use the DSLR the same way we used a 35mm film SLR. Many people replace their 35mm SLRs with DX sensor DSLRs so it is only natural to compare the two.

[pankaj purohit]

I am still happy with my 35mm films SLR and still don't want to switch over DSLR or MF because it still full feeling my needs. It depends what and how you need, I don't always be in hurry so I don;t need to switch to DSLRs and don;t want huge details like MF so I don't go for that because if I switch to any of these mediums then I will definitely have to compromise some areas and may be those areas may not suite me, if I switch to DSLR then I will have to invest in large amount and I don't feel that I will get that much of benefit in ratio of investment. And if I upgrade to MF than I will have to carry a huge and heavy equipment setup with I don;t want that.

 

I strongly feel that the equipment choice vary by person to person and need by need.

[matt_needham]

Yay!!! 35mm film beats imaginary digital camera when photos are viewed through a microscope. Wait a minute...

[kelly_flanigan1]

Whats cool about this is the 40D is now 26 megapixels. Thus my 35 megapixel 7x10cm Phase One scan back is now 64 mp "if full frame" and my 50 megapixel scan back is now 91 "if full frame". :) And my Blackberry's 2 megapixel camera would be about 2 to 4 Giga pixel "if full frame".

[mauro_franic]

All, this should further CLARIFY THE SETUP, for the ones struggling to understand it:

 

http://shutterclick.smugmug.com/gallery/6560295_hraSq#418129352_UCwSt-O-LB

[mauro_franic]

I'd rather error on assuming people mean well and need help understanding, than assuming they are ill intended and already understand. Plus the former is the only one I can help with...

[mauro_franic]

That is my intention Les.

 

I will post the other films' results as soon as I receive them.

[stan_belyaev]

Thanks Mauro,

Nice comparison.

Almost everyone knows that low speed 35mm films easily out-resolve/outperforme small digital sensors, although not everyone admits due

to ignorance, stubbornness or commercial reasons.

Canon admits it in its brochures.

Kodak and fujifilm say the same thing.

Here is an interview with the disigners of Superia Reala/Provia.

It is in russian but on can use easily translate it using one of the web programs.

 

http://www.fujifilm.ru/fujifilm_world/technology/reala/interview.htm

 

I think that the results of your test are valid. The simple explanation of this results is that almost all digital cameras have anti-aliasing filter

with cut-off frequency of about 40-50lp/mm. Therefore any good 35mm film should outperform digital cameras by definition. Eventually

bigger size digital sensors win due to very high number of pixels but one has to expect to pay a price, which is much more that the price of

typical DSLR.

 

What is the point to argue with someone who thinks that a graphic editor generated lifeless images have anything to do with real

photography. Let them pay thousands for upgrades every year. The countries are in crisis (including Japan) and the industry needs our

support.

[kelly_flanigan1]

Why not call the test a test of a cropped 10 megapixel dlsr versus a piece of 35mm film or cropped piece of film that has the same area as the 10 megapixel sensor? *Then* one is not making up some 26 megapixel 40D; its really what the test is about a 10 megapixel 40D; ie reality.<BR><BR>There are many applications were one cannot get closer to the subject; ie one is shooting sports, birds; the moon, race cars and you cannot magically get closer in an instant. <BR><BR>Thus if one has a 50, 200 or 500mm lens the object is same size one the sensor or film. It might be tiny; the football player might be catching a long bomb pass and you are using a 500mm lens. Thus many folks really are using fraction of the cropped dslr's image or 35mm film's area for a newspaper or magazine. What matters in these practical applications is the results. Unless its the twlight zone one cannot halt the pass and shoot several frames to make a 26 megapixel image; one cannot walk on the the field and get closer; or change lenses; or even zoom. As a practical matter one really is comparing two digital files; one is scanned with a scanner; not even a high end drum scan. Its like commercial; sports and newspapers have mostly gone digital and landscape photographers still can enjoy film; since mountains and trees dont move much. :)

[mauro_franic]

Thank you Stan.

 

Kelly, I agree you should choose the tool that matches your needs best.

[daniele_chiesa]

How much easier this all would have been if Mauro used a 5D... Mauro, don't you have one?

Too much room for denial with this setup...

BTW, Mauro, when you get the others films scanned, it might be a good idea to start a new thread for those.

[mauro_franic]

Sadly I don't have a 5D. It would have been a nice add-on to the test. Obviously it would have resolved a lot less detail than the 40D or 35mm film IN THIS SETUP.

 

I will follow your advice and start a new thread once I receive the other films.

[mauro_franic]

I thought of sharing this example I shot with TMX last weekend to add some humor and have grouchy people loosen up.

 

Entitled: "How can something that is dead look so good: Film."

 

http://shutterclick.smugmug.com/gallery/6523007_CpbPd#414593695_QxkHH-X3-LB

[Alan Johnson]

Evidence I have supports the contention that T-max 100 film outresolves any current digital sensor.

 

But if the question is asked 'how much enlargement of film is allowable to achieve a grain-free result?',Roger Hicks and Frances Schultz have an answer in their book Quality in Photography p166: "With 35mm we find we can achieve a result which is reminiscent of a contact print from 5x7 by using Ilford Delta 100 or XP-2 and enlarging just 5x...Frances can go to 11x14 because she favours the 6x9 format."

 

So the comparison for resolution and for grain gives different answers.

[StuartMoxham]

Yep small prints from B&W film can look real nice. I made quite alot of 6 to 7 inch square prints from 6x6cm med format for a family album and made a lot of 3 1/2 x 5inch prints from 35mm. The tonality in those prints was real nice.

[mauro_franic]

Last week I printed the pictures I took with Ektar 100 35mm which is less fine grained than TMX. I printed them all at 11x14 and one at 16x20 on Epson Velvet Fine Art paper. None have visible grain on the print and tonality and detail like you mention in mind -blowing.

 

You can download and print yourself to experiment (give me 1 minute to remove the download protection):

 

http://shutterclick.smugmug.com/gallery/6499685_dJwsh#412832226_BhGTs

 

 

Also none of my prints from Velvia or TMX (6x7 film) show visible grain. I print at 16x20, 17x25, 20x30 and 24x36.

[Alan Johnson]

I dont have a resolution test result for color print film so cannot add anything on that.

 

Regarding grain, Hicks & Schultz say "....in color we find the half tone effect less obvious....this means that enlargements as big as 10x are entirely feasible from the best current (this was published in 2000) color print films,typically ISO 100 or less..."

[mauro_franic]

Not sure I understand what "entirely feasible from best current" means, nor I can assert or deny other peoples results.

 

From my own tests, the 11x14 prints from 35mm Ektar didn't show grain on Velvet Fine Art paper, that is higher than 10x. (You can download them and print them too if you have the same paper).