« Peter Turnley Prints | Main | Start With the Work and Work Backwards (The Stillborn Leica Challenge) »

Tuesday, 03 December 2024

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

I used to shoot 35mm, 6cm x 6cm, 4in by 5in film images. To my mind 1-inch sensor cameras are equivalent to 35mm film cameras, micro 4/3 and APS-C cameras are equivalent to 6cm by 4.5cm film cameras, and full-frame-sensor cameras are equivalent to 6cm by 9cm or even 4in by 5in film cameras. Whatever the format, digital cameras nowadays are very very good.

I didn’t buy a faster computer (Mac Studio M2 Ultra) to replace my old Mac Pro tower because I needed faster results (though faster processing is obvious), but in large part because the software I use (Mac O/S and other) keeps improving and ultimately will require hardware that can support all of it. These fancy Macs are no longer internally upgradable, and paying now will save in the long run. My Mac Pro lasted 14 years by continuously upgrading internals; and I hope the M2 will have similar longevity as spec-ed out of the box.

"If your work isn't good enough, it's unlikely the sensor is to blame.”

Mike, I do not consider myself a particularly good snapper, but I agree with this, though somehow in reverse.

I have a Foveon powered Sigma Quattro (apsc), and a Leica Q2(ff). I rarely use the former, and I love the latter.

However the camera, that I made my favourite snaps with, required hardly any setting up, it just required me to spot a great composition. As I said, not a great snapper, and that is completely down to me, so my Hasselblad SWC was pretty much wasted on me.

My point though is that the simplest camera was my best camera.

Unfortunately I was lured by Leica into buying the Q2, for all the reasons you state, but sort of the reverse, apart from the weight thing.

The Sigma is really fiddly, the Leica makes wonderful files, that print really well, but the SWC is point and shoot, and that’s it… job done, it really shows me up.

Are you sure that digital full frame corresponds to 120 film? I've seen a lot of prints done from 120 film and they are very far from what my cheap aps-c Pentax gives me when my cheap Tamron zoom nails focus (rarely). I have the impression that digital full frame has the resolution of sheet film, 5x7 or larger, but without the depth of field limitations. That is, much more than needed.

So where would you place APS-C? Lot of APS-C cameras out there - older DSLRs and newer mirrorless.

Looking back, I have a feeling that the 24MP APS-C sensors were perhaps the sweet spot for that size sensor. Any more (e.g Canon's 32 MP sensor as used in the 90D and now the R7) and you run into issues with the photosites being very small.

As someone who shoots m4/3, aps-c, FF 35mm (60MP) and 44x33mm medium format, I can say in practice I don't really find any image quality difference between them. I'm sure that if I were printing 50" wide prints, I'd find small differences, but in my favoured size of 12"x12", I find the image quality indistinguishable. This is still true when I print 150MP square stitches from my Fotodiox Rhinocam Vertex rotary stitch adaptor and Pentax 645 lenses. There is only so much real detail a printer can lay down on paper per inch, and that seems to be the limitation. At the 12" tall print size, all my cameras out-perform the printer/ink/paper.

I use MF camera primarily for long exposure work and it is excellent at this. The MF camera has the ability to shoot multi-minute exposures without needing LENR (which would double exposure time), it has shutter speeds to 1 hour and doesn't need a cable release or Bulb mode. A great convenience in an inconvenient genre.

My m4/3 has mostly been supplanted by my A7Riv which is an oversized m4/3, aps-c and FF camera all in the same body - a body which is smaller than my m4/3 camera. I particularly enjoy using it in FF mode with my 85mm f/1.4 and in m4/3+ mode with a 18-135 superzoom. I do love a good superzoom for handheld work. I even have a superzoom for the MF camera, an adapted Tamron 28-300mm Canon EF mount lens. I find it useful, even if MF purists would throw up their hands in horror. What works, works.

I agree, sensors are good enough these days for almost everything. People worry too much about the quality of gear, it's all good.


I am old school who grew up during the era when 35mm (24x36) format was considered "miniature".

APS-C is nice but that makes my vintage lenses built for 35mm photography longer in focal length by 1.5x.

Now with a relatively cheap FF Canon camera like the EOS-RP, and with the right adapter, I am able to use my vintage lenses all over again at their designated focal length.

Doesn't choice of format depend on how the camera will be used -- what the final product will be with respect to print or "on line" for example? Posters or 3 x 5 keepsakes?

Left out of this posting, any mention of APS-C. I use(d) full frame when I could afford my first model the Nikon D700, after using the 18-24mm APS-C Nikon cameras, which I always looked at as inferior, probably as a carry over from decades of film shooting where size really mattered.

I went back to APS-C with Fujifilm based on size and weight, but also the improved sensor and processor which gives me better image quality than that 12mp Nikon D700 camera.

If razor thin selective focus is not needed, I'll pick APS-C as the ideal digital format.

It is often said to be a limitation of Micro 4/3 that is harder to achieve shallow depth-of-field with it than FF, but during the days of shooting 135 film, most lenses for 24x36 weren't all that sharp until stopped down to f/8 I find Micro 4/3 lenses are quite sharp at f/4, which gives the exact same depth-of-field, so it's definitely a corresponding format to me. Add in image stabilization that is so good I can leave my tripod at home for most things, and things sure are better than they used to be!

Mike, Where does APCS fit in this framework?

Rene

I sort of agree with your designations of digital vs film formats...however, M43 is vastly superior to any 35mm film. "Full frame" digital seems to me to be at least equal to 4x5 film. So I guess I'd move them all up a notch for equivalency.

Of course, this only applies to image quality. I always considered the shallower depth of field of medium format as an advantage, not a problem. I don't think even the "larger than full frame" digital cameras get anywhere close to the shallow focus of medium format film, let alone view cameras.

Agree regarding sensors, my Nikon D750 sensor (2014) was plenty good enough for me. I have no complaints with my Fuji X-T5 sensor, 40mp is more than I need and I'd be happy with 24-26mp.

"My monochrome files look a lot like, and print a lot like, 4x5-inch Ilford HP5 Plus."

Do you add grain? (Just curious)

[No, but you can't see grain with 4x5 HP5+ at normal print sizes. --Mike]

Famous quote: “if your photographs are not good enough, you’re not close enough.” Capa didn’t say anything about sensor size.

Mike, I find that your comparison of digital sensor formats in relation to film formats aligns pretty much with my own rough approximations. It is quite extraordinary just how good digital at its present state of the art can be. And that even extends to older digital hardware when it is paired up with current processing software.

During the covid lockdown I became curious about the Fuji X sensors, and in order to keep expense to the minimum, picked up a Fuji X20 compact to play with. Loved the image quality the camera was capable of producing when given lots of light; it was very pretty from the front, and to my eye looked like a mini version of a screw-mount Leica. The idea of a Leica in my pocket, with a very versatile zoom to boot, was enticing. The lens seemed to be up to the challenge. But I didn't like the optical eye-level finder at all and the rear finder was fixed rather than being of the tilting variety, which I find to be a great knee-saver for us old guys.

Did my research and found out that Fuji had replaced that model with the X30, which had an excellent evf and a tilting rear finder. The lens and sensor were unchanged. Eureka! Took a couple of years to track down a clean copy at a rational price, but worth the wait.

Earlier this year I decided to find out just what this lens/sensor combo was capable of achieving under optimal, but real world, shooting conditions. I chose a couple of images from 2022 shot at base ISO (100), at the sharpest aperture (f/4.5), and at a very fast shutter speed (1/1000 and over) to eliminate any possibility of motion blur. The images are very sharp, with no evident noise and good gradation. One image was at an equivalent focal length of 54mm, the other at 89mm, within the broad sweet spot of the 28-112 f/2 to 2.8 zoom. The 12 megapixel sensor is a 2/3" type, which is 1/4 the area of a m4/3 sensor, which is in turn 1/4 the area of a full frame sensor. So quite small.

I proceeded to attempt a 2x uprez, trying out the several different algorithms available for output in Iridient Developer. In each case there was much juggling of sharpening, contrast, and noise reduction at both the input and output stages. Iridient offers a very broad range of controls for these factors. The files then went back to PhotoNinja, from which they had originally come, and more sharpening-contrast-noise reduction adjustment was done.

In the end the experiment was successful. Both files were sufficient to print, at very high quality, a 21x28 image area @ 300PPI. At 240ppi, which would be quite reasonable for this print size, the result is a 24x32 printed area; given a 3 inch border top/bottom and 4 inches on the sides for handling purposes, you have a 30x40.

To me that's mighty impressive result for a camera that weighs one pound and can be carried in a small pouch hung on my belt. In the mid-1980's I was involved in a project to produce a suite of 30x40 gallery quality prints from 35mm B+W negs. The logistics and technical problems that arose were surprisingly difficult and the costs were extremely high.

My understanding of the relative quality of the images created by small vs large sensors is that the differences that can most readily be seen at the print stage are primarily the size that can attained by a given file, more pixels = larger print. The quality of the X30 image is quite good, but small. Some of that difference can be made up by uprezzing the small file without giving up any print quality. I expect that using state of the art AI uprezzing, which I have not yet tried, would enable even larger prints without a perceived loss in print quality. The fact that more pixels = more detail would be a limiting factor, but it is my understanding that AI is able to somewhat mitigate that problem even now, and at this early stage of its development is improving quickly. The latest full frame sensors have the advantage of several stops more dynamic range than the X30, but I believe that even the 12 year old Fuji tech has more range than the color films that it replaced.

One important factor for me in comparing sensors is the size of what I would describe as their exposure sweet spot. Other commenters have touched on this as well. With the little Fuji, for my typical usage case, that sweet spot is f/4 to f/4.5 at base ISO of 100. Changing the aperture by even a third of a stop goes outside of the optimum range and a decrease in resolution can be easily seen when viewing the file at 100%. Increase the ISO by one stop to 200 and noise will pop up in the shadows. This sweet spot is pretty much as restrictive as it could possibly be. Jump over to an APS-C size Sony sensor and it takes four-thirds of a stop aperture change to see an appreciable difference in resolution and a three stop
bump to ISO to see much noise come into play.

I've seen examples of full frame and larger camera output that truly amaze me with their capability to produce very large prints, but I have no personal experience with them. I shoot primarily out on the streets, so compact size and minimal weight are important to me. Also, at many of the locations where I shoot here in LA it would be very foolish to be carrying thousands of dollars worth of equipment. My collection of smaller sensor cameras, (1", m4/3 and APS-C), some of them now ten years old, have been sufficient. The Fuji, with its ability to punch above its class, has been added to the rotation.

After using the Pentax K-1 for years, I got a K-3lll. Then, recently, an Olympus OM-1. I haven't mastered the OM-1 yet, with all its computational modes. All three cameras are nice in their own way, but here's the sensor-related differences:

Detail- With high quality Olympus Pro lenses, this Micro Four Thirds camera has this aspect of image quality nailed. They're sharp at 100%, and very good at 200% view in LR. With the Pentaxes, I judge equivalent sharpness at 300%, but difference is nothing any ordinary person would see online

Dynamic Range - This is where the Pentaxes shine. They give a perfectly malleable RAW file that allows me to use 100% of highlight reduction and shadow lift with no ill effects. The exposure latitude is hemispheric. The Oly will blow highlights on winter ice or cloud highlights if I'm not careful.

Noise - The K-1 is best of the three, of course. The OM-1 can be noisy even at three-digit ISOs... or not. I've had stage performance around 1600 that showed almost no grain, but evenly-lit indoor scenes can show grain as low as 640. Current Adobe noise reduction is an effective leveler here, if you take the time.

There's one other difference between these camera systems and formats that matters to me: aspect ratio. My K-1 offers OVF crop lines for APS-C and square format. I find that quite useful, especially with primes. The K-3lll has 1.3x crop lines intended for video, and 1:1. But none of them have crop lines for 4:3, which is my most favored aspect ratio. I ran through too many orange boxes of Agfa 8x10" paper, back in the day; that shape just looks more like "art." Other makers have avoided the 4:3 shape since it's become a trade name for their Micro 4/3 competitors. Most of my 2:3 images get cropped to 4:3, especially the verticals.

The OM-1 shoots in any common aspect ratio. Most modern mirrorless cameras do too, I suppose. But it's the one that does 4:3 as a native format, using all the lens' capabilities and viewfinder real estate. That could be the tie-breaker for me. Having three systems does't make much sense to me. If I were to sell one, it would be the APS-C K-3lll, keeping the FF and MFT because they're so different. In film terms, FF is print film, accurate and forgiving, and MFT is slide film: punchy, colorful, demanding.

As an addendum to the discussion: Stephen Shore did a series on the (then) new Hasselblad X1D (50MP). He said in several interviews that he felt the 50MP sensor was the equal of 8x10 film.

The comments to this entry are closed.

Portals




Stats


Blog powered by Typepad
Member since 06/2007