Jump to content

Canon A640....10MP too much?


Recommended Posts

I have a Canon A620, which I love, but I had a chance to handle and fool around

with the new A640 at the store today. Very nice little camera and it feels more

substantial and well built than the A620, though I noted it only goes down to 80

ISO compared to the 620's 50 ISO.

 

I read the two or three reviews on this camera, but they don't seem to address

my main concern: Is the 640 with 10MP overkill, or are there any real benefits to

10MP over the 620's 7MP in a sensor this size?

 

Sorry if this is a silly question, but I'm not really up on the technical

details of the current digicam sensors.

Link to comment
Share on other sites

My basic rule of thumb is that in order to extract all the detail that a sensor can capture, you need to shoot at an f-stop value equal to or smaller than the pixel size in microns (and you need the lens to have very few aberrations at that aperture).

 

The pixels on the A640 are 2 microns (3648 pixels in 7180 microns), which means that in order to extract the best out of such a sensor you'd need an excellent f/2 lens. Since the lens of the A640 is f/2.8 at the wide end and f/4 at the long end, the pixel count in the A640 is a bit overkill.

 

If you want 10MP that really translate into a very high image quality, you need a Digital Rebel XTi and some good lenses (pro zooms, or single-focal lenses).

Link to comment
Share on other sites

Edward:

 

Norman Koren has plenty of information about various aspects of resolution, including a big blurb about diffraction at the following URL: http://normankoren.com/Tutorials/MTF6.html

 

In a nutshell, if you could have an optically perfect lens, it'd get less and less sharp as you stop it down because of diffraction. With a digital camera, there'd be a point from which that loss of sharpness you be such that it would affect the final image quality.

 

What remains is to figure out at which point diffraction starts to be actually measurable in the real-world (Norman has plenty of good math out there, but I don't want to discount experimental data).

 

I use the measurements done at photozone to experimentally find the "critical" aperture. The most interesting lens to use here is the 35/1.4L: http://www.photozone.de/8Reviews/lenses/canon_35_14/index.htm

 

Scrolling down to the MTF data, you can see that from f/2 to f/5.6 the sharpness of a 35/1.4L on a DRebXT is pretty much the same, which means that it's primarily limited by the sensor in that range. At f/8, the values start to fall, meaning that diffraction is starting to be so severe that pictures shot at f/8 are less sharp than the sensor can resolve. In fact you can check through the various tests that good lenses on the 350D are consistently sharper at f/5.6 than f/8.

 

The pixel size on the 350D is approximately 6.4 microns, and the sharpness starts to significantly fall between f/5.6 and f/8, which makes it a good numerical match.

 

In fact, based on the numbers at photozone for f/8 and f/11 of a few good lenses, it's even possible to somewhat quantify how many pixels are "wasted" because of diffraction: the loss seems to be approximately 15% per stop for the first two stops.

 

Going back to the original question, the 10MP A640 isn't likely to take sharper pictures than an 8.5MP camera when used at the wide end, and sharper than a 7MP camera at the long end - and that assumes that the lens is diffraction-limited wide-open, which is probably quite far from the truth.

 

That's not highly scientific data, so you should take it with a grain of salt, but I believe that it's a good rule of thumb.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...