Steve's Digicams Forums

Steve's Digicams Forums (
-   What Camera Should I Buy? (
-   -   MegaPixels do not matter - NYT article (

interested_observer Nov 22, 2006 9:34 AM

I ran across this posting about Megapixels do not matter. The author took the results from a 5, 8 and 13 MP cameras and created a set of 16x24 posters - only one viewer in Time Square could tell the difference....

slipe Nov 22, 2006 11:40 AM

Actually, I'll tell you about one thing right now. We did an episode on digital cameras.

I would be more comfortable with someone who does photography as his main pursuit rather than a self proclaimed geek who did a single show on digital photography.

One print had 13-megapixel resolution; one had 8; the third had 5. Same exact photo, down-rezzed twice, all three printed at the same poster size.

The results differ from mine sufficiently that his results are suspect in my mind. Printing 13 X 19s I can see a definite difference between photos starting with 5 Mp and photos starting with 7 or 8 Mp.

A 16 X 20 with 5 Mp gives an image of 125 PPI. I don't know exactly what his workflow was, but I always see improvements in prints when I have a source that generates over 125 PPI.

You have to start with a 6 Mp image in 3:2 format to end up with 5 Mp from a 16 X 20 crop. Which brings into question where he got the 13 Mp image at 16 X 20. You get close to 14 Mp from a 1Ds MkII and just over 11 Mp from an EOS Mk5. If he actually started with a 13 Mp image directly from a camera he must have cropped a MkII image a little more.

People seem to overestimate the value of an upsample. If you don't have enough pixels an upsample does little. So where he got the original 13 Mp image is important to the results. If he upsampled to 13 Mp and then downsampled to the other sizes I generally agree with his results. But starting with an unresampled 13 Mp image directly from a camera his results are puzzling. I'll probably have to watch the show to find out.

Gozinta Nov 22, 2006 11:53 AM

His test means little either way unless we know more about the equipment used.

mtngal Nov 22, 2006 2:50 PM

I read through the comments at the end of his blog. Most mentioned the difference in camera equipment, and he responded at the very end (at least it was at the end when I read it this morning) that he didn't want to use different cameras and different files because that brings in the question of glass and equipment - he was concentrating only on megapixels. That makes it sound like he used one file (or 3 different pictures from the same camera)then resized accordingly. Don't know if that's true because he didn't actually say that, however that's about the only conclusion I could come to if he really did take out any camera characteristics out of the equation.

romphotog Nov 22, 2006 6:53 PM

5.0 MP - (2576x1932);
4.4 MP - (2576x1716), 3:2;
4.0 MP - (2304x1728);
3.2 mp - 2048 x 1536

1.8 MP - (1552x1164)

doubling resolution requires quadrupling the number of megapixels.

8mp is _not_ double of 4mp.

4sqrt = 2

8sqrt = 2.83(not double of 2)

16sqrt = 4(double of 2)

16mp is double of 4mp.

100 = 10 x 10

200sqrt = 14.14 (not double of 10)

400sqrt = 20(double of 10)

400 = 20 x 20


4mp = 2000 x 2000

8mp = 4000 x 2000(_not_ double resolution of 4mp)

16mp = 4000 x 4000(double resolution of 4mp)

All times are GMT -5. The time now is 1:55 PM.