CCC Forum

Photokina used to be THE show where manufacturers would vie with each other to launch the best product. We are still seeing that, although three major "in development" announcements is unprecedented. What we have not seen is more interesting - a chase for megapixels. Nikon might have been expected to launch a DSLR with a 42 MP sensor, but after the D5 and D500 earlier this year is having a fairly quiet Photokina. Sony seems content to let Canon maintain its bragging rights with its 50 MP full frame sensor, instead putting its 42 MP sensor into the new A99 DSLR. Apart from upgraders who have an investment in A mount glass, if ever there was a camera with limited appeal the A99 has to be it. Yes, Canon, Fuji and Olympus have upped megapixel counts, but within sensible limits. Fuji in addition is going head to head with Hasselblad in the new mirrorless medium format segment where 50 MP is fe rigeur.

One explanation for the lack of a new mega sensor is the earthquake in Japan earlier this year, which is known to have hit production. Maybe plans are simply on hold. We shall see, but I prefer to think that the megapixel race is finally over. 50 MP images might look good on the screen and provide generous cropping options, but the reality is that such sensors are more of a hindrance. They demand the best lenses and show up the slightest camera movement. Some years ago I spent a day with a 37.5 MP Leica S2, which has no image stabilisation, and found that handheld with the standard lens, a shutter speed below 1/250 was likely to result in a less than optimally sharp image. It might look like an oversized DSLR, but it cannot be used like one. There are a few people who really need such high resolution and they will be the ones making really large prints. For everyone else, we have reached peak megapixel.

After posting the above, I found this on the Amateur Photographer website. I had forgotten about Canon's 120 MP monster which it is showing at Photokina and, amazingly, is just APS-H size. At present it is a proof of concept and a demonstration of what Canon can achieve, but even their rep says they are unsure of the demand and thinks it might be limited to specialist use. The Raw file is 211 MB, so disk space is going to disappear at an alarming rate and the image will require a fair chunk of CPU power.

http://www.amateurphotographer.co.uk/la ... demo-95831

I have just rediscovered this article by Keith Cooper, who is behind the excellent Northlight Images website, a mine of information for anyone interested in the various aspects of post capture processing. There is also a rumours section, which Keith emphasises is just for fun and is an entertainment in itself when looking at what new camera gear goodies might be coming our way.

In the article, Cooper shot the same scene using three different Canon cameras of varying vintages using the same tilt shift lens. The cameras were the original 1Ds (11 megapixels), 1Ds III (21 MP) and 5Ds (%1 MP). All have full frame sensors and Cooper used the 1Ds III for his commercial work over a period of 8 years before upgrading to the 5Ds. In the article, he notes differences about the way they perform and in particular how they handle shadow noise. What he really wanted to know was how images from the three cameras compared when printed. For the test, he produced three prints from each camera, at three sizes - 13"x9", 19" x 13" and 29"x19".

The article is quite long and much of it details his workflow in producing the prints. The bottom line was that at the smallest size, which is roughly equivalent to the size of most prints submitted to club competitions, there was effectively no discernible difference. The 5Ds did improve matters in the larger print, but not by much and it really needed a magnifying glass to see the difference. In a real world test, no one was able to determine which camera had been used. The 5Ds does have some advantages over its predecessors, but these are largely incremental apart from the ability to crop more tightly.

To me, this shows theree things:

• Megapixels are not that important and it is other improvements in sensor technology which yield greater benefits.
• Those who do claim that more megapixels are better are usually pixel peeping rather than examining prints.
• Sensor technology matured several years ago.
• Good image processing technique is key to getting the best result, whatever camera was used to capture an image.

It is worthwhile bearing in mind that those who never print their work do not really need many megapixels at all. Few people will post full size images on the web and in a DPI competition the club's standard 1400 x 1050 requires just 1.5 MP. Even the latest PAGB requirement of 1600 x 1200 requires less than 2 MP. Both figiures assume that a picture is presented with a 4:3 ratio. Anything other than that will require even fewer pixels.

http://www.northlight-images.co.uk/arti ... rison.html

I came to a slightly different conclusion, that the people this guy asked to review his prints did a piss-poor job of finding differences. It's very clear to me And how does pixel peeping differ from examining prints, surely they are peeping but just on a different medium

The "who needs more megapixels" argument isn't just about prints though. There are other benefits, such as the ability to more aggressively crop, that means more is better than fewer. Whether or not it's something an individual needs comes down to what they shoot and how they present it.

And the guy actually pixel peeps his prints using a pocket microscope

http://www.northlight-images.co.uk/revi ... onclusions

davidc wrote:I came to a slightly different conclusion, that the people this guy asked to review his prints did a piss-poor job of finding differences. It's very clear to me And how does pixel peeping differ from examining prints, surely they are peeping but just on a different medium

Cooper is not saying that the print quality is the same regardless of camera. Examined closely, as the crops of the largest prints clearly demonstrate, there is a benefit accruing from the higher resolution camera. At the smallest size, he says there was no difference. But no one, not even judges*, examines prints that closely. At normal viewing distances, it is not possible to discern between them. Our eyes are not capable of seeing that much detail without a viewing aid to boost magnification. Prints have to be made very large before the advantage of higher resolution kicks in. There are definitely some people working professionally who need every megapixel they can get, but for most it is a non issue.

davidc wrote:The "who needs more megapixels" argument isn't just about prints though. There are other benefits, such as the ability to more aggressively crop, that means more is better than fewer. Whether or not it's something an individual needs comes down to what they shoot and how they present it.

No one is disputing that, although higher resolution brings its own issues as I have previously described in my initial post.

* Not yet, anyway.

davidc wrote:And the guy actually pixel peeps his prints using a pocket microscope

http://www.northlight-images.co.uk/revi ... onclusions

Maybe a sensible precaution for someone selling his work and wanting to achieve the highest standards?

As a follow up, it is worthwhile recounting part of what happened when I went to the recent Wayne Johns seminar which was part of the recent Fuji X-T2 launch. He showed us some large prints made from images taken on Fuji cameras which were somewhere between A3 and A2 in size. I happen to be very short sighted and without spectacles the only point at which my eyes can focus is around 4" from my face. In general that happens to be something of an inconvenience, but it does have one distinct advantage; I can examine things very closely with a naked eye. Wayne Johns is a fashion photographer and the shots were of female models with the focus point on their eyes. Looking at detail in the eyelashes, I can confirm that those prints were indeed as sharp as anyone could wish.

The kicker was that Johns had been away and had only just received an X-T2 from Fuji, so did not yet have any prints from shots taken with the camera. As you would expect, he has a X-Pro2, but the first prints he showed us were made with a X-T1, which has a mere 16 MP. Some X-Pro2 shots came later and even with my close vision I could not see any difference. At that point, the Fuji reps present should have been having kittens. I believe the argument was supposed to be if that is what can be achieved with 16 MP, imagine what the new 24 MP sensor can do. An alternative view is if 16 MP is that good, why bother with a new camera? I did ask about sharpening and printing. The former had occurred using Johns' own settings, which he did not disclose, and he is able to send files to be printed which are a quarter of the final size although he did not know how that was accomplished.

That last answer is very similar to the one I got from Damien McGillicuddy, who is an Olympus ambassador, about the 16 MP images he was getting from an E-M1. He said that prints were being made with a longest edge of 4' or 5' from his shots.