Scanning Techniques...?

That would be really cool if there was a program for the public to do color ID from a scanner.

Wert,
Let's look at this issue without any technology for a minute. You and I agree to identify the color of the same stamp and we will both use the Pantone COLOR BRIDGE® Coated chart. (http://www.pantone.com/pages/produc...id=1000&ca=1). We each go out and buy a fresh, new one (new so that we can be sure that have not aged and color changed over time). You match the color you see with a specific shade and then send me the stamp. I then match it to what I feel is the closest match. Chances are fairly good that we will both be on the same color chart 'page' but chances that we both select the exact same shade are not that high. First, lighting makes a huge difference is matching colors. And then of course our color vision might not be the same.

So how can we sync up and eliminate these variables? Pantone has a answer in its CAPSURE™ product (http://www.pantone.com/pages/produc...id=1032&ca=1). This device removes these variables. We now have no dependencies upon lighting, human eyes, graphics controllers, monitors, or software. The device is embedded, as long as we both use the same device we can match a color correctly each time. (Even if you have your connected to a Linux PC and I have one connected to a Window tablet.)

Moving on to your idea; I have little knowledge of which manufacturer the various local paint companies use and how they may differ between one another. Color comparator equipment is manufactured by many companies (Thomas Scientific, Fisher Scientific, etc.). But IF you local paint store and my local paint store have the exact same color comparators, then yes your idea would work.
don

I believe there is free software available on Sourceforge.net that will generate an ICC color profile but the key to doing this is that you start with a "known" color reference chart. You would scan this, then calibrate your monitor and printer to match to the correct shade. Pantone also has some free software tools to help match to their color standards.

And lastly, check with your scanner manufacturer to see if they have any calibration tools.
don

There have been many threads on SCF about scanners, scanning and
scanning techniques in the past including these two

https://goscf.com/t/10659&whichpage...ing,problems

https://goscf.com/t/30390&whichpage=1

but it's always good to find out new ways to make better scans.

Wert, I've tried most of the stuff you mentioned and basically
IMO the best results come with the stamp directly on glass
and lid open.

Only problem is that mint stamps curl and therefore
this causes the scan to have grey areas where the stamp
is not pressed flat to the glass.
Also I found that the heat from the scanner light makes them curl even more.
Plus its hard to keep a bunch of stamps lined up perfectly on the glass
compared to placing them straight in a Hagner.
And yes I know about the Picassa straightening tool,
which I use sometimes but not often.

The most important thing I believe is the type/make
of scanner.
IMO Epson produces the best scanners, especially the V500
V550, V600.
Scans from those are about as true to real colour as can be plus
one can see the actual texture of the stamp paper and on engraved
stamps the ridges jump right out at you.

Perfect example are scans from nethryk and revenuecollector
both use a Epson V500 I believe.

Overall, the third one looks best.

I don't think it is possible to make statements about scanning results because of the many variables mentioned already here: You scan your stamp with your preferred scanner and view it on your monitor, which may not be properly calibrated, and tweak it until the final result looks, to your eye, to be an accurate representation of the stamp's colour. Then you send it to someone else to look at, who may be using a monitor that is accurately calibrated (or worse than yours), using software operating with a different colour space or gamma setting. What's critical, and has been said here by others, is to have a system that removes all those variables. Sometimes when I scan things I will place the closest colour chip from my SG colour chart beside the stamp, so another viewer can compare how that "known" colour looks on his or her monitor and compared to his or her SG colour chart. As for software, I use a dozen or more apps for photo work (not Photoshop as it is overkill for me and too expensive) and there are many more that will easily handle these basic functions, and some are free. You only have to look at ebay listings for the same stamp to see how much variation there can be - particularly frustrating when you are looking for an elusive shade.

quigngt the only trouble is the stamp looks fantastic but is not the true colour/shade or darkness of the actual stamp...Look at the example below...You would say WOW Wert, you stamp looks great, but the true colour is the one with the lid closed and not open..That is deceiving to some people, but make it look good..So as 51studebaker says, I would use enhancing software to lighten it up, but still it is deceiving..

LID OPEN


LID CLOSED

Wert,
Your scanner CCD array detects the light that is emitted from the scanner lamp which is reflected off the stamp being scanned. But the real delta between the 'lid open' and 'lid closed' state is found in the scanner algorithm software. It is in this software that your scanner will try to average the light levels (including ambient light) in its calculations.

So it would be correct to say that the scanner will produce the most accurate scans with the lid closed. This make sense if you think about it, a scanner manufactured cannot control ambient light. This is why they bother with the extra cost of putting 'lids' on scanners.

And of course be aware that the light source of your scanner also changes over time, the bulbs will lose intensity. This is why the manufacturer recommends that you calibrate every few months.

So much of the imaging we all do comes down to the software. This point cannot be made enough when we start talking about details (subtle color shades, small hairlines, faults, etc.); if we are relying upon images it is important to know how the image has been generated. The more 'layers' of software that have been applied to an image, the more uncertain our opinion about the actual stamp should become.

But if the desire is to simply generate the 'prettiest' stamp image possible, throw whatever imaging filters you have at it and see how good you can make it!
don

True...Very true..My scanner is older, and that why I rely on software to produce a product that allows me to do analysis/flyspeck or whatever else the stamp commands me to do to get the information I am looking for...The lid closed scan is closer to the actual colour/shade of the stamp as it sits in front of me...Not pretty, but to me software is to anaysis, not to produce a pretty product..I never sell stamps, but if I did, I could well imagine I would use all available techniques to enhance my stamp for sale..I dont like that though...I should really invest in a new scanner.

Probably a naïve question, but would a digital microscope have the same color issues? I think they use a CCD array too.

Dan

danstamps54...A USB microscope is built I would imagine to the same standards as the next one..Surrounding light in your house could effect it...Removing the front plastic covering affects it as I found out and of course the settings in the software program you use..Contrast, tint, etc...All these things make your picture capture and my picture capture have different results..

What is missing from all this is are the 'black balance' and 'white balance' algorithms.

You enshroud the sensor in 'complete' darkness - in a camera, you might just place your hand over the lens - and the signal (well, noise, really) from each pixel is, to coin a phrase, The New Black (digital zero).

Next, you show the sensor a target you know to be white and, from that point, the signal in every pixel in every image is weighted to that pixel's version of white.

The problem is that, in any consumer-level scanner, this is all done once, by a robot, at the factory ... if it is done at all.

Even if this sort of end-to-end testing is done, and done correctly, back at the factory - they might just decide to trust the sensor & the source out-of-the-box, or test them separately, before assembly - we know that the sensor ages, the source ages, etc, so the results will vary over time.

Unless, of course, there is some sort of dynamic white balancing built into your scanner. There might be an algorithm that looks at the nice white background, attached to the lid, to use as a white reference. (That would explain the funny color variations when scanning against a white vs black background, or when the white background is removed by opening the lid.) Alternately, there might be an optical setup that diverts a small part of the source illumination to a tiny reference target deep inside the bowels of your U$D 120 scanner, allowing for continuous white balancing.

Or, most likely, every scanner gives a different color result ... and that is *before* any image processing algorithm is applied by whatever software you run ... and that is before this is all badly skewed by the type of monitor, and its age, and its quality, on which the image is viewed.

The digital color business is a tough business.

Cheers,

/s/ ikeyPikey

ikeyPikey...The only trouble is you and me and everyone else has a different scanner..Big problem..Not only that, when you and me (and everyone else) sets the white balance, etc. etc....We do not set our scanners exactly the same....Guess we all need to buy the same scanners to even hope we can get close to conformity...BUT THERE IS OUR MONITORS...

I'd like to assist an Australian military history author, on Australians involved in the Spanish Civil War.
I have an expensive Philatelic catalogue of the Civil War, with a suspect glued spine, which I do not wish to risk on the flat scanner platen.

Anyone have any suggestions? are hand held scanners suitable? expensive?
Thanks

Rod- A camera with close-up focus enabled?