Things looking less good for DSLRs?

[William W]

It occurs that there may be more to this than just Canon deciding to decamp from producing their EF lenses. . . . That said, there an enormous volume of great EF glass out in the marketplace and sitting in various stockpiles. That is a lot of glass that is not all suddenly going to disappear off the face of the earth.

Yes, I agree on both points: it is a downturn in production - and - there are a lot of EF Lenses out there, all of which can be used for a long time.

Same point I made on the other conversation.

WW

[DanK]

Trev,

Thanks very much for the kind comments.

In fact, I think very few people notice. After all, most people rarely print, or never do, and 20 MP is far more than enough to display online.

Even for printing, I have to admit that it doesn't seem to matter very much to most viewers. I've had a few prints that I thought were clearly not up to snuff that got hung anyway. One is a fairly severe crop taken with a 15 MP 50D and then printed A3+. I think it's soft, but no one else seems to notice, and my wife had me hang it in our house. The other is a shot that is slightly soft because of operator error. Again, no one seems to care. I have an A3+ print from it in a gallery, and one person has requested an A2 (17 x 22) print and didn't change his mind even after I sent him an 8 x 10 of 1/4 of it to show what it would look like at that magnification.

I also have an A3+ print that I made from a substantial crop from a 12 MP MFT camera that has gotten very positive comments and is in a gallery.

So this is why I think I may be chasing my own tail much of the time. I prefer having the leeway provided by more pixels, but I suspect it rarely matters much to most people viewing the photographs. On the other hand, once in a while I encounter someone who really does pay attention to the quality of the print.

Dan

[revi]

Isn't viewing distance important when considering large prints?
Yes, if you look at a print from 30 cm, it should ideally have a resolution of about 300 dpi.
But do you normally look at an A2-sized print from 30 cm (12'', if you prefer)?

With that in mind, I'd expect a larger version of an acceptable print to be acceptable as well
(as the viewing distance should increase with the size of the print).

[DanK]

There's another variable in this: print heads have a specific pitch. For example, Canon printers, which I believe have a native resolution of 300 dpi, have a nozzle pitch of 600 dpi.

So one can't print at any resolution one wants. Rather, the software and firmware have to translate somehow between resolution determined by file and print sizes and the nozzle pitch of the printer.

When printing from Lightroom, this is controlled in two places. One is the "quality" parameter in the printer's firmware. The second is a print resolution option in the Lightroom print module.

I always have the latter turned on. I recently ran a test to see whether print quality is improved by setting the resulution to the nozzle pitch (600 ppi) rather than the "native" 300. The answer is no: quality was lower when set to 600.

So for my printing, I get 300 dpi regardless of the resolution of the image. The issue for me is what happens when I have too few pixels for 300 dpi. More specifically, how well does resizing, or using Super Resolution, stack up against having real pixels? That's the question I have been trying to address in this thread.

However, there is another option, which I have never tried, which is to turn off control over resolution in Lightroom. I have never tried this and have found only brief descriptions of it, so I may not know what I am talking about. However, from what I have read, the software keeps the resolution determined by the file and print sizes, as long as the result is between 72 and 720 ppi, which is the software's limits. Outside of those limits, it resizes the image. If the image were, say, 150 ppi, the mapping to the nozzle pitch would be simple enough. But say, for example, the native resolution of the print would be 123 ppi. It seems to me that this mapping could get quite messy, and perhaps visibly so. I'm not inclined to spend time or ink trying this

This is a long way of saying that for me, this boils down to a few questions:
1. How good a substitute are imputed pixels, imputed in various ways, for real pixels, and how does this vary as a function of the type of image?
2. How often will most people notice the difference?
3. How often will I notice the difference?

One last thought about viewing distances: I don't think people pay much attention to recommended distances. Watch people in an art gallery. Many stand at a distances that is clearly not optimal. Some--I'm one of these--often walk forward and backward, to apprehend both the gestalt and the details. When I hang an A2 print, I can't assume that people won't come closer than 73 cm. Still, I take the general point, which is that people may often be far enough away that all of this is moot.

[pnodrog]

I tend to judge a photograph from a comfortable viewing distance and for a lot of subject matter I don't bother to pixel peek. However for some types of photography (and some of Dan's do fall into this category) I am sufficiently curious to examine as much detail as possible. This pixel peeking can add a little bit to the enjoyment regarding the technical aspects of the photograph but certainly won't alter my initial judgement as to whether it is a photograph worth looking at.

It needs to be a very tasty cake before I worry about what the icing looks like and if it tastes superb who cares?...

[lunaticitizen]

Dan, if high resolution is important why did you choose the 5D Mark IV instead of the 5Ds R?

[DanK]

Leo,

First, because I wasn't trying to maximize resolution. Doing arithmetic like that here, I figured that an increase of roughly 50% from the 22 that I had with my 5D III would be about the sweet spot: enough resolution to avoid visible problems when I crop substantially and print as large as I do. I told someone at the time that if my 5D III had around 30 MP and a little more dynamic range, I'd be a happy camper. The 5DIV offered both, plus a number of other goodies.

Second, there are tradeoffs. All other things being equal, larger photosites perform better than smaller ones. If you compare the dynamic range and noise for the 5D IV and 5DS, you'll see that the IV is better with respect to both, particularly noise. The S also has limited low light performance. So for my purposes, getting more MPX than I think I need at the cost of more noise, less dynamic range, and more limited low-light performance would be a bad trade.

When the R5 and R6 were released, I was hoping that they would have a model in the 30 MPX range. I doubt they will. One needs roughly 45 MPX for 8K video (which interests me not at all). On the other hand, the data I've seen indicate that Canon's sensor design has improved so much that the drawbacks of the 5D S have been largely eliminated.

What this thread suggests is that even the increase to 30 MPX might not matter all that much at the sizes I print, especially now that Super Resolution has been introduced.

[lunaticitizen]

I see. Do you have any example on your website that would have been unacceptable if it had been shot using a 5Ds R? I'm honestly wondering why the camera got such a bad rap about its DR and noise. I understand the numbers; I just want to know how it translates in real life.

[DanK]

I see. Do you have any example on your website that would have been unacceptable if it had been shot using a 5Ds R? I'm honestly wondering why the camera got such a bad rap about its DR and noise. I understand the numbers; I just want to know how it translates in real life.

I would just be guessing, as I never used a 5DS. Since one of my two main reasons for wanting to replace my 5DIII was to increase dynamic range, and because I wasn't interested in more than about 30 MP, I didn't look seriously at the 5DS other than reading a bit.

At any one time, smaller photosites generally mean poorer performance on some dimensions. This is offset to by progress in sensor design. For example, look at the graph below, which is from photonstophotos.net. The black line is my 5D IV. The light blue is the 5DS. The green is the R5. The noise of the 5DS is appreciably greater throughout the ISO range, while improvements since the IV was introduced allowed Canon to increase the pixel density by 50% from the IV without a consistent increase in noise. Notice also that the native ISO range of the 5DS is much smaller.



If you look at dynamic range, the differences are apparent at the low end of the ISO scale, which is where I usually shoot. The IV beats the S by about 1 EV at base ISO, while the R, despite its smaller photosites, beats the IV by roughly another stop.



The bottom line is that the 5DS was pathbreaking at the time, the highest-resolution FF body on the market. That entailed serious compromises that didn’t matter to some people but do to me. The R5 has changed the calculation for me: I don't see any need for the additional 15 MP compared with what I have, but the data suggest I would pay no performance penalty for it. I opted not to go for the R5 primarily because of cost; I bought the IV while it was on special last June and the net cost of switching to an R5 was literally 3 times the net cost of upgrading to a 5D IV.

[Donald]

I'm honestly wondering why the camera got such a bad rap about its DR and noise. I understand the numbers; I just want to know how it translates in real life.

I've never heard that. The 5DS is a brilliant camera and I don't see any difference with the 5DMkIV (which I have converted to IR).

[Manfred M]

I've never heard that. The 5DS is a brilliant camera and I don't see any difference with the 5DMkIV (which I have converted to IR).

The "bad wrap" always seems to come from people that look at camera specs and test results and make pronouncements based on data. "Real photographers" figure out the limitations of their gear and go out and make strong images.

The only time I get really concerned about camera and lens performance is when making large prints, where I am pushing my equipment and photographic skills to a much higher limit than the 2MP images we see on the internet.

[Donald]

The "bad wrap" always seems to come from people that look at camera specs and test results and make pronouncements based on data. "Real photographers" figure out the limitations of their gear and go out and make strong images.

Agreed, which is why I've steered clear of this Thread up until now. It excites some people, but such discussions involving tables and test results etc., turn me off totally. My biggest concern is the beginner who sees such a discussion such as this and thinks they must have to be able to converse like this in order to be a good photographer. I certainly once though that way when I started out. Let me use this post to assure any of them who may be ploughing through this thread thinking that they'll never be a good phoptographer:- NO YOU DON'T.

Get out there and take photographs and then come home and learn how process RAW files so that you get the image you wanted when you fired the shutter.

[DanK]

Wow.

I started this thread simply to inform people about Canon's decision about supplying EF lenses.

This segued, as threads often do, into a discussion of something else: a discussion of which cameras best suit some people's needs. Doesn't seem inappropriate to me. As Manfred posted:

The only time I get really concerned about camera and lens performance is when making large prints, where I am pushing my equipment and photographic skills to a much higher limit than the 2MP images we see on the internet.

Indeed. And its the former use I consider when deciding what equipment I want. That's what this discussion was about: detail when printing fairly large.

That led to what I thought was an appropriate and informative discussion about whether more resolution than 20 MP matters when printing 17 x 22 (roughly A2), particularly given the advent of tools like Super Resolution. I don't know the answer to that and would like to. That is, the discussion suggested to me that my priors might be wrong. People who are uninterested needn't have read further.

Re the 5DS: I for one didn't even bring it up until Leo asked me why i didn't buy one. I answered him, with a clear statement that I had no first-hand experience and showing him some data I had looked at.

I, for one, never made any comments suggesting that everyone needs to pay attention to this or that one needs to share these interests to be a good photographer. In fact, I specifically mentioned that I have successful prints made with equipment inferior to that discussed here. I didn't suggest that beginners buy equipment of the sort discussed here, and when asked, I have generally advised them not to.

Actual data matter. I have shot with cameras with good noise characteristics and bad noise characteristics, and the difference is clear. That's not to say that one can't be successful with the latter. I have won prizes with images taken with a 50D, which is a relatively low resolution and noisy camera that handles low light poorly. Notwithstanding that, my 5D IV produces visibly better images under some conditions and allows me more flexibility, e.g., allowing me to raise ISO more than I could with the 50D. Similarly, for my uses (for example, urban night photography), dynamic range--that is, the data I showed in one of the charts--actually does matter in practice. Ditto, measured differences in lens characteristics. I can get images with my 100mm prime with edge-to-edge sharpness that I simply can't get with one specific zoom that I often use. Charts like this show real optical phenomena. They will matter in some cases but not in others.

[xpatUSA]

Originally Posted by Manfred M The ["bad rap"] always seems to come from people that look at camera specs and test results and make pronouncements based on data. "Real photographers" figure out the limitations of their gear and go out and make strong images.

Agreed, which is why I've steered clear of this Thread up until now. It excites some people, but such discussions involving tables and test results etc., turn me off totally. My biggest concern is the beginner who sees such a discussion such as this and thinks they must have to be able to converse like this in order to be a good photographer. I certainly once though that way when I started out. Let me use this post to assure any of them who may be [plowing] through this thread thinking that they'll never be a good [photographer]:- NO YOU DON'T.

I take the opposite view. In photography, as in any other endeavor, knowledge is power and, the deeper the knowledge, the more powerful is that knowledge. While such knowledge does not necessarily make a person a "Real" or a "Good" photographer, it most certainly enables one to take better photographs.

Get out there and take photographs ...

For one whose main interest is in the technicality of photography, I've always found that commonly-given advice quite irritating.

... and then come home and learn how to process RAW files, so that you get the image you wanted when you fired the shutter.

Presumably recommending "how to process raw files" to be on a suck-it-and-see basis with no technical consideration.

[Tronhard]

I think that there is a place for the two perspectives to exist. The issue is when one thread theme morphs into another. Perhaps when that happens our moderators can perhaps mandate starting a new thread on the subject and move the first divergent post to that thread, leaving the original one to carry on in its context, but allowing the technical discussion to get oxygen in its own space?

[Donald]

Presumably recommending "how to process raw files" to be on a suck-it-and-see basis with no technical consideration.

Got it in one, Ted. Try it sometime. You'll be staggered by the amount of freedom it gives you.

ps - 'ploughing' was spelt correctly. Thanks all the same.

[xpatUSA]

Got it in one, Ted. Try it sometime. You'll be staggered by the amount of freedom it gives you.

Good one, Don!

ps - 'ploughing' was spelt correctly. Thanks all the same.

Ah, sorry, divided by a common language yet again. My US spell-checker took exception to it but allowed "plowing" ... ho hum.

[Manfred M]

I think that there is a place for the two perspectives to exist. The issue is when one thread theme morphs into another. Perhaps when that happens our moderators can either mandate starting a new thread on the subject and move the first divergent post to that thread, leaving the original one to carry on in its context, but allowing the technical discussion to get oxygen in its own space?

When it comes to equipment, it is challenging to sever the two sides, as the technical (and as Dan pointed out, price considerations) versus what we do with the gear.

My issue with concentrating on the technical side is that it defines an upper limit that rarely gets utilized by most photographers; whether that be the "sweet spot" of a lens, the high resolution / quality the sensor can record, etc. The data we see is lab-based data , generated using cherry picked examples and used under ideal conditions. As Dan points out, often has little or no impact on the impact the final image and even when it does, the image is still good.

Yes, sometimes it can have an impact and I can see the difference as I do own a number of high-end pro lenses that I shoot on what was a high end body (back about 3-1/2 years ago when a newer, better body was introduced). That being said, I can honestly say, this is only the case when I make a large format print. I can't tell (unless I pixel peep) on a computer screen.

The only parameters I ever looked at before buying a lens was the focal length and maximum aperture.

What is important to me when I capture an image is that I have full control of key parameters; so I use either flash (mostly studio flash, even when shooting on location to freeze motion) and / or a very sturdy tripod and head. On the occasions I hand-hold, I use good shooting technique (ones we developed back before there was either in body or in lens image stabilization) to minimize unwanted camera movement.

Strong composition and images with that "wow!" factor are only loosely linked to the equipment we use.

[Tronhard]

Hi Manfred:

My point is this. The title of the thread was quite clear. When it morphed away from that to a detailed technical discussion of the merits of specific equipment, then I think a line had been crossed that meant that the new direction deserved its own space. That would allow those wishing to follow either or both to have a clear point of discussion. I agree with Donald that such a deviation as we have seen here took it well outside the parameters of where, say, a relatively new person might look for advice on the direction that the camera market is taking relative to DSLR vs MILC gear.

[Cantab]

I think that there is a place for the two perspectives to exist. The issue is when one thread theme morphs into another. Perhaps when that happens our moderators can perhaps mandate starting a new thread on the subject and move the first divergent post to that thread, leaving the original one to carry on in its context, but allowing the technical discussion to get oxygen in its own space?

I agree.

I understand the dangers of newcomers to photography becoming overly focussed on technical data, etc. But CiC is big enough to allow technical discussions to occur, even if those discussions may not have real world implications for the vast majority of photographers.