Philoking.com

I have read many articles about how megapixels do not matter recently, the latest being on the New York Times so I thought it was time to clear the facts up. I am sure in certain situations, this statement is true, but let’s get to the brass tacks of why and how pure megapixels are hugely important.

Let me start this by saying that most of these articles are written with point and shoot consumer digital cameras in mind. There are a few things about these cameras that make resolution not quite as important, lens quality and sensor size. Let’s first talk about the sensors in these cameras. The sensor is the component that actually takes the photo. I am going to reference a photo.net article here to explain to you the importance and effects of sensor size. Bob Atkins takes in to account several cameras, all with a 3 megapixel pixel count. A standard 35mm film frame is 36mm x 24mm, that is considered full frame. There are only a few cameras, most notably the Canon 1Ds and Canon 5D that actually have full frame sensors, nearly all cameras in production, and I would venture to say all consumer cameras are vastly smaller than full frame. The following image shows the difference in the amount of real estate the camera has to capture image detail from the glass, no matter what the resolution.

So why is this important? Cameras with smaller sensors use shorter focal length lenses to get the same angular coverage as cameras with larger sensors do with longer focal length lenses. So, if you have a 28-105mm zoom on a Canon D30, a 10-37mm zoom on a Nikon 995 or a 7.4-28mm zoom on a Minolta X1, you get approximately the same shot. The smallest, the Xi is going to have to capture a whole lot of information on a very small sensor, using a pretty poor lens. The Coolpix is going to have a marginally better lens, and more room to capture a clean image. The D30 is going to have the largest sensor of the group and would be the cleanest lens to capture it’s image with.

So what does all this mean? It’s all about physics; the lens takes the outer world, and compresses it to an image small enough to be captured by the sensor. It bends the lightwaves to allow the camera to capture a much larger point of view. In this case, glass quality is vastly important. The D30 with a nice L series lens is going to perform this infinitely better, so a 3MP image from this camera is going to be incredibly clearer than a 3MP image from the Minolta. It’s just physics.

So when does pure pixel count matter? When you take a realistic test, of like situations. I have a Canon Rebel XT for example. If I were to set the camera at it’s different resolution settings, and shoot the same image 3 times, on a tripod, at 3 resolutions, 3MP, 5MP and 8.2MP, I would dare say the quality of even an 8×10 would be drastically different on each. The test from the NY Times article is flawed in that in every case, the image would have been digitally resized to fit that print size anyway. If you were at 3MP, either Photoshop, or your printer would have calculated the resize so it would be scaled to a “poster” size print. In the case of the 13MP camera, you can guarantee that if a consumer camera is boasting that type of resolution, the camera is natively resizing the image up itself. So what you have is cameras trying to print an image that is too large for them anyway, so whatever was used to print them, is resizing them to a poster size.

It could be true if…
A quality print is usually made at 300dpi or dots per inch. A poster us often printed at 150dpi or dots per inch. Lets say to be fair that the print was made at 16×20, a standard print size. At 150dpi, your camera would have to capture an image that is 2400 pixels wide. A 3MP Camera captures an image that is 2048×1536. If you had NO crop, and your poster was printed in landscape mode to get maximum resolution, you would STILL be short roughly 860 pixels. So at 3MP, your image would need considerable resizing. Now, if you had an 8MP camera, your image would only need about 100 pixels of image to make that print. So yes, in theory, if you were making a 16×20 print at 150dpi, and you captured a 3MP image and a 8MP image with the SAME camera, on the SAME lens, it would be a small difference between the two.

But when will it matter?
Lets look at this same scenario with a 16×20 at 300dpi, say a photo you actually wanted hung on your wall at home. We will be using the same situation, a Canon Rebel XT, same image, shot on a tripod at 3MP and 8MP. Your same 16×10 now requires 4800 x 6000 pixels to print a 1:1 print where there is no need for the software to resize your image. Now your 3MP camera is having to basically triple it’s pixels, literally guess and make up image for the missing pixels. The camera is now over 3200 pixels short of it’s target print size. In this scenario the 8MP camera is about 2300 pixels short. Still quite a bit of image to make up, but if you straight double the image, your really close. Due to the way image resizing works, you will find it resizes to even magnifications quite a bit better than odd ones.

The following chart shows how much resolution you need for standard print sizes at 300dpi.

As you can see, if you want true clear prints, the resolution needed is pretty large. I would bet larger than your camera. The real point to this post is that you cannot make a blanket statement like “Megapixels Don’t Matter.” Sure, you can create scenarios where that is true, but in reality, it’s not even close to accurate. In the world of prosumer and professional digital SLRs, quality lenses, and quality printers, it’s a very important issue.

When you are shopping for your camera, wether it be consumer or higher end, I would give you the following advice:

• Do your homework, look at sites like http://www.dpreview.com and make sure that the resolution you are getting is the actual resolution of the camera’s sensor, not the interpolated or software resized image the camera’s computer outputs.
• Go to sites like http://www.pbase.com and search by camera model. Look at real world images other people have produced with that actual camera in general conditions, not the studio shots that have been enhanced to sell it.
• Check out http://www.cnet.com, and read reviews that actual users have written about your camera (also check who online has the lowest price!)
• Don’t pick camera a just because camera b doesn’t have a free memory card or some photo application bundled with it. The camera itself is what matters, get the right camera and don’t be sold by some cheapie bundle.
• Stick with the brands that know cameras. There are a few exceptions, but I would say if it doesn’t say: Canon, Nikon, Pentax, Leica, Kodak, Sony, Fujifilm, Olympus or Minolta on it, I might look elsewhere.
• Forget “digital” zoom capability, go for real zoom. If it says 20x Digital Zoom, it’s unusable so why bother.

Ok, those are my digital camera tips. If you wonder if I have any experience on this subject, check out some of my photography at http://www.philoking.com/coppermine.

JB