Scott #63 Color Variations

I have found that scans of stamps while very useful at 1200 dpi for plating purposes are not accurate for color. A good example is that Sinclair's two scans of the same stamps are materially different.

If we could include a RGB value taken directly from the stamp, we could enter that value into the site that returns a color image that is more representative of the actual color.

Hi Njs900,
Your discovery sounds interesting and I think it is worth pursuing; by avoiding dealing with images you have effectively removed many of the color challenges that face hobbyists. This only a handful of variables that you will now have to deal with including ambient lightning and time.

As you probably know, colors reflect differently under various ambient lighting wavelengths. So you will have to define and control the ambient lighting so that it does not affect the reading you get. Whatever the ambient lighting may be, the key is consistency from reading to reading. Note that this consistency includes shadows; so be careful with overhead lightning and the way you are positioned when you take a reading.

Time is tricky, stamp colors are a combination of the paper, the ink, and any chemical changes that the environment has exposed the stamp to. So your readings will be for what the color is today and may not be close to the color it was 20 years ago or what the color might be might be in 20 years from now.

Lastly, the 'sample size' of the device you are using to get the color sample is important to know. How wide is the sample? Is it looking at 1mm, 0.1mm, 0.01mm, 0.001, or ???? of the face of the stamp? The larger the aperture the more the device must be averaging the actual array of colors it has detected. A smaller aperture means that moving the device reading just a fraction will result in a different RGB color (since the color varies across the face of the stamp).

If you can resolve the challenges defined above then you are getting very close to being able to say, "On this day, under XX ambient lighting conditions, and taking a reading from this part of the stamp, the RGB color is 255,192,76.

It would be huge to be able to detect how the ink chemistry and color changes over time and environmental conditions.
Don


Edit:
To illustrate how important ambient light can be, consider the phenomenon of metamerism(word of the day). Metamerism is when two colors that are not actually the same appear the same under certain lighting conditions. This is illustrated below

I will post sample size, light source, device and program details hopefully tonight so others can experiment as well and begin to compare notes.

One of the problems is that stamps fade, paper yellows, atmospheric elements (and non-archival mounts, acidic paper, etc.) affect the appearance over time. I've always thought a non-destructive analytical test (near IR, for example) that differentiates based on an elemental/compositional basis was more likely to yield conclusive results than a color measurement approach.

BTW, this is the reason the approach probably wouldn't work at all for unstable colors like 3c pinks, rose pinks, pigeon blood pinks, etc. What's pigeon blood one year can be common rose the next if left out in the sun a little too long... But compositionally it might still be possible to differentiate them...

And that brings up another question - should ID be based on what it was when issued, or what it is today?

Permit me to say that NBCo. used two kinds of presses, hand operated and steam power (assist). The steam presses were intended solely for the 1c and 3c values, while the other six values were done on hand presses.

At times the 1c stamp exhibits plate wear as a result of higher pressures encountered in steam-assist presses. Such worn impressions have a distinctively pale shade going towards the blue-grey end of the spectrum.

I also believe that on occasion hand presses were used for the 1c. In these cases the stamps have a warmer shade of blue (63b) and distinctly different impression quality where the fine cross-hatching around the medallion is not as crisp or wide.

Two more "milky blue" stamps.



They are over inked and one has a New York cancel.

The closest match seems to be 45'' or 45''' in Ridgeway. A BG-B hue with about 60% to 75% gray. The gray comes from adding white to the very black underlying pigment.

Douglas, you and I seem to have very similar interests. You can reach me at *** Private information removed by Staff. Please do not post your address/email/phone number etc. *** if you would like to compare notes.

Dark Blue -- The actual stamps are darker than this image. The bottom two have PF certs.




The principal components used to print #63's are Prussian Blue pigment with added white.
(see Brittain, 65 Chronicle 239)

Prussian Blue itself is quite dark and is described why Ridgway as 12.5% color and 87.5% black. It most often has a green undertone stemming from not a difference in the chemistry of the compound but from variations in particle size, PH value for other differences in the manufacturing process.

The Dark Blues are pigment with very little white.

Don, Please feel free to correct any of the above.

FYI...Posting email addresses is not allowed**, but folks can message you now using the forum email system.
Don

** When a forum allows email addresses to be posted it attracts automated spiders and bots to crawl the forum harvesting them for spam purposes.

Good day all, I see this was an old conversation, but very educational. I recently acquired and shared a page of what I thought was all 63s, I was told that on the bottom two rows there are three that stand out and might be 63a. after reading this I'm sure I need to do more research, but regardless of what my eyes tell me. a certificate will be the defended the factor at the end of the day. I still wanted to share the page, and 2400dpi scans I took of the three mentioned to me. along with two that are one covers. any and all information, either on identifying, or a few that look amazing for grading would be much appreciated. and thabk you all

Hard to tell from a scan, but I do not think you have an ultramarine 63a in that lot - they look like pale blue 63s. A $30 gamble on a cert for a $1600 might be the way to go. I've uploaded this before in another thread, but check this out to see if it helps. My stamp labeled "ultramarine" is a PSAG-certified 63a.

This is one of my favorite stamps - and you have some pretty examples!

@PhilaZilla - great page. Thank you for sharing it.

Mine too.
I am glad to have stumbled on to this discussion. Very enlightening info on all the different shades of blue. Until now, I thought there was only regular blue and dark blue. And had always hoped this first one was a dark blue. But, now more than ever, I am convinced it is not.




- chis

Here is one that I'm not sure of. The detail is amazing and the color is definitely dark dark blue.