I get perfectly acceptable 16x20's with my 5mp camera using PSE's bicubic interpolation.
Olan Mills, the portrait studio uses the same camera for their 24x30's.

RonkII, being new to photography I don't know what PSE's bicubic interpolation is. Can you please put in layman's terms for me? Are you speaking of PhotoShop software capabilities?

Yes!
Shutter Bug even had an article on how one particular program puts an end to the mega pixel conundrum.
I was taught to shoot a bit loose and crop on the easel, but the basic compositional elements are on the film, now, media card. Cropping is necessary to allow for a particular print size to be rendered from a particular image. I have never believed that cropping makes up for poor vision at the taking stage.
Using a zoom lens is not the same as walking closer to the subject; it allows only for refinements to the composition. Kodak's "Pocket Guide to 35mm Photography" shows the difference pictorially.

Digital camera bodies become obsolete much quicker than film cameras, as sensors and the camera's processing engines get better and better. Lenses are forever. Buying a great lens is for the long term. My superb manual lenses - acquired more than two decades back - are serving on my new dSLR with distinction.

It is very difficult to evaluate image quality numerically, though it can be done to some degree with MTF curves. Begin talking about MTF curves and a customer's eyes will glaze over - mine included. Megapixels don't mean a whole lot, but they are at least easy to understand and ten sounds like a lot more than six so most people would think ten is 40% better. It is, but only if everything else is better too. In the case of consumer cameras, tiny sensors jammed with 10MP of photosites (individual pixel sensors within the main sensor) may produce far worse quality than one with only 5MP where spacing between the photosites is not so tight. Cramped sensors tend to generate a lot more noise.

Comparing an 8MP dSLR with a 10MP dSLR with the same size sensor, the 2MP difference is nearly irrelevant. However, it does help marketing departments sell cameras.

Actually 10mp is not 40% better than 6mp either, because 10mp is width*height, for example a 10mp D80 or K10D has a resolution of 3872 x 2592, where a 6mp D70 or K100 has a resolution of 3008 x 2000, so you are comparing 3872 pixels with 3008 pixels or 2592 pixels with 2000 pixels, so not that much of a difference after all.

The other important factor is that the photosites (the bits that light hits to take the photo) on the 6mp camera will be that small bit bigger so more light will hit each photosite, so less noise at the same iso. This also helps compensate.

So get yourself some nice glass and console yourself with the knowledge that whatever camera you buy it will be obsolete in a few years where the glass will still be good in 20.

Actually there is a significant difference depending on what you want to do with the file. The (approximate) one third larger pixel dimensions puts you a lot closer to having 35mm slides made from your digital files, one of the main purposes I use a DSLR for when shooting artwork for clients.

But on the other hand, I can earn a living selling fine art photographs taken with a two megapixel camera.

It's all in understanding the limitations of your equipment and making things work.

Given a choice in today's world, I would never purchase a six megapixel DSLR. Ten is now entry level. Six won't give you as many pixels to work with and won't have much resale value when you decide it's not enough.

Quote:
RonkII, being new to photography I don't know what PSE's bicubic interpolation is. Can you please put in layman's terms for me? Are you speaking of PhotoShop software capabilities?

FYI (For Your Information), PSE above most likely referred to PhotoShop Elements (the entry-level version of Photoshop).

Interpolation generally refers to the process of adding extra dots between the existing dots in order to get a higher total number of pixels (resolution).

Bicubic (don't know the technical details of the term) is a specific method of adding in additional pixels that theoretically gives a slightly better image quality than other interpolation options, such as (in GIMP [the Gnu Image Manipulation Program], a free alternative to Photoshop with similar features): Linear, Cubic, Sinc (Lanczos3).

For example, one could theoretically take a 800x600 image and interpolate it into a 1600x1200 image by adding an intermediate pixel between every original pixel. To do that, it has to figure out what color value to make the intermediate pixel(s). The different types of interpolation calculate the precise number of pixels and color/shade/tint/saturation of colors to put into the final larger image, in an effort to not degrade color quality, etc. in the process of making the image physically larger (double the pixels). It may also do a certain amount of smoothing, or texturing so that blowing up the image doesn't cause "jaggies" to show up at the larger size (or moire to be too apparent when shrinking certain images with repeating/intersecting patterns like fences, brickwork, or textured cloth).

Hope that helps, and is relatively close to accurate.

Cheers,
~Michael Gmirkin