Newsprint
Newsprint
Newsprint

Scanning FAQ

Where can I find in-depth knowledge about Photoshop?

A good all-around Photoshop book is:

Note: Revised editions of books carry new ISBN numbers. Bookstores can now search by name and author, so do not be alarmed if the ISBN is different.

The coverage of calibration, and the calibration images on the included CD, are worth the price of the book alone. If you have always wanted to be a Photoshop master, this is your starting place.

If you are new to scanning, or need a refresher course on such things as resolution, tonal correction, and halftone theory, a great book is:

Another good resource is:

While some of the techniques are a bit over the top, this remains a good, solid reference directly applicable to scanning for print. While we do not recommend every technique in the book, the author does try to pull out all the stops on some dreadful originals. Valuable for the emphasis on learning to think in channels, and understanding the relationships between channels.

Do my scans look bad because my monitor is not calibrated?

In a word, no. Check out our monitor Calibration FAQ.

More resolution is always good, right?

No. For halftone reproduction, twice the line screen equals more than adequate scan resolution (Note, this is influenced by enlargement or reduction). For Fricke-Parks Press, this will mean a final image resolution of between 200 and 300ppi, depending on what kind of paper your job will print on.Too high a scan resolution can make your image appear ‘soft’ (not in super-sharp focus). For complete coverage of how to calculate resolution, check out Real World Scanning and Halftones listed above.

Can I scan slides on my flatbed scanner?

Probably not. Some inexpensive scanners come with ‘slide adapters’ that should be dropped into the nearest trash bin. The problem is with the dynamic range of slide film, the range of values between the whitest white and the blackest black. It is very large, beyond the ability of most inexpensive scanners to ‘see’.
That’s the bad news. The good news is there are a new generation of dedicated slide scanners that produce scans good enough for much of the print market. Units from Polaroid and Nikon are available for a reasonable price.
While quality from these units is good, fine artists, photographers, and high end users will continue to use drum PMT scanners. More good news: the price of image capture for even medium format originals can be as low as $60, a real bargain (traditional color seps from a PMT drum scanner used to cost hundreds of dollars).
Of course, the inexpensive tip is to obtain an 8x10 print from the slide. You can then scan it on your scanner (and frame it afterwards). Be sure to avoid ‘studio’, or other prints with a texture on the paper... you want ‘glossy’.

I want to enlarge a portion of a photo, but it looks terrible.

Assuming you are scanning a regular photo print, there is a limit on how large you can go. Many scan gurus refuse to enlarge photo prints at all! This is a little extreme, and a few percentage points will not really hurt. Anything over about 150% will show defects. Attempting 300 or 400% is hopeless. The problem is not in your scanner, but in the print itself.

Color photo paper is not really a high resolution medium, at least compared to film. The solution is to have an enlarged print made, and scan it. This way, the enlargement is made using the full amount of image detail present on the negative or slide. Crop the area you want with post-it notes, and have an enlarged 4x5 print made. Scan it, and you are done.

”Scan an enlarged print” works for enlargements... can it be used to improve all my scans?

Possibly! The easiest way is when getting your film developed... always opt for the slightly more expensive larger print size. The cost of this is almost zero on a per image basis, and the larger the original, the better.

An other trick is to start with an original that is 200% of the final image size. Scan it at 100%, but at half the resolution you need. In Photoshop, use the ‘Image Size’ command (with ‘resample’ turned off) to reduce the image to the correct size (which also makes the resolution correct). Reducing the image in Photoshop means any artifacts or ‘fuzziness’ introduced by the scanner optics is reduced in size. The result is a sharper image. This is especially true with a scanner that uses ‘interpolated resolution’. See the next question for more details.

This trick of scanning and reducing will make it possible to have really sharp images, even from an inexpensive scanner.

I am buying a scanner, and I’m confused. What is scanner resolution?

Resolution is a measure of how many samples per inch the scanner is capable of. The scanner samples 300 times in an inch, for example, and creates a 300ppi scan. When marketing departments enter into the scene, things get more complicated. There are several kinds of resolution measurements made by manufacturers, and each measures something different.

The simplest, and most important, is optical resolution. This is how many samples the actual CCD (the scanner’s ‘eye’) is capable of. A common measure among bargain scanners is 300x300, and among better scanners, 600x600 or more. During a scan, the CCD is moved across the bed of the scanner. Take a CCD that is 600 ‘eyes’ wide, step it across the bed 600 times an inch, and you get a 600x600 scan.

There is a semi-misleading practice of claiming, for example, 1200x600 optical resolution for a scanner with a 600ppi CCD. Someone realized you could make the CCD take baby half-steps across the bed, and claim 1200x600 optical resolution. After all, it is taking samples 1200 times an inch, in one direction. The problem is that those samples are so large, they overlap! This does not give you 1200ppi of information. Now, it may or may not be any better than a straight 600x600 scan, but essentially it is a 600x600 scan.

As remarkable as it seems, there is an other measurement that is even more questionable, and that is ‘interpolated’ resolution. Since we are not studying calculus, for our purposes the word ‘interpolated’ can be thought of as meaning ‘made up’. The scanner takes a 600x600 scan, and guesses the values of all the pixels it takes to resample it to 1200x1200, for example. Heck, Photoshop can do that. Not only is interpolated resolution not good for photographs, best quality is usually obtained by scanning at the ‘lower’ optical resolution.

When comparing scanners, use the true optical resolution of the CCD.

What other factors affect the quality of a scanner?

Dynamic range: In actuality, there is only one aspect that concerns us, and that is how well the scanner can see in the dark parts of our original. This is usually rated as Dmax, which is a logarithmic measure (small differences in number equal large actual value differences). Expensive traditional PMT drum scanners excel in this area, commonly rated at Dmax 4+.

The most expensive CCD scanners (unit costing tens of thousands of dollars) might come in at 3.7 or so. Most desktop scanners are lucky if they can achieve 2.2.

Unfortunately, manufacturer’s claims for this are all over the board, and some are just not true. Many would have you believe a $150 scanner rates the same as a $200,000 drum machine. No way. The good news: most of this is really only needed for scanning slides. Photographic prints are usually well within the range of good desktop scanners.

Since the claims are unreliable, the situation is difficult for the comparison shopper. One note: manufacturer’s ratings usually are consistent within their own product lines. If one scanner is rated better than an other, it may well be better if both scanners are from the same manufacturer.

Mechanical ruggedness: This is usually overlooked by most people without experience with traditional optical equipment. The S2000 Linotype-Hell CCD flatbed scanner, for example, is the size of a clothes washing machine, and weighs hundreds of pounds. All that vibration-dampening mass helps produce sharp scans, and the rugged construction keeps the optics in good shape. Of course, it sold for tens of thousands of dollars new, but the same principles hold true for lesser machines as well.

I have seen desktop scanners so flimsy that when making a scan the vibration had the scanner jittering across the desk. If the CCD does not advance accurately, if focus is not stable across the scan bed, or if the design allows dirt onto the underside of the glass, what kind of scans are going to come off it?

This is a hard judgment to make. Always try to see an unboxed display model, working if possible. You can’t expect too much from a cheap scanner, but it should seem solidly constructed, heavy for its size. The old HP scanners were wonderfully solid.

Quality of the CCD/electronics: The bane of CCD scanners is noise in dark areas of the image. Appearing as multicolored ‘speckles’, it is especially damaging when scanning a dark image that needs to be lightened. Unfortunately, there is no manufacturer rating for this.

All CCD scanners will produce some noise when pushed to the limit of their Dmax. The only way to know how much noise you will get is by experimenting with a scanner. Try searching on the Web for the manufacturer of the scanner you are considering. Many private enthusiasts post test results to the Web. You should be able to find something about your potential purchase, and maybe even get sample images.

Drivers/scan interface: A good scanner can be marred by lousy software. Not so long ago, a major prepress industry manufacturer unveiled its new superscanner at a trade show. The software was so inadequate, by the end of the demo the factory reps were apologizing to the audience. Not even they could make it work. You don’t want to get stuck with bad software.

Reviews usually comment on the software as well, so do web research on this as well. Both HP and Microtek are usually well regarded for their TWAIN drivers. The more popular models even have individuals who post icc profiles to the Web.

What about color theory, and more helpful information?

Try these links to other resources:

Color tutorial from Adobe: