Finding a scanner's "sweet spot"

[G. Couch]

Most consumer grade scanners have what is known as a "sweet spot"- an area where the scans will be at their sharpest, and with the least amount of noise. For most scanners this a narrow area running down the middle of the bed, but it can vary greatly from scanner to scanner. Higher end scanners will generally have much larger sweet spots.

A simple way to find your scanner's sweet spot can be found here

Basically, you place a sheet of paper in your scanner and scan the entire bed at 72-100 dpi. In Photoshop, use the Equalize command to exaggerate the differences in the scan and the resulting image will give you a good idea of where your scanner's sweet spot is!

Here is the result for my scanner, an Epson 2450...pretty much the entire upper half of the bed is the sweet spot.

[Ed_L]

Aha! A very useable post. Thanks Greg.

Ed

[DJ Dubovsky]

Very handy info Greg.
DJ

[DannyRaphael]

Greg...

I just tried this out and was dismayed at how "unsweet" my scanner is... Then I had a glimmer of hope.

I actually cleaned the glass and tried again. To my relief (and edification), now I've got a sweet spot too! This will change how I scan from now on. Excellent tip.

Danny

[jeaniesa]

Thanks Greg. I'll have to add this to my list of "installation steps" when I get my new scanner!
Jeanie

[Ed_L]

Just FYI, I scanned a piece of cheap inkjet paper. The scan looked absolutely terrible after using equalize on it. Another scan was done with matte heavyweight paper. This scan was much better. But I did notice that if I took the paper out of the scanner, then inserted it for a re-scan, it looked different each time I did that. Is it possible that very small dust particles could be moved around each time I did that? The glass was cleaned before scanning the first time.

Ed

[Steve Taylor]

I'm curious as to why there is a need to put a piece of paper in there at all.

[G. Couch]

Ed - Do not worry too much about dust. You are basically looking for large areas of dark and light. I get slightly different results with each scan but the lower part of the bed is always darkest as is the extreme left hand side. Try a few different scans and make note of the areas that always remain lighter...this will correspond to the sweet spot.


Steve - I think the only reason for the paper is because of the glossy surface of most scanner lids. I tried mine with just the lid down with no paper, and the reflected light caused a great deal of distortion.

[PixelMover]

Ok..I scanned the paper (full scanner bed size) @ 78dpi, tried every filter, gimmick, tweak and mode I could find in PS (including equalize, of course), but all I came up with was a completely evenly colored image, ranging from pale grey to bright green, but no dark/light differences.. Did I miss something?

[G. Couch]

Maybe you have a very high end scanner?

[jeaniesa]

Hmmm - given that you just posted a question about Linux drivers for a Scitex, I think that perhaps you DO have a high-end scanner!
Jeanie

[Doug Nelson]

Since the flatbed scanner sensors that read the top of a page are the exact same sensors that read the bottom, I wonder why scanners have a 'sweet spot'?

[Blacknight]

lack of equal quality throughout in the glass??

[G. Couch]

Originally posted by Doug Nelson
Since the flatbed scanner sensors that read the top of a page are the exact same sensors that read the bottom, I wonder why scanners have a 'sweet spot'?

Good question! I have heard several people talk about scanners having a "sweet spot", but no one ever really explained why. None of the seb sites I found mentioned much more than "scanner inconsistencies". I guess lower end CCDs tend to degrade a bit toward the edges? I'll see if I can track down some more info.

[BigAl]

Since the flatbed scanner sensors that read the top of a page are the exact same sensors that read the bottom, I wonder why scanners have a 'sweet spot'?

Light intensity decays as a 1/r^2 function (where r is the distance between the source and the detector), so it is easy to see that any irregularities or uneveness in the track that the source/detectors follow can have a marked effect on the amount of light reaching the detectors.

This can be clearly seen in the dark backgrounds of some of the pics in the "Scanning Objects" thread.

[PixelMover]

Originally posted by jeaniesa
Hmmm - given that you just posted a question about Linux drivers for a Scitex, I think that perhaps you DO have a high-end scanner!
Jeanie

My flatbed scanner is a Heidelberg/Linotype Hell Jade. Ok, ppl see this as a high-end scanner, but this is actually the cheapest, most basic one they have (600x1200dpi, 24bits SCSI). That's why I was interested to try this 'sweet spot' thing. I thought some parts of the scanner gave better results than others and thought this was a good way to check it...
Got it secondhand for $200 including transparency adapter. (At the dutch equivalent of E-bay).

As for the Scitex, I swapped it for a used Nikon F4 and lens

Hmm...anyone here who wants to swap a Sony 24" TFT screen for a 1986 Honda Civic? Only 128.000km !!

[Stephen M]

As noted, different substrates may show different levels when equalized. Perhaps try different types of white paper, as well as black. Matte or less glossy may work better.

Two good links on consumer scanning can be found here:

http://digitaldog.net/files/Scanningtutorial.pdf

http://www.thelawlers.com/FTP/B&W%20Scanning.pdf.sit

The fine detail test suggested by Andrew Rodney using currency is a good approach too (legalities though). GAFT and other print based organisations would have dedicated scanning targets, for both colour/tone and detail.

A Umax 'prepress' scanner I used to use had a 'zero' point marked on the ruler at the front of the scanner - this spot was where the manual suggested the best focus was obtained, tests also showed this...so tranny scanning was done one at a time, while reflective was done with more originals and batching so that more productivity could be gained.

This sweet spot issue can be critical when you are attempting to batch scan many originals. Some scanners let you set separate scan crop frames and apply separate endpoints/gradation etc to each scan frame and save the files off to disk or live to Photoshop for post processing. This is a huge productivity bonus, but if the flatbed has serious probems over the entire scanning area - quality may be sacrificed or the batch may be smaller than you would like if you try to stay within the better area.

I used to use a CreoScitex EverSmart Supreme high end flatbed (US40K) - which had patented XY positioning...the scanner head moved both vertically and horizontally to focus in on parts of each image, then it would stich the strips into one final scan. Sounds scary, but I never saw any issues from this technique. With this line of scanners you can use the full surface of the bed and the biggest plus is the batch scanning workflow. Prescan the whole bed, set your scan crops, adjust your first scan and then send to the scan qeue, adjust the next prescan crop while the first is scanning/writing to disk etc.

I currently work for a publisher who uses a drum - I am not sure that a high end flatbed could match the pace/productivity of a good drum and operator.

Now I just use a humble Microtek flatbed and SilverFast software - since the drum does most of the work (I am not a drum operator and scanning is not the focus of my job).

Scanning is such a deep area to go into - and any discussion involving scanning should include this link:

http://www.scantips.com

Have fun.

Sincerely,

Stephen Marsh.

[G. Couch]

I think BigAL is on the right track with his explanation. I did some more research and found this in one of my books, Scanning the Professional Way-

"The fluorescent lamps used in most flatbed scanners are susceptible to fall-off, a phenomenon that produces better illumination near the center and less at the outer edges of an original."

Obviously higher end devices have better light sources, higher quality CCDs, A/D converters, etc... I think electronic noise plays a part as well. Certain areas in the scanner are more susceptible than other areas to interference.

[WideAngle]

I am using a Microtek 6100pro with silverfast which I guess is sort of midrange (about 400CAN dollars). see the pic for the sweet spot after applying equalization, I assume an even gray is good? There is some noise on the edges though.
Wide angle

[kaulike]

Posted below are before and after shots of my Epson 3170. The first was by laying a sheet of heavyweight matte paper and hitting the button. The second was augmented by pressing down on the lid while scanning, as I noticed the first time that the dark spot was caused by a slight warp in the paper. I thought the difference was remarkable.

[raniday]

Hmmm, maybe scanners have improved since this thread was begun. I scanned a sheet of white paper on several settings and just got a completely white sheet of paper. Sweet.

[kaulike]

Try loading the result into Photoshop and use the Equalize command. I think you may be surprised at the results.

If you are still getting a completely blank document, make sure you don't have the scanner set to scan with a transparency adapter. *Any* scanner will have varying results across its scan area.

hope this helps!