Some people eat, sleep and chew gum, I do genealogy and write...

Saturday, November 26, 2011

Digitizing Photos -- How much resolution is enough?

Let's suppose that the cost of hard disk storage is not a concern and since I just bought another 3 Terabyte drive for about $100, I can make that assumption. Most of the concerns we have had over the years in digitizing photos involved file size. In fact, the JPEG (Joint Photographic Experts Group) standard was developed as a result of the need to conserve storage space. JPEG files are essentially compressed files. JPEG is a trade-off between image quality and image size. But getting a 3 T drive for $100 started me thinking. Why should I be concerned about file size? Shouldn't I be more concerned about the quality of the digitized image rather than living in the past when storage was limited and expensive? JPEG is not only a file compression method but also ends up being "lossy" as in loss of detail and quality. If you want to see what is meant by lossy, go to Wikipedia:JPEG

Most of the literature on digitization talks in terms of dots per inch (DPI) which no longer has any real meaning outside of laser printing and similar applications. A more descriptive measurement is pixels per inch (PPI) but this measurement is for images expressed in pixels. Technically, resolution or sharpness of an image is characterized by a parameter called Modulation Transfer Function (MTF) also known as spatial frequency response which is the response of an optical system to varying intensities of light. This becomes very technical very fast, but is a way to measure the resolution of lenses and other optical devices such as scanners based on the resolution of a standard scale of lines per inch or lines per millimeter.

I am really talking about two different things. One is the resolution of a digital camera and the other is the resolution of a scanned image of developed and printed photograph.  If you are taking a digital photograph it makes sense to capture all of the information you can from your camera. In photography, the easy answer is saving everything in Camera RAW.  Raw files are so named because they are not yet processed and therefore are not ready to be printed or edited with a bitmap graphics editor. With Camera RAW (not an acronym) you get the maximum pixel count of the camera because a RAW image is everything captured by the camera. End of story. But here is the catch, not all cameras save images in Camera RAW, most cameras default to JPEGs. The quality of the image you get with a digital camera depends entirely on the quality and the resolution of the camera's sensor in pixels (Megapixels) and the quality of your lenses. If you want better pictures, you buy a better camera and lens. For example, a Canon 28-70mm f2.8L lens is very highly rated but the lens alone costs around $1,300.

The easy answer for scanning existing images is to scan everything into TIFF (Tagged Image File Format) files. TIFF files are uncompressed and lossless and most scanning software will allow you to save your scans in TIFF format. In either case, the file sizes are considerably larger than the equivalent JPEG file. The main file size concerns today come from sending files over the Internet or posting images online. In either case, it is a rather simple matter to re-save the larger image as a JPEG and use the smaller files size for online sharing. For example, I have a Camera RAW image on my computer that is 22.5 MB in size, if I develop the image, so to speak, and save the same image as a TIFF file, the size of the image jumps to 63.1 MB. But if I save the same file as a JPEG, the image drops to 5.3 MB. That isn't magic, there has to be a loss of information from the file to make that big of a difference.

Why should I care? I have a problem with the idea of losing data. I can't go back and take a picture over again and I don't want to scan pictures again. With photos from my camera, I can go back to the same place or I can gather the same people but the picture can never be duplicated. Children grow too fast, people die, people move away.  How may of you wish that you had a better picture of your grandmother or someone else in your family? Once the data is lost, it is lost. So I always have an incentive to purchase as high a quality camera and lens system as I can afford.

Scanning introduces a whole different set of concerns. Following my thought line about photography, shouldn't you just scan everything at the maximum level your scanner can handle? Well, not really. A photograph already has a limited amount of detail i.e. information. Scanning at higher and higher resolutions may not always produce a higher quality image. Wait a minute. Isn't that a contradiction of what I have already said? Yes and no. Think about it. Once a photograph is printed in whatever format, The amount of information in the picture is already established. Although most modern photographs had a reasonably small grain size, you can enlarge any photograph to the point where the individual grains of the photo are visible.

There is another limitation, the ability of the human eye to resolve details. Higher and higher resolution may exceed the ability of the eye to see any difference. The human eye can resolve only about 6 to 8 lines per mm from about 10 inches.

So there has to be some point at which scanning an image from a photograph reaches a practical maximum level. Would you be surprised to know that there are also physical limits caused by the interference of the light waves also known as diffraction. See Resolution and MTF curves in film and lenses by Norman Koren and Image Detail, How much detail can you capture and scan? by R. N. Clark. Diffraction limits the amount of detail any imaging system can obtain due to the nature of light waves. There are two systems for measuring the diffraction limit of an optical system; the Dawes Resolution Criterion and the Rayleigh's Resolution Criterion. Look these up if you want to know more.

What does all this boil down to? A sharp photographic image will contain about 400 pixels per inch although you might go as high as 600 ppi.  You can scan at a higher resolution but there will be little, if any, increase in the resolution of the final image. As a final note, scanning printed or handwritten documents has an entirely different set of criteria.






1 comment:

  1. Great piece. I've wondered about this, particularly as it relates to scanning. Thanks for laying it out without burying me in acronyms or too much technical detail.

    ReplyDelete