zoom lens vs. megapixels

[LumenLux]

Would you expect a more clear photo by using high power zoom lens or computer blowup of high megapixel image of same subject area?

It has been a long time since I gave serious thought to a Kodak camera. However, I found myself intrigued by two new cameras by Kodak. The P850 is 5.1 megapixel, 12x optical zoom 36mm - 432mm (35 mm equivalent.) The P880 is 8.0 megapixel, 5.8x optical zoom 24mm - 140mm equivalent.

Assuming both lens are similar optical quality, consider the following. If I am maybe 25 meters from a moose, with which camera will I get the clearest, most detailed, 11x14 inch photo of the moose filling the entire frame?

I did not have a moose handy, so I made an inside-the-house, short range test. At 5 meters, I took photos with each camera at longest zoom length. I then selected the same portion of each photo and printed to fill a 4x6 inch print.

The segment of the 12x, 5 megapixel, was printed at 153% of native size. To make nearly so large a print of the same segment from the 5.8x, 8 megapixel, required me to print it at 315%. The quality of the 12x zoom print is far superior!

Another comparison: Using the same segment of each original photo, I printed the 5.8x zoom, 8 megapixel, at 100 % size. The print is 24 mm x 36 mm. I then printed the 12x zoom, 5.1 megapixel, at 100% size. The print is 58 mm x 87 mm. The area of the 12x zoom version thus is 5.8x as large.

Can someone more adept at the numbers tell me if this looks right? From what I observe, I am surprised how much better the optical zoom photos seem to look vs. the higher megapixel versions of the same subject segment.

And, if the above is all correct - how would the results be effected if the desired final print of the moose is to be 5x7 inches, 11x14, or 20x30 inches?

[centreofengland]

Hi, This sounds correct to me as you have explained it.

The important thing you have not mentioned in your very detailed description is how many pixels long and high was the image from each camera that you 'printed out', but I can tell by the fact that you have said the 12x zoom one is printed at 153% original and the 5.8x zoom at 315% original size that the 12x zoom image has more pixels in it.

As you can probably see by what you have said, you don't have as many pixels in the image from the 5.8x zoom camera so you have had to spread the pixels you do have across the paper more during printing (by 315%).

Therefor in this situation the 12x zoom is better as you can get 'closer to the subject' optically and get more pixels representing the image.

People seem to get confused by megapixels and think that you need lots and lots, this is not true. You can print stunning 12" photos from a 2.1mpix camera, you can not tell that its a digital print.

The MOST important thing is the quality of the lens and the optical zoom range, as you have seen for yourself having a higher megapixel camera has not helped because you can not get close enough to the subject (the moose). The P880 will give you 672 extra pixels in the length of your image (P880 is 3264 pixels width versus P850 with 2592 pixels width), but ONLY if you can fill the frame with the subject at the shooting time. Any cropping you do to an image in th ecomputer after will reduce the resolution.

My Olympus 5050 will produce great 18"x12" prints (5mpix) but they look best if you fill the frame so you have as many pixels as possible. I have just ordered a 30"x20" print online from one of my photos to see what thats like, looking forward to seeing it this week hopefully.

So to sum up, if you can not get close to a subject (wildlife) then the 12x zoom is worth a lot more to you than higher megapixel camera. If you CAN always get close to the subject then the higher megapixel camera is better.

Hope this helps explain, any questions let me know.

Andy

[Lin Evans]

There is an easy way to determine which is the better way to go, assuming similar quality pixels, lens, noise ratios, etc., which is called the Figure of Merit.

The answer depends on the resolution of the higher pixel count sensor and the degree of magnification of the lower pixel count sensor. Here's a link where you can get the formula, read a comprehensive explanation and simply plug in the variables.

http://www.mav-magazine.com/Jun2000/FigureofMerit/

Lin

[Alan Lyons]

Hi All,

here is a forumle from the past, The bigger the negitve, the more you can enlarge. I assume the same applies to the digital image. The more pixeles used to render an image the more detail will be available in the enlargement.

Alan

[LumenLux]

Thank you Andy, Lin, and Alan.

Andy, your confirmation and additional info is helpful. Alan, of course you are right - so far as it goes. Lin, I will pursue the link - sounds like a useful tool to make accurate comparisions of specific alternatives.

"Mr President, a follow-up question:"

If a user were to put an auxillary lens on the cameras - could one expect a wide-angle, or a tele, to degrade less? Specifically, the P880 has 24mm - 140 mm, and will accept an auxillary lens for additional zoom-in. The P850 is 36 mm - 432 mm, and will accept an add-on wide angle lens. It seems it would be better to add the wide-angle to the 12x zoom lens than to add the tele to the already wide-angle 24 mm ?

[Joe Webb]

The results you achieved also sound right to me: i.e. that the extra optical zoom gave better results than the 8 mp enlarged. Assuming that the camera is well held to eliminate any camera shake, I would always opt for the 12x optical zoom. For example if you take the 8mp image and enlarge the central 50% the resulting image will contain only 4mp (less than the P850). This is the same effect as using a digital zoom in-camera.

As to the "follow up" question: The first thing to remember is that these are lens adapters, this is not the same as changing to a different lens on an SLR. Lens adapters are usually of a lower optical quality than the original lens, so both will result in some loss of quality. However I would expect that the drop in quality would be much less than the extra enlargement needed to compete with a tele adapter.

Which would drop more quality, tele or wide? I don't know (!) but I would have a guess.... We know that the weak area of a lens is usually the outer edges, with lower edge definition and possible distortion. Given that the wide adapter would make more use of the edges, it seems logical the wide would show a greater drop in quality - unless someone knows better......

The best advice must surely be to use the original lens as much as possible only adding an adapter if you really need it to achieve the desired shot.

Best wishes,

Joe.

[Lin Evans]

Hi All,

here is a forumle from the past, The bigger the negitve, the more you can enlarge. I assume the same applies to the digital image. The more pixeles used to render an image the more detail will be available in the enlargement.

Alan

Hi Alan,

One would think that this would be a general "rule of thumb," and in most cases it probably is, but there are some caveats which must be considered.

If pixel count were a true measure of resolution this would be true, but though pixel count contributes to resolution there are numerous other issues which impact resolution. First one must be aware of just what is meant by the term "resolution." The amount of detail in a digital image is the result of the combination of pixel count, pixel quality, amount of noise in the image and lens quality. True resolution is usually measured by the number of line pairs per mm which can be discerned from the capture. A six megapixel capture from a good quality dSLR with a decent lens will generally produce more detail and thus higher resolution than a corresponding image from a seven megapixel fixed lens digicam. A 3.4 megapixel capture from a Sigma SD10 with a decent lens will produce slightly more detail and higher resolution than a 6 megapixel capture from a typical CFA (bayer processed) dSlR sensor. So even though the SD10 only produces a file of 3.4 megapixels the capture was made with 10.3 million photo detecting sites (pixels). The directly captured pixel level detail (separate sensors for RBG at a given site) and quality are cleaner and more accurate than the corresponding pixels produced by bayer interpolation from the larger file of a CFA type sensor.

The bottom line is that when comparing resolution and detail we need to be attuned to not only pixel count but pixel quality, lens quality and noise performance - all of which impact resolution.

Best regards,

Lin

[Alan Lyons]

Yeah, like I said, You only get out what you put in