Breaking the Myth of Megapixels

[paulbk]

This from the New York Times..... "The myth that the more megapixels a camera has, the better the pictures, is a big fat lie."
Breaking the Myth of Megapixels
Comments....?

[Ray]

This from the New York Times..... "The myth that the more megapixels a camera has, the better the pictures, is a big fat lie."
Breaking the Myth of Megapixels
Comments....?
[a href=\"index.php?act=findpost&pid=99789\"][{POST_SNAPBACK}][/a]

Very amusing!    There is a myth here, but it's not so much do do with the fact that more megapixels don't matter, period, but more to do with the question, 'matters to whom?'

Assuming that the tests and comparisons were carried out competently, using sound technique and methodology, the results would tend to indicate that the average person in the street, taken at random, does not think about or consider such matters as sharpness and resolution in images. The few who were able to separate the lower rez images from the higher rez images were probably part of the relatively small minority who might occasionally or frequently view such sites as Luminous Landscape.

I bet, if the author of this article had asked the passers-by who agreed to take part in the test, 'What do we mean by the resolution or sharpness of an image, most of them would have demonstrated they didn't have much of a clue.

A similar situation exists with hi fi audio. You could compare the most accurate loudspeaker that money can buy, a speaker having inaudible distortion and a flat frequency response from 20Hz to 20kHz, with an el cheapo boom box. Ask people at random from the street which they prefer, and a good proportion would claim to prefer the boom box.

[Raoul]

It would be good fun to make the amount of money these bypassers spent on camera equipment part of the equation. How many people having spent less than 500 $ can see a difference? how many in the 500-1500$ bracket ? more than 1500 $?

Actually my hearing is worse than when I was 15, but my stereo equipment is way better than at that time...  

[BJL]

Pogue's summary seems reasonable: not that "more MP is never relevant" or even that "more than 5MP is never relevant" or "more than 5MP after cropping is never relevant". Instead he says that

“For the nonprofessional, five or six megapixels is plenty, even if you intend to make poster-size prints.”

That is a bit imprecise, because some "non-professionals" are very demanding amateurs or "photographic connoisseurs" and this comment does not apply to them, and sometimes people will scrutinize poster sized prints closely (say to examine a face in a group portrait) and then should easily see the differences. But the gist of it seems fair. I would guess that, with the large print size used, most viewers stood back at a distance that fits the image into a comfortable field of view, two feet or more, and at that "angular image size" I can easily believe that 5MP is about all that most people readily resolve.

Let me add my own anecdotal evidence: many people who look at my selected favorite images from my 5MP camera comment on the good quality, even though they are surely used to seeing lots of snapshots at 5MP and up, and eschewing false modesty, I see the same when I look at many other people's compact digicam images. Immodestly, I suggest that some of that comes down to my "better than snapshot" efforts with focus and steadying of the camera; more modestly some of it probably comes from the fact that my 5MP comes from a well designed DSLR (Olympus E-1) with 6.8 micron pixels, and lenses (Olympus 50-200/2.8-3.5 and 14-54/2.8-3.5) that are likely better than most digicam lenses.

And the IQ differences are in fact often clear in on-screen viewing of the entire image (not 100% crops) where screen resolution is less than 5MP. On-screen like this, sensor resolution is probably irrelevant from about 4MP up.


Dare I suggest that for many snap-shooters, camera shake and/or compact camera AF imprecision are often limiting image sharpness to no more than about what a more skillfully used 5MP sensor gives? And that for their purposes, like on-screen image display, this does not much matter?

[thechrisproject]

I remember reading about the original test.  A lot of people argued over the methodology he employed.  It's nice to see this followup article.

[Dale_Cotton]

Breaking news! Ansel Adams, the iconic American landscape photographer long believed dead, is in fact still alive and residing incognito in a safe house near Boswell, New Mexico along with Elvis Presley and a mysterious figure some believe to be Jimmy Hoffa, although this has not been confirmed. We coaxed Mr. Adams out of hiding last week to get his views on digital photography in general and the recent megapixel controversy in particular.

"Yes, it's true: I have been actively involved in photography through the digital revolution. I started with a 2.6 megapixel Nikon D1. Then I upgraded to a Nikon D2X when that model was released - but did not find the extra megapixels to be that important. I now shoot a special edition, titanium body, 3.1 megapixel Coolpix 990, rigged with a custom-made accordion lens shade and a carbon fiber Velbon tripod, specially manufactured to my own specifications. I find this lightweight combination gives me exceptional image quality in prints up to 20x30 inches; and, in any case, everyone I show my work to seems to be able to tell the difference between Half Dome and the Grand Canyon, which of course is all that really matters.

"Having experimented with many tools, my digital 'darkroom' now consists of the very spartan combination of QImage and an Epson 9600 printer. Of course, I experimented with Photoshop and similar software in the beginning.  But if there's one lesson I've learned over the years, it's that a good photographer gets the image right in the camera. Fussing about in post-processing is strictly for losers. [Chuckles] I remember how I used to spend hours dodging and burning every print under a dim red bulb in my darkroom - how embarrassing to look back on the follies of one's youth!"

[Kirk Gittings]

Ellis Veneer is a good friend of mine and a dedicated writer/tester on photography. he does thoughtful work.

About MPs........I don't really care if only a small minority of people can see the difference.

I CAN SEE THE DIFFERENCE

Those same crowds of people probably couldn't tell the difference between a 16x20 enlargement from 35mm film vs. 4x5 film too. So what! I can see the difference.

It is my standards I am trying to meet. Not the McDonalds eating masses.

[Jack Varney]

Am I missing something in the NYT article? Pogue states that they printed at 400 dots per inch to a size of 16X20 inches. Certainly the printer must have had better than 400 dot per inch capability. Did Pogue mean that they printed with an image resolution of 400 PPI?

If he meant 400 PPI all of the images would have to be uprezzed by a factor of four or more. Would this not impact the credibility of the test? Or have I  overlooked something?

[thompsonkirk]

Why is Pogue pretending that all point-&-shoot & DSLR pixels are equal?  

He's skipped the basic issues of pixel pitch, or of how big the sensor & its pixels are in the first place.  And he forgets about noisy files at higher ISOs.  These are the basic problems of small sensors with too many pixels, vs. larger sensors with the same number.  

My guess is that most passers-by on the street could see which print was made from a file with a poor signal-to-noise ratio (without necessarily having the words to talk about it).  They'd say it looks kinda gritty over there in the shadows.