It's the MTF advantage, stupid

[ondebanks]

Edouard, I found your title for this thread not offensive but comical because it is flawed.

Shashin was correct to point out the obvious directly. MTF of smaller format lenses (the good ones) are much higher than the ones used for the medium format lenses where the requirements are a lot lower.

Give it a rest.

I do get what Edouard is trying to discuss. But you're right, his thread title is misleading.

A better thread title might have been "It's the lp/ph advantage, stupid". What larger format systems can, and should, deliver is greater total resolved detail, which is often encapsulated in a single figure as line pairs per picture height; lp/ph.

Ray

 

[edouard]

I do get what Edouard is trying to discuss. But you're right, his thread title is misleading.

A better thread title might have been "It's the lp/ph advantage, stupid". What larger format systems can, and should, deliver is greater total resolved detail, which is often encapsulated in a single figure as line pairs per picture height; lp/ph.

Ray

yes, or maybe just:
"Size does matter!"

;-)

 

[edouard]

I don't know ... I'm feeling quite not very comfortable as I have the feeling I was the one forcing edouard into this...

However, I just posted two pictures in the "Artful Pictures..." section. Both of them are not exactly safe for work - nothing too extreme of course - but you may want to take a look and guess which one is done with either a Hasselblad or a Nikon System.

Please guys: I don't want to start any discussion about the need of MFD nor the replacement of a H with a N. I just wanted to share MY decision and why I did so.

I fully understand, that there are a million reasons why a MFD is still better in some cases than a DSLR (and vice versa) as like for Fashion <-> Sports.

S.

Yes, interesting subject: I think the higher micro-contrast & optical resolution of larger format is mostly impressive on pictures with a lot of details the brain is used to process, so more visible in group photography and "wide scenes" than on macro for example.
For boudoir photography, let's say that the brain got focused on the few parts that are detailed enough even with less micro-contrast ;-) ;-) + with post-processing it becomes difficult to tell (although it won't be high fidelity anymore if too processed). I'm sure I can't tell the difference (+ on small sized images)!

This made me also think: One could argue that a 2x surface increase like 35mm vs current MFD backs is not large enough... unlike 35mm vs 6x6 (3.8x increase = ~50% reduction in spatial frequencies). To me, film full 6x6 still looks more impressive than 645 MFD, but I hate going to the dev shop, waiting and scanning.

You shouldn't switch to small format but to 8x10 instead ;-) ;-)

 

[edouard]

You shouldn't switch to small format but to 8x10 instead ;-) ;-)

or even better: life-size direct-positive prints ;-)
Uncomfortably Close: Richard Learoyd’s, Presences - LightBox

('have seen few of his giant photos at Paris Photo 2011, 'never seen something that impressive / disturbing!)

 

[fotografz]

I don't know ... I'm feeling quite not very comfortable as I have the feeling I was the one forcing edouard into this...

However, I just posted two pictures in the "Artful Pictures..." section. Both of them are not exactly safe for work - nothing too extreme of course - but you may want to take a look and guess which one is done with either a Hasselblad or a Nikon System.

Please guys: I don't want to start any discussion about the need of MFD nor the replacement of a H with a N. I just wanted to share MY decision and why I did so.

I fully understand, that there are a million reasons why a MFD is still better in some cases than a DSLR (and vice versa) as like for Fashion <-> Sports.

S.

If you are only posting sub 1 meg, 8 bit jpegs in sRGB color space on the internet, then you do not need a FF 35mm DSLR either Stan. A $800 camera and decent lens, coupled with advanced lighting know-how (ambient or strobe) can fool many eyes ... (although I suspect Guy nailed which is which, and why, on your comparison thread). Next time also shoot it with a Canon Rebel and a 85/1.8 ... it could save you a lot of money

Again, I am forwarding the notion that it is impossible to sort out each individual need or desire regarding IQ because the components are a huge jumble of different requirements ... thus making ANY definitive proclamation "for or against" anything merely a personal preference based on your own criteria.

IMO, that is different from settling for "good enough".

Where I live, the commercial studios that once thrived are all but extinct ... yet in the midst of this a few have made an amazing success of it, and continue to grow. Since I used one of them for advertising stuff in past I asked the photographer how the hell he pulled it off. He said that while others reacted to the market by diminishing their product (what I call profit by attrition), he ramped it up, took the risk, and formulated a business strategy of flawless craftsmanship coupled with understanding the client's business, trained his shooters how to achieve it quickly, and sold all that to clients who otherwise may have been strictly price buyers.

Result: he has built an amazing new studio, has 9 photo stations running 24/7 ... shoots 95% of the work with Hasselblad H cameras two of which are Multi-Shots. When he tests gear it is to the criteria that's winning over his clients, and more importantly keeping them. He has one 35mm DSLR for reference shots.

Personally, I stepped back from a much loved 60 meg set-up to my dual shutter capable S system as I further retire from the fray. While the sensor is smaller, there are zero concerns regarding the optics which are all new digital specific ASPH or APO designs, consistent from focal length to focal length, and each is spectacular. It fits my current needs to a T, while providing the versatility, shooting pleasure, and IQ I've grown accustomed to. :thumbup:

- Marc

 

[ondebanks]

But you MTF argument is false. A smaller format at a given aperture will actually have better MTF over the image simply because of greater DoF.

Well, MTF is a property of the lens; not of the input scene, and not of the output image. The lens' MTF is not a single function either. Although normally plotted for a system in perfect focus, it changes (worsens) as the PSF changes on either side of focus. So, bringing DOF into things isn't really fair; because the scenario you gave ("a smaller format at a given aperture") gives an advantage to the smaller format system thanks to its greater proportion of scene convolved with a (nearly) in-focus PSF - but it doesn't deliver the same picture due to the DOF differences which would be clear to the viewer.

To equalize the pictures, requires equalizing the DOFs, which means stopping down the larger format lens more. Then one gets into this familiar argument:
"equalizing the DOF -> stopping down more -> more diffraction -> blurrier PSF -> detail lost in direct proportion to the format size -> conclusion: all sufficiently well sampled formats deliver the same detail for a given picture; there is actually no point in using larger formats".

Where that conclusion falls down is that it assumes that all lenses are diffraction limited at all f-stops; that for smaller and smaller formats, there will exist faster and faster lenses which are still diffraction limited wide open.

In practice, we are much more likely to find that our larger format lenses - even cheaper old ones - are happily diffraction limited at their modest and small f-stops than the smaller format lenses are at their wider f-stops; thus delivering the same picture DOF but with higher levels of in-focus detail.

Unless you are simply photographing flat planes, then MTF is not a reason for buying anything.

Flat planes...or the inside "surface" of an infinite sphere (astrophotography! :chug...or if it's a non-flat-plane nearby subject but you are not looking for deep DOF, and just want lots of resolved detail where it is in focus.

Also, MTF is dependent on pixel pitch,

Well, sort of...if you choose to use a certain pixel pitch to normalize everything. But the standard units are lp/mm spatial cycles, which has the advantage of being independent of pixel size, or indeed of any kind of sampling. MTFs existed prior to digital imaging!

Ray

 

[wentbackward]

I don't know ... I'm feeling quite not very comfortable as I have the feeling I was the one forcing edouard into this...
...
S.

Stan, you've clearly balanced all factors and made your choice, that's why I don't think there's a MF/SF debate. With your talent, you could easily take those photo's with a D600 and they would still be stunning and probably not much in it between the D600 and D800, and the same argument could start D600 vs D800. It's a never ending argument and pointless.

One of the nicest images from my recent trips can only be printed at about 18" wide because it was taken with a waterproof P&S. So be it, I was in a river in waders. IQ was not even a consideration, walking away with something I could print, from the middle of a fast flowing River in Norway was the requirement.

 

[Tibor]

or even better: life-size direct-positive prints ;-)
Uncomfortably Close: Richard Learoyd’s, Presences - LightBox

('have seen few of his giant photos at Paris Photo 2011, 'never seen something that impressive / disturbing!)

Wow, I really like these. Amazing stuff.

 

[Frederic]

Regarding Marc comment and reaction to my post I agree there's no one standard in professional photography. However in such highly competitive markets my feeling is photographers are fully aware of their clients needs. Those who don't cease to exist pretty quickly.

Knowing which tools are good enough to produce one's vision and meet tech criteria really says nothing about a photographer high standards and general attitude toward work. Good enough equipment for a given task doesn't have to translate into good enough work. It's a business decision first and foremost, and maybe more of a German/northern EU one indeed.
Whether we like it or not on a metaphysical level is a different matter and a distinct debate.

 

[Shashin]

Well, MTF is a property of the lens; not of the input scene, and not of the output image. The lens' MTF is not a single function either. Although normally plotted for a system in perfect focus, it changes (worsens) as the PSF changes on either side of focus. So, bringing DOF into things isn't really fair; because the scenario you gave ("a smaller format at a given aperture") gives an advantage to the smaller format system thanks to its greater proportion of scene convolved with a (nearly) in-focus PSF - but it doesn't deliver the same picture due to the DOF differences which would be clear to the viewer.

To equalize the pictures, requires equalizing the DOFs, which means stopping down the larger format lens more. Then one gets into this familiar argument:
"equalizing the DOF -> stopping down more -> more diffraction -> blurrier PSF -> detail lost in direct proportion to the format size -> conclusion: all sufficiently well sampled formats deliver the same detail for a given picture; there is actually no point in using larger formats".

Where that conclusion falls down is that it assumes that all lenses are diffraction limited at all f-stops; that for smaller and smaller formats, there will exist faster and faster lenses which are still diffraction limited wide open.

In practice, we are much more likely to find that our larger format lenses - even cheaper old ones - are happily diffraction limited at their modest and small f-stops than the smaller format lenses are at their wider f-stops; thus delivering the same picture DOF but with higher levels of in-focus detail.



Flat planes...or the inside "surface" of an infinite sphere (astrophotography! :chug...or if it's a non-flat-plane nearby subject but you are not looking for deep DOF, and just want lots of resolved detail where it is in focus.



Well, sort of...if you choose to use a certain pixel pitch to normalize everything. But the standard units are lp/mm spatial cycles, which has the advantage of being independent of pixel size, or indeed of any kind of sampling. MTFs existed prior to digital imaging!

Ray

Ray, so systemic MTF is not real? I have a book that tells you how to calculate it. I find it really hard to benefit from the lens output without putting a sensor in the way. And you are going to tell me that in a 100% crop of my 6um RX-1 image at f/11 is going to be different from my 6um 645D with a 35mm at f/11? I would argue the MTF is identical.

The point I was trying to make to make is that there are too many factors in which to make absolute statements.

Actually, that wasn't the point I was trying to make. I think the OP is confusing GetDPI with DPreview. Opening a thread by calling people stupid is sad. Why can't photographers get along. Why are they so condescending toward each other--and I plead guilty as charged, but as a reformed sinner...(Hypocracy is my favorite form of government.)

And there is several problems with equivalency models. First, you need to bias the model. You can get a bunch of factor to equalize, but there will be at least one that is not. Most simply bias the system in favor of their point of view. The second problem with an equivalency model is that it assumes their one is the "correct" way of making a photograph. The old larger-format-are-for-shallow-DOF argument really never stopped large-format shooter stopping down for a large DOF--never seen shallow DOF in any of Adams work. I would argue DOF is not what gives large- and medium-format their "look," but their focal length, or rather their limited choice of FoV. Likewise, if you want the "look" of a particular format, shoot that format--never known a photographer thinking, "lets see, I am working in APS-C and I want this to look like it was taken in 35mm and so I need to convert this focal and aperture..."

Photography is subjective. It is an art form, at least once you stop shooting brick walls and test charts. The resulting image is the final word and the viewer perception of that. I think you are going to find it hard to find a direct relationship between that and MTF. And no system is going to get passed the limits of the human visual system--at what point does MTF become irrelevant?

The camera is a subjective choice of the photographer. It is not superficial. Photographers understand all the factors that camera choice gives. For the photographer, the relationship to the tools is important. The problem is each photographer comes up with a different solution. Just as there are different types of photographs. Now, I have a Phase One back and a 645D (and my D800 is sleeps in its box except for quick and dirty work). I am certainly not going to say forget MFD, on the contrary, it is a great format, but the MTF argument is too simple. In a dark bar, what MFD is going to do what I can do with an OMD with a 12mm f/2 lens? "But," I hear you say, "MFD is not a street camera!" But you would be wrong. I have done street with MFD. I am happy with the limitations. But I understand the benefits of other systems. I am thinking the OMD is going to have better MTF in the bar. Or maybe not, if I am lucky. (And I am talking about systemic MTF.)

And naturally, no one would use MFD for astrophotography.

I guess I am tired of the my-camera-is-bigger-than-yours arguments. In a way, this is no different than the D800-killed-MFD arguments that keep coming around. Personally, I would not worry whether my images are optimized for MTF, but are worth showing. And I think one of the reasons, and one of many, that images of MFD and other shooters of high-end equipment is "better" is because of the skill of the photographers to make compelling images, skills that were learned over time and a great deal of effort. There is nothing wrong with being technically competent, I would actually recommend it, but then what?

:rant off:

 

[johnnygoesdigital]

MTF has nothing to do with why I switched from MFD to the D800. That decision was based on reliability of the electronics and mechanics. This whiz match of who's more technically challenged feeds trolls, and quite frankly I think that's what the OP wants.

 

[fotografz]

All a matter of subjective opinion.

I switched from Nikon to Sony because I simply didn't care for the Nikon optical aesthetic look and feel ... and still don't. 24, 36 meg or even 50 meg is not relevant. The only D800 shots I've found remotely appealing where done with Leica R glass ... and I'm NOT going back to adapted manual focus lenses ever again.

Besides, IMO Nikon still hasn't made a camera that can equal the out of camera color of the A900 which can be had for 1/3 the price of a D800.

- Marc

 

[johnnygoesdigital]

Agree on the A900, nice camera, especially with Zeiss lenses, but I don't mind adjusting the color to my liking as I don't often use the OOC color anyway. Too bad they don't still make the A900. Although this is a MF thread the comparisons keep coming... Subjective indeed...

 

[fotografz]

However, I do not "subjectively" mix up my A900 with MFD ... even my S2 camera which is my main preferred camera for most color work.

It isn't a matter of "technical comparisons" which you seem obsessed with. IMO, it is a matter of creative preference. Various combinations of sensors and lenses that provide some aesthetic look and feel that one may favor over another. I still like what the fat pixel backs produced when used on either a Contax 645, or a Mamiya RZ ... just something special about the look. 10 year old tech which is far more pleasing to my eye than anything I've seen clinically produced by a D800.

 

[fotografz]

To elaborate, I also tweak what the camera captures. When I say I like the OOC color, it means it requires less PP work to achieve the aesthetic look I'm after. I've found that the more PP you have to do, the more danger there is of the image looking manipulated. Of course that depends on the skill of the person doing the manipulation, yet who wants to work every image so much to get what you want, when there are solutions that deliver a much closer file OOC?

For me, the lenses are main criteria. The sensor and output in concert with optics that produces an aesthetic that comes closest to what I am creatively after. Many people adapt lenses to a different camera to achieve that, which I also once did ... but as mentioned, I no longer am interested in that type exercise because it defeats the underlying purpose of a swift, mobile modern 35mm DSLR AF camera.

When Leica first introduced the S2, I tested it and found that it wasn't ready for prime time. Once they tweaked the camera/sensor to work with the all new lenses, it came into its own in a big way. Very natural look and feel that other people then took note of.

We all find our way ... and the only real comparisons are for ourselves and what we are creatively looking for from any given camera.

- Marc

 

[Guy Mancuso]

Just a side note. Color also has a lot to do with profiles and raw converters. One camera maybe awesome with a certain raw converter but look like crap with another. I've seen this more times than I can count. Testing everything when I got the D800 was pretty dismal when I tried all the raw converters until C1 came out with a profile for it. That's just one example. Phase products like Hassy products will always be better with there dedicated raw converters. IMHO Nikon and Canons own converters where never very good even with there own dedicated software. Color is a very flexible comparison among converters and also a very hard one to compare against each other. In most cases it will come down to user effort and time put into it. It's really hard to say one sensor or cam is better than another without finding the best converter for it. Having said that its also up to the OEM to build the proper internal profiles and algorithm of each body produced. There are a lot of variables here. Just look between what the same sensor in different brands produces. Sony and Nikon are a classic case of this using identical sensors in one or more of each others cams and look completely different from the profiles built into them.

 

[ondebanks]

Ray, so systemic MTF is not real? I have a book that tells you how to calculate it. I find it really hard to benefit from the lens output without putting a sensor in the way.

Ah, you're right of course. I was neglecting the contribution of the pixel sampling grid to the systemic MTF. But I do regard that contribution (approximately) as a sort of top-hat windowing function - no effect on the MTF at low frequencies and then a sudden collapse in the MTF at spatial frequencies above the sampling frequency, other than nasty aliases. This is where I feel the MFD systems have the advantage. Within that window, what they deliver is either of a higher modulation than with smaller formats (comparing e.g. 4000 pixels of 9 microns to 4000 pixels of 6 microns in image height), or it is of the same modulation but there's just considerably more of it (comparing e.g. 6000 pixels of 6 microns to 4000 pixels of 6 microns).

And you are going to tell me that in a 100% crop of my 6um RX-1 image at f/11 is going to be different from my 6um 645D with a 35mm at f/11? I would argue the MTF is identical.

No dispute with you there, Will - as long as the lenses are the equal of each other!

I guess my original point, which we have strayed from due to my inaccuracy re. the systemic MTF, was that if you zoom out of that 100% crop to the full image, you would be getting a different picture with the 645D (a wider true angle of view). To get the same picture as the RX-1 (angle, composition, DOF) you'd need a 45mm lens on the 645D, and an f-stop of f/14. But it's a better quality picture because the 45mm image is magnified more than the 35 mm image falling onto the same size pixels. And even if both images are fully diffraction limited in the plane of focus, the better sampling and additional light capturedin the MFD case allows for better results in deconvolution processing.

Ray

 

[Jan Brittenson]

There's no big, linear MTF advantage just because you increase size. Aberrations scale with everything else, and while some are easier to suppress as size increases, others are more difficult. And aberrations are what causes MTF to drop. In fact, aberrations are often expressed in percentage of image height when discussing optical designs, because this makes them size-neutral. A bigger size requires an increase in DoF that imposes a geometric loss of light. The only benefit is sensitivity, which is why as you progressively increase sensor site size you increase image quality. At least up to 24x36, because current MFD equipment doesn't yield any sensitivity advantage like one sees going from a phone to a compact P&S to APS-C, and to 24x36. Then it kind of stops.

The big difference is the target audience, not something inherently technical. MFD targets the studio photographer, where the key is hue neutrality. If a studio photographer wants a different skin hue they'll tell the makeup artist. If they want a different color background they swap out the backdrop. For a change in contrast, they adjust the lighting. They need a camera and lens which doesn't have any characteristic of its own to speak of. They need flat color response because they need to be able to mix and match hues and saturation levels; they need a fairly neutrally rendered face next to a highly saturated fabric. This is the market MFD caters to. There's no sensitivity benefit because the target market doesn't care - if you need more light you fix this by adding more light...

 

[dougpeterson]

Some aberrations, especially those dependent on mechanical tolerance/precision limitations do not scale with size. They scale/vary by process/machine.

And the idea that MFD is for studio shooters is definitely many years out of date. I don't know if it's the majority or just a large minority, but a huge chunk of our users do most of their shooting other-than-in-a-studio.

 

[edouard]