Micro 4/3 When Will Micro Four Thirds Equal Medium-Format Film?

Random thought - while the look of an image is influenced by many things - even film grain, my main concern is the lens perspective. So if I am able to crop at will, and I have a 35 mm lens and I'm trying to duplicate the perspective of a 50 mm lens just by cropping, it wouldn't be the same, unless .... (unknown)

Yes. A couple of years ago I was on a country road in Ontario with a brooding sky and nice red barn. I had my friend Gary make a couple of passes on his Honda. I took one photo wide (24mm EQ)and the other longer (~65mm EQ):

25247973239_818fa02d69_b.jpg

Two passes, two focal lengths
by John Flores, on Flickr

When you zoom in on the wide shot it looks a lot like the longer shot:

25615538015_7ef96286ef_b.jpg

Zoom in on the wide shot and it starts to look like the other.
by John Flores, on Flickr

The shots differ a little because I was crouching more for one than the other. But otherwise the relationship between bike , barn, trees, and road are identical because I didn't move; my relationship to them is unchanged.
 
I guess in theory you could do street photography using a bunch of video cameras working on automatic and then sift through the results, given arbitrarily large computational resources. Of course, that eliminates the part of photography I enjoy most. For most of us, the goal is not simply to get a great image. It's participating in the process, waiting and looking for that decisive moment.

Yes and, as many of us know, the best photographs are often not technically perfect (sometimes not even close to it) but ones that invoke a response on the part of the viewer.
 
much in the way that one could ask a computer to spit out infinite permutations of musical notes. If it were truly infinite, the computer would eventually spit out Beethoven's 5th Symphony or The Rolling Stones' "Satisfaction". It really is all about the creative process.
 
much in the way that one could ask a computer to spit out infinite permutations of musical notes. If it were truly infinite, the computer would eventually spit out Beethoven's 5th Symphony or The Rolling Stones' "Satisfaction". It really is all about the creative process.

This sound suspiciously like the infinite number of monkeys + infinite number of typewriters = all the works of Shakespeare argument. Bob Newhart had some insightful things to say about this.

Cheers, Jock
 
Why wouldn't it be the same? Apart from resolution that is. So long as both focal lengths are free from barrel or pincushion distortion, there shouldn't be a difference in perspective, assuming the subject distance is the same.

If you change the subject distance in order to have the same subject size in the frame, then yes, perspective changes. But with the same subject distance, the perspective is given, and the only thing that a wide or tele lens does is show you more or less of that given perspective.

Having said that, at wider than 50mm (equivalent or absolute? Im not sure on this one, but my gut says equivalent), lenses start to render the edges slightly stretched. But even then, cropping to only keep the center should give the same composition, including perspective, as using a lens with a longer focal length.

And in terms of Ken Rockwells 175mp figure: his logic is flawed in several ways, but i'm not sure if it's worth describing why. If anyone wants to know, I think I already typed it out somewhere on this site, years ago.

Yeah - good point - thanks.
 
This sound suspiciously like the infinite number of monkeys + infinite number of typewriters = all the works of Shakespeare argument. Bob Newhart had some insightful things to say about this.

Cheers, Jock

Well, that's my secret - shoot 1000, keep 5.
 
Why wouldn't it be the same? Apart from resolution that is. So long as both focal lengths are free from barrel or pincushion distortion, there shouldn't be a difference in perspective, assuming the subject distance is the same.

If you change the subject distance in order to have the same subject size in the frame, then yes, perspective changes. But with the same subject distance, the perspective is given, and the only thing that a wide or tele lens does is show you more or less of that given perspective.

Having said that, at wider than 50mm (equivalent or absolute? Im not sure on this one, but my gut says equivalent), lenses start to render the edges slightly stretched. But even then, cropping to only keep the center should give the same composition, including perspective, as using a lens with a longer focal length.

And in terms of Ken Rockwells 175mp figure: his logic is flawed in several ways, but i'm not sure if it's worth describing why. If anyone wants to know, I think I already typed it out somewhere on this site, years ago.

Clarification: I agree on the perspective, but not on Ken Rockwell. Ken has the best measurement credentials of anyone I know, and many sites see posts that denigrate him because he reveals some unpleasant truths. His "Your camera doesn't matter" article really got a lot of folks upset - especially those who live on camera ads. But the point was golden. Digital is good - I use it exclusively, but I can understand how it doesn't capture *everything* that film does. My Leica D-Lux 4/3 is way, way below the quality of a good 35 mm color slide.
 
I looked up my old post about this, which turned out to be a reply to another post by you, Dale :p so here goes what I said:
-I doubt any lens for 135 film resolves 175mp worth of detail. With the introduction of the Nikon D800 and its 36mp sensor, it turned out that only a handful of lenses is sharp enough to consistently outresolve the sensor. The reason no-one ever noticed the unsharpness in loads of lenses is that film never got above 36mp either. So you can imagine what the chances are of any lens resolving 175mp. Besides, at 175mp, any lens no matter how good will be diffraction limited by f/5.6.
-He uses Fuji Velvia 50, the sharpest of all films - most films don't come close to this.
-He bases his calculation on the lines/mm for a 1000:1 contrast, wayyy stronger than what you'll encounter in any setting other than astrophotography. For normal situations, a contrast of 1.6:1 is used, giving Fuji Velvia 50 a resolution of 80 lines/mm and making the resolution "only" 22mp - or neatly at the limit of what people report for real-world prints from film.
-the "lie factor" he's talking about, due to Bayer interpolation: I don't have any hard data to back this up, but I don't think it's as strong as he states. A good sensor/lens/software combination can give critically sharp images at 100%. Also, at the 1000:1 black/white contrast he used for his calculations, I'm pretty sure it doesn't really matter what colour filter is added to an individual pixel, because the contrast will be visible anyway. At more reasonable contrast rates, sure it might make a difference, but as I showed above, even Velvia 50 is only at 22mp if you use reasonable contrast.

as for KR, I didn't read his article about why the camera doesn't matter; I found his article about wide angle lenses very useful. However, I also saw several things that were plainly wrong on his website. His About section answered why; basically, it's part being human and not having a fact checker, and it's partly his "sense of humor", as he enjoys 'making things up for fun'. He says: 'I have the energy and sense of humor of a three-year old, so remember, this is a personal website, and never presented as fact.', and as such I have since treated his website. Useful at times, but not something to automatically be relied on. Which led me to fact-check his 175mp claim, leading me to my comments above.
 
I looked up my old post about this, which turned out to be a reply to another post by you, Dale :p so here goes what I said:
-I doubt any lens for 135 film resolves 175mp worth of detail. With the introduction of the Nikon D800 and its 36mp sensor, it turned out that only a handful of lenses is sharp enough to consistently outresolve the sensor. The reason no-one ever noticed the unsharpness in loads of lenses is that film never got above 36mp either. So you can imagine what the chances are of any lens resolving 175mp. Besides, at 175mp, any lens no matter how good will be diffraction limited by f/5.6.
-He uses Fuji Velvia 50, the sharpest of all films - most films don't come close to this.
-He bases his calculation on the lines/mm for a 1000:1 contrast, wayyy stronger than what you'll encounter in any setting other than astrophotography. For normal situations, a contrast of 1.6:1 is used, giving Fuji Velvia 50 a resolution of 80 lines/mm and making the resolution "only" 22mp - or neatly at the limit of what people report for real-world prints from film.
-the "lie factor" he's talking about, due to Bayer interpolation: I don't have any hard data to back this up, but I don't think it's as strong as he states. A good sensor/lens/software combination can give critically sharp images at 100%. Also, at the 1000:1 black/white contrast he used for his calculations, I'm pretty sure it doesn't really matter what colour filter is added to an individual pixel, because the contrast will be visible anyway. At more reasonable contrast rates, sure it might make a difference, but as I showed above, even Velvia 50 is only at 22mp if you use reasonable contrast.

as for KR, I didn't read his article about why the camera doesn't matter; I found his article about wide angle lenses very useful. However, I also saw several things that were plainly wrong on his website. His About section answered why; basically, it's part being human and not having a fact checker, and it's partly his "sense of humor", as he enjoys 'making things up for fun'. He says: 'I have the energy and sense of humor of a three-year old, so remember, this is a personal website, and never presented as fact.', and as such I have since treated his website. Useful at times, but not something to automatically be relied on. Which led me to fact-check his 175mp claim, leading me to my comments above.

I'm sure you understand disclaimers and such, and obviously you know the technical aspects of resolution, at least on a 2D surface of digital images. I was part of an NDT team starting in 1984, to put a film project on computer, preceding those wonderful tomography machines we have now. Film is amazing, especially when you have them collated just so.

Anyway, disclaimers aside, Rockwell's skills are phenomenal, and I highly recommend comparing his analyses of actual measurements compared to other folks who do similar work. A piece of film is really nothing like a digital image, but you'd have to work with it to understand. The problem with all of the discussion of lines per mm etc. is assuming that everything in the hypothetical 175 mp flows through that 2D plane, as if the famous double-slit experiment behaved like classic physics. Rockwell is waaaay up there, and his disclaimers and humor undoubtedly provide him a degree of safety.
 
I can’t understand pixel count comparisons with different film formats. I don’t see much discussion about the pixel density, rather than total number of pixels on a sensor. In the darkroom, I was able to get a grainier image by altering chemicals and processing times and then manipulating enlargement/cropping. The chemical density of silver-based emulsions was the same across film sizes (for the same brand and formulation). Increased tonal gradation, sharpness, etc, was better for larger format film because a print of a particular size required less enlarging with a larger film size. And, of course, there were all of the factors brought to consideration about the chemical formulation of the emulsion on the paper. Hundreds of possible factor variations required an artistic determination, rather than a formulaic decision.

We can say that these artistic decisions involve ‘computations’ that are not accessible to conscious awareness. I see a parallel in the current contest between AlphaGo and the best Go players. The humans make game play decisions based on analysis, intuition, feelings, and other psychological (unconscious) factors. AlphaGo simply computes — and wins. A game though, is not art. In the former there is a winner and a loser, the latter there is no winner (except, perhaps through consensus).
 
Back
Top