IQ180 Vs. 8X10 film article

[Fine_Art]

Sorry, not buying it. It's well known you have to scan at over 4000 dpi to get the most out of fine grained films like Velvia or Ektar. Why do a test now with discontinued films? Especially scanned at 745dpi? Nonsense.

Since the author is a biologist he needs to look at the film under his microscope to see the real detail. I have. I also use a 12x telescope eyepiece to inspect details in slides.

Roger Clark estimated 35mm slide film (velvia)  overall image quality at comparable to 16MP.

"digital megapixel equivalent (35 mm film) = 10 * (lpm1.6 / 80 lpm)2             (eqn 1)

digital megapixel equivalent (6x4.5 cm film) = 31 * (lpm1.6 / 80 lpm)2             (eqn 2)

digital megapixel equivalent (4x5 film) = 150 * (lpm1.6 / 80 lpm)2             (eqn 3)

digital megapixel equivalent (8x10 film) = 600 * (lpm1.6 / 80 lpm)2             (eqn 4) "

That puts 8x10 at about 800MP
http://clarkvision.com/articles/film.vs.digital.1/index.html
http://clarkvision.com/articles/scandetail/index.html#testarea3

Personally I scan at 4000dpi. I know people with the old Minolta 5400 scanner get better quality than me. I also find Noise Ninja (or any other noise software) incredibly good at removing film grain! For that reason I expect a 35mm slide to be about 20-22MP digital equivalent. What is lost in some image detail is gained in full RGB capture at every point.

Showing fuzzy 745dpi scans does nobody any good.

[Fine_Art]

Here is a simple test almost everyone can do with their DSLR. Get out your best macro lens to stitch together 4 or 6 shots as required for your best slide film. You may need to reverse a wide angle lens to pick up the grain. Make the grain cover a few pixels. Run your noise software. Sharpen. Downrez to where the image looks clear at 100%. How many MP do you have? That is over an area 840 sq. mm. An 8x10" is 204x254mm = 51816 sq mm.

For me based on 22MP for the slide, that is over 1300MP for an 8X10".

[theguywitha645d]

Actually, I have seen film under a microscope, not just the eyepiece, and there is no useful image. Sorry.

Thank you for taking the time to do the math, but you have one little problem--the images that were posted do not support your hypothesis. It is hard to argue with results. Could the film have gotten a higher result, yes at wider apertures. But in this test, and it is valid, the DoF was to be set at similar levels.

I think you might be a little confused by the results. These are at 100%. In no way is that a realistic viewing distance. Your perception of the actual image would be very different. While the different processes would add different qualities, both images will appear very detailed.

As a long time proponent of film photography, I really cannot agree with your inflated resolving powers for film nor your equations for calculating them. I don't know who Roger Clark is, but his site does say that the outcomes of this test are quite expected. He seems to think digital cameras resolve better than film. Which also seems to suggest that the math you got from him is also wrong. Perhaps it is time for him to update his website.

[dchew]

Don't forget that the constraint on image quality when shooting 8x10 film might not be the film, but the ability to focus accurately enough to resolve all of what the large sheet is capable of. The calculations above certainly don't take that into account.

This was a real-world format comparison, not a film resolution test.

Dave

[theguywitha645d]

Here is a simple test almost everyone can do with their DSLR. Get out your best macro lens to stitch together 4 or 6 shots as required for your best slide film. You may need to reverse a wide angle lens to pick up the grain. Make the grain cover a few pixels. Run your noise software. Sharpen. Downrez to where the image looks clear at 100%. How many MP do you have? That is over an area 840 sq. mm. An 8x10" is 204x254mm = 51816 sq mm.

For me based on 22MP for the slide, that is over 1300MP for an 8X10".

Actually, that is not a test for resolution. Grain are not pixels. You can image more grain than the grain can resolve. Yes, you can make 4000dpi scans of film, but that does not mean the film is resolving that. Nor can you simply extrapolate up to get 8x10 resolving power. Large format camera don't work at the same resolving power as their smaller format cousins.

[Mr. Rib]

There were numerous threads like this, some accompanying the release of a new back, some pop up randomly in time.. and they all end up with the same conclusion:

despite all the calculations and estimations of mp count of a 8x10, xxxxx digital back is resolving significantly worse than your 8x10, should the 8x10 be exposed properly, focused correctly and you were lucky enough to have the film flat enough, which is not obvious and always is a problem with this format.  

xxxxx stands for your digiback model name, whatever it is and will be for probably next few years.

[Fine_Art]

Actually, that is not a test for resolution. Grain are not pixels. You can image more grain than the grain can resolve. Yes, you can make 4000dpi scans of film, but that does not mean the film is resolving that. Nor can you simply extrapolate up to get 8x10 resolving power. Large format camera don't work at the same resolving power as their smaller format cousins.

So your lens doesnt resolve to the color film?

If you do a 3000dpi scan vs a 4000 dpi scan you can clearly get more detail from a good slide film. Countless people have done it.

By choosing 745 dpi you are saying the max resolution of the film is the equivalent of 1027x704 for a 35mm slide. Anyone who has seen a slide projected knows the detail is an order of magnitude greater that a small PC screen.

If you know you became diffraction limited in this test such that your system could only resolve 745 lines per inch, it is hardly a useful test.

[Fine_Art]

Hows about this:

at 750dpi you are 6000x7500. Divided by 1024 that is 43.9MP

You couldn't possibly have proven film can't match 80MP by scanning at 44MP!
The scan resolution was a mistake. Just rescan it.

[theguywitha645d]

So your lens doesnt resolve to the color film?

If you do a 3000dpi scan vs a 4000 dpi scan you can clearly get more detail from a good slide film. Countless people have done it.

I have done it too. The final resolving power would go up as the systemic resolving power is going up. But that is not saying the film is actually resolving that nor that a camera sensor with the same pixel density would be getting the same result, it would be doing better.

By choosing 745 dpi you are saying the max resolution of the film is the equivalent of 1027x704 for a 35mm slide. Anyone who has seen a slide projected knows the detail is an order of magnitude greater that a small PC screen.

And how can you determine resolving power of a system by simply looking at a projected image? None of this actually contradicts the results of the test. There is not much of a difference in pixel resolution, but the IQ180 image is clearly ahead. I see nothing in the scans to suggest the film has not already passed its limit.

If you know you became diffraction limited in this test such that your system could only resolve 745 lines per inch, it is hardly a useful test.

You are welcome to make your own tests and post them. However, you have said nothing nor posted anything to support your argument. I see nothing in the test that suggest that the film would produce a better result under those conditions--it is a valid statement. If what you are saying is true, then a direct comparison of a digital and film image will reveal that. I have scanned a lot of 35mm and medium format film with a 4000dpi Nikon film scanner. From what I have seen, there really is no doubt that on simply the technical issue of resolving power that digital sensors are doing better.

If you have evidence that gives a different result, I think everyone here would like to see it.

[theguywitha645d]

Hows about this:

at 750dpi you are 6000x7500. Divided by 1024 that is 43.9MP

You couldn't possibly have proven film can't match 80MP by scanning at 44MP!
The scan resolution was a mistake. Just rescan it.

Since I shoot with a 40MP camera, the film is soft even at that resolution. But it might be easier for you to show us examples.

[EricWHiss]

Forget the pixels and peeping - make a analog print from the 8x10 and one from the IQ180 and then ask normal people to pick which they like best.   

[ErikKaffehr]

Hi

If a good scan is soft than it will also be soft if resolution is doubled. I presume that authors finds that lens softness is clearly visible at the scan resolution used, and interpret it quite correctly that the scan is limited by the imaging system of camera.

It is quite possible that more performance could be wrangled out of the 8x10", the question is how? It seems from the experience Jack Flesher reported that focusing is highly critical and film flatness and positioning are real issues.

Best regards
Erik

Since I shoot with a 40MP camera, the film is soft even at that resolution. But it might be easier for you to show us examples.

[Fine_Art]

Since I shoot with a 40MP camera, the film is soft even at that resolution. But it might be easier for you to show us examples.

Fair enough. I don't shoot 8x10, I can provide slides or 645. That is a huge amount of data to upload so it would be easier to have people test their own systems. How's about a piece of Velvia backlit, with a tape measure on top? People can see the mm markings then extrapolate for the area.

Yes Eric, I understand that in this test there might be softness. The issue for me was the color film showed no sign of grain. I took that to be it is not near its max resolution. The B/W clearly did so i agree it's maxed out.

If you review the articles by R. Clark, who is an expert (bio at the site) as well as a very good photographer, you can see he has drum scanned 4x5 Velvia already. His comparisons show digital has superior contrast in the details. Digital also shows some fake detail. On the flip side film has the color accuracy of 4 colors at every point. That gives a realism to the images.

I think there is no question 24mp DSLRs from the big 3 are beyond the best 135 film in everything but color. I will completely agree with anything that scales that up based on area. I cannot agree about 8x10 vs 80MP given an 8x10 is basically a gigapixel camera. To beat that in digital you need a scanning back or massive stitching.

[macz5024]

Hi

The internet can show images as precise as some worn out newspaper paper - therefore I would like to add this link:

http://www.markuszuber.com/8by10.html

Markus

[yaya]

Hi

The internet can show images as precise as some worn out newspaper paper - therefore I would like to add this link:

http://www.markuszuber.com/8by10.html

Markus

Thank you Markus for the link. I hope this will help this thread come to an end:-) (it probably won't but thanks for trying!)

Yair

[design_freak]

Thank you Markus for the link. I hope this will help this thread come to an end:-) (it probably won't but thanks for trying!)

Yair

+1

I agree 

[gazwas]

I looked at the first test image and that instantly made my mind up which was better....

[Graham Mitchell]

I have read many comparisons between film and digital, and it always seems to end the same way, whether it's a 16MP DSLR v 35mm film or an IQ180 v 8x10 film: someone posts samples of scanned film versus digital and the digital is more detailed and noise-free. Usually the film buffs defend it by saying that the scanner was to blame or the image wasn't focused properly or the film wasn't flat, etc, but this seems to happen every time.

[issa]

I think there is no question 24mp DSLRs from the big 3 are beyond the best 135 film in everything but color. I will completely agree with anything that scales that up based on area. I cannot agree about 8x10 vs 80MP given an 8x10 is basically a gigapixel camera. To beat that in digital you need a scanning back or massive stitching.

It is not about 8x10 might transalte to a gigapixel, it about the quality of pixels in the first place.

I use 5x4 and have just received an IQ160, whilst I agree the colour of velevia/provia is great, the quailty of the file from the IQ160 is way better than 5x4. with the added advantage of  better detail in the shadow and highlights for PP, I can always add that velvia magic in PP if needed. It is not all about mega pixels, a canon 5D with 20MP will beat 35mm drum scnned even if it 100 mp.

[TH_Alpa]

True.

I also don't understand the real meaning of doing such a test. I simply remember the numerous (very) large prints I have seen from 8x10", up to 4 x 3 m, and I as well remember the resolution/sharpness I could see with my eyes in front of it. Being it flatness of the film, unprecise focus, not adjusted camera, stability of the camera or other vibrations, etc ..., this "softness" was always there, at close observation of these prints.
Something which I certainly did not see with digital sensors of the current generation and the best lenses used novadays with these digital sensors.

Thierry

... etc, but this seems to happen every time.