Why the Megapixel Race Needs to End

[soundsop]

http://blog.wired.com/gadgets/2008/08/why-the-megapix.html

[apu]

I work in computer vision, and there's an entire sub-branch called "computational photography" that aims to do more interesting things with cameras. The idea is that by capturing more information at the hardware level -- perhaps in some 'coded' form -- we can do much more with the captured data. Things like getting better quality images, getting depth information, capturing HDR images in a single image, etc. Many of these techniques require trading off spatial resolution for their benefits, thus making the increasing number of megapixels actually useful (i.e., so that when you lose 50% of your resolution to capture some other information, you're still left with a decent size image).

A good page describing this field is http://www1.cs.columbia.edu/CAVE/projects/what_is/

[rw]

Do not conflate orthogonal issues. Pixel count increases, all else being equal, are good. Any visual artifacts caused by that increase are the result of other factors. The author's real problem is with the design of individual sensors, not the quantity of sensors.

"Good enough" does not mean "stop innovating."

[soundsop → rw]

Pixel count increases, all else being equal, are good.

The problem is that all else is not equal. Recent pixel count increases have come at the expense of pixel area, since the total sensor area is not increasing. Thus there are fewer photons per pixel, decreasing the per pixel signal-to-noise ratio.

[jmah]

Calm down, it'll end like the MHz (GHz) race did.

[Herring → jmah]

Incidentally, why was the MHz race a swindle?

[josefresco → Herring]

The MHz race wasn't a swindle, when MHz mattered it was a simpler world were you could compare processors without having to lookup meaningless product numbers like T850 or TL-60.

The only thing you needed to remember was the Celeron/P4 distinction (and the AMD equivalent)

Now each chip company spits out new products at higher and lower MHz which makes it even more confusing to consumers. Is a P4 4 GHz chip faster than a Core Duo 2.16 GHz, is a typical question I get from friends/family.

[mechanical_fish → Herring]

I wrote half of the answer and then decided that it was better just to say:

http://en.wikipedia.org/wiki/Clock_rate

[bprater]

Heard of downsampling? There is a reason that HD video looks better on a standard television.

[shadytrees]

That's one damn erotic campfire.

[vaksel]

they should still have the race for the professional cameras, but for your basic point and click its way overboard.

[aschwo → vaksel]

Even when considering professional cameras, it's more about the size of the sensor than raw megapixel output:

"Bigger chips mean so many good things. Big chips mean shallower depth-of-field at equivalent settings. Big chips mean big pixels, which means more light hits each photosite. More light means less gain, less noise per pixel. Less noise means more dynamic range. Bigger chips mean better pictures."

http://prolost.blogspot.com/2008/07/panasonic-is-my-hero.htm...

http://rebelsguide.com/dl/sensorSizes_06_cheatSheet.png