Certainly a newbie question: If you shoot a lower resolution image (1024x768) on a high resolution camera (Olympus c-2000z) should the picture be any better than one shot with a camera with a maximum resolution of 1024x768? Set aside, for the moment, whether the lens and color system on the higher res camera would make a significant difference.

Thanks

-- Michael Drexler (drexler@earthlink.net), June 25, 1999

Answers

Michael, there are too many variables to be able to give an accurate answer to that. Beyond the obvious differences like the lens and color that you mention, are issues of the algorithms that the manufacturer uses to massage the ccd data. Raw data from the detector is not just stored in memory. That information will have heavy manipulation done on it, even before any compression is done to store it as a jpeg or other format. Even if stored as a tiff that data has still been altered before being stored.

Don't try to read too much into resolution figures. They can be used as an indicator of picture quality, but not as the last word.

-- Steve (tuna-boat-captain@ibm.net), June 26, 1999.

CAUTION! The Digital Photographer General has determined that the reply below can be hazardous to your health and may cause headaches or nausea! :-)

I'm with Steve, that's a tough one! This question got me thinking a bit.(machismo for quite a bit of thought actually...) I think because of the way the ccd's in most digicams are striped with filters that taking a lower resolution picture on a higher resolution digicam might yield a very slightly better picture because the layout of the colored elements in each pixel is the same. WHAT??? Ok, let me explain what I THINK:

In a typical digicam the ccd elements are covered by colored filters to make them responsive to one shade of light: typically red, green, and blue. The layout for a 3x3 pixel array usually looks something like: (I hope this lines up correcetly:

RG RG GB GB

RG RG GB GB

Now when you look at that you probably notice two things right off:

1. Hey, there are more green elements than red and blue ones in each pixel.

2. HEY! I thought you said that was a 3x3 (9)pixel array! I only see 2x2 (4)pixels, are those camera guys cheating me again?

Well, yes there are more green pixels, probably necessary to gather a balanced amount of light, since the human eye is more responsive to yellow-green light. Darwinian adaption to a yellow sun, don'tcha know? Good guess, or does someone have a better reason? :-)

And Yes, Virginia there is a Santa Claus. Go back up and look at that 3x3 pixel array again. You'll probably think the pixels look like this:

RG RG GB GB

RG RG GB GB

When it really gets interpreted by the camera like:

RG-> GR <-RG GB-> BG <-GB || || || VV VV VV GB-> BG <-GB RG-> GR <-RG ^^ ^^ ^^ || || || RG-> GR <-RG GB-> BG <-GB

Notice how the camera "borrows" the inside elements from each pixel to create a "virtual" pixel between each "physical" pixel.

Voila! 9 pixels from what appeared to be only 4! Neat trick huh? Well, it does work, despite how many of you purists out there are probably still grumbling about getting ripped off! :-)

My thinking is that since there are enough pixels in the array to form a "digitally unzoomed" 640x480 array of pixels(without having to steal the "inside" two elements of each pixel to form another virtual pixel that exists between each "physical" pixel), and since the Red and Blue elements are oriented the same way in each, that a lower resolution picture taken on a higher resolution camera might look just a bit better than the same size image captured on a lower res. camera.

Also, I wouldn't recommend a bad camera just because it has a higher resolution, but if it has all the features you want and the quality is also there the higher resolution images will generally let you do more with them before they get grainy or blocky(blowing them up). Also beware of camera's that use interpolated resolution in their specs. Always look for optical or image sensor resolution.

Anybody else want to take a swing at this one? It's kind of interesting and partly conjecture on my behalf. I think I'm going to get an aspirin now, and maybe go take some test shots to see how this actually compares. :-)

-- Gerald M. Payne (gmp@francorp.francomm.com), June 26, 1999.

Well, it obviously didn't line up to well! What's with the alignment software for this forum anyhow? Is it just trying to remove blank lines?

Well anyhow, I'll try again, here's the first illustration:

RG RG

GB GB

RG RG

GB GB

here's the second:

RG-> GR <-RG

GB-> BG <-GB

|| || ||

VV VV VV

GB-> BG <-GB

RG-> GR <-RG

^^ ^^ ^^

|| || ||

RG-> GR <-RG

GB-> BG <-GB

the idea being that you get virtual pixels by combining the values of the inside strips of adjacent pixels.

I hope this looks better.

-- Gerald M. Payne (gmp@francorp.francomm.com), June 26, 1999.