Questions relating to the correct ID of stamp colours?

[midas42]

I find it very difficult to identify stamp colours, and I am sure I am not alone. I have a SG colour guide card but a lot of colours are not on the card. I have included three stamps from the Bahamas and have no idea whether they are lilac, violet, rose lilac, or deep violet, pale or dull lilac etc.

I have also included a picture of two Denmark stamps, that are supposed to be brown. I have found this website that allows one to upload a stamp and they will then give a report on the percentage of the different colours on the stamp. However if you look closely you will see that parts of the top stamp are brown and parts are a dull green.

As technology progresses, I wonder if it would be a good idea to include an analysis of the colour of a stamp when cataloguing it, or perhaps give a colour analysis when describing a stamp. I can see that there could be a difficulty if the cancellation is overpowering, but maybe there could be some merit in giving a particular stamp a colour identity.

Some thoughts on this problem would be appreciated.

Alan Parry

[Gismo]

I can't help you with the Bahamas stamps, but the danish stamp with dull green colour is lighted by the sun or some chemicals.

/Michael

[Allanswood]

There is no way that a website can take a submitted scan and tell you the true colours of a stamp. They can tell you what they think are the colours of the scan, but not the stamp.

They even counted the cancellation as a colour.

[Global Administrator]

I have included three stamps from the Bahamas and have no idea whether they are lilac, violet, rose lilac, or deep violet, pale or dull lilac etc.

You have a SG Colour Guide, and you cannot sort these???

[chaddy]

I am colour blind, so shades are almost impossible for me to distinguish. I use 3 different colour guides for Commonwealth stamps.

1 - a very old guide which has what look like coloured stamps in. Sorry that I don't know the make but I have misplaced it at the minute. This is good for older stamps, because the ones on the guide have faded over time, just like a normal stamp. However, it doesn't have a vast range of colours.

2 - the Stanley Gibbons colour guide. This has a greater variety of colours but, because they are on plastic, are not as accurate as the previous guide.

3 - when all else fails, I ask for the wife's opinion. Obviously her word is final

[kenlondon]

Would it be possible to post a link to this colour-matching site please? It looks like a good concept.

I would have thought that the problem with the cancel being included in the colour analysis can easily be solved by getting the user to draw a selection-box around a sample area of the stamp which is of uniform colour. PC applications such as Thumbs Plus have a similar kind of method of sampling with a view to doing a copy and paste of a colour (see this colour x here, I want it rolled out to all parts of the image with this y colour), but I don't think you can produce a report that shows a breakdown of the constituent shades.

[Allanswood]

It can't work. It sounds "pretty" but... no.

Try sending them the same stamp scanned with a white background and then black, grey etc.
Then try the same stamp with 2 bright stamps beside it, then 2 dark beside it.

Then see what auto settings your scanner has to compensate and what auto brightness, contrast etc and see how that goes with someone elses scan of the same stamp etc etc etc.

And what about the DPI? Get in closer and closer and the scan colour will change for the same stamp.

And is your scanner certified as a true colour scanner calibrated to a known accepted standard?

Even looking at a few sample scans of the same stamp on your own screen will show the different shades.

Then of course the testing superiority needs to have a certified example of every known shade for the stamp you have to be able to say its "vermilion", or "bright red" or "scarlet blaze".


As a side point there already are machines that can analyse the ink on a stamp and then tell you the composition of the elements and pigments that make up that particular ink, so that can be compared to yours and it then doesn't matter about the cancel or age or a bit of fading.

[midas42]

Would it be possible to post a link to this colour-matching site please? It looks like a good concept.

Hi kenlondon

The web site is stampsmarter.com. A lot of interesting info on colours and lighting.

Regards, Alan Parry

[midas42]

Thanks to all who have replied to this post. I still think that as technology advances we will eventually be able to give a stamp a "colour ID" using the codes for red green blue. If I then scan my own stamp it should come up pretty close to the colour ID. Don't hold your breath; this might take some time!

Alan Parry

[kenlondon]

Alan: Thanks for the link.

I hear what's been said about calibration of scanner, etc., etc. Humour me for a while, but I would be interested to see what results others get, using the same mint stamp, using whatever scanner you have.

I've chosen GB Jamboree stamp as it is pretty common, and has just the one bright colour (yes I'm aware that it is unlikely to be "pure" red). These are the results I get using a Canon Lide 400 scanner.

My thoughts are that, much like a camera can be calibrated against a grey card, a scanner has the advantage of having its own little micro-environment, with a white background which the scanner should be able to calibrate itself against. Putting the stamp against that background should aid the scanner in that process.

Jamboree colour analysis

[midas42]

Alan: Thanks for the link.

I hear what's been said about calibration of scanner, etc., etc. Humour me for a while, but I would be interested to see what results others get, using the same mint stamp, using whatever scanner you have.

I've chosen GB Jamboree stamp as it is pretty common, and has just the one bright colour (yes I'm aware that it is unlikely to be "pure" red). These are the results I get using a Canon Lide 400 scanner.

My thoughts are that, much like a camera can be calibrated against a grey card, a scanner has the advantage of having its own little micro-environment, with a white background which the scanner should be able to calibrate itself against. Putting the stamp against that background should aid the scanner in that process.

Hi Kenlondon

I have done an experiment using your scan of your stamp. I have copied the first colour on your scan with RGB code c03048, which translates to decimal 192,48,72. I then inserted a square shape in a word document and set the colours of the shape to the decimal equivalents of your colour. I then copied and pasted your colour into word next to my shape and to my aged eye they appear to be equivalent. See screen clip.
My scanner is Canon 4400F with all settings set off, so there are no enhancements etc.

[kenlondon]

What we need to do next is ask chaddy's wife to give her opinion.

http://stampsmarter.com/features/Color_Extract.html

This is the link midas42 referred to if others want to try this out, select a file from your pc, and press the Process button.

As mentioned previously by Allanswood, this will be no good for stamps whose images are made up of lots of dots of different colours, unless using colour separation techniques to split them up... life is too short for that.

[Allanswood]

These are 12 scans of the same single strong red coloured stamp scanned with 3 different backgrounds and at 4 different DPI settings 300/600/1200/2400

No settings adjusted on my scanner.

So which one is the true to life colour?

And if you know you have the same stamp sourced from a different location and slightly more or less aged thus scanning a different shade of red - how would anyone know that they are a match when to those with good colour vision there is such a difference in the scanned results?

How many shades do others see here? But it's the same stamp.

Australian KGV 1d Red 12 scans same stamp

[kenlondon]

Why not put each of those 4 dpi settings that have no background (which the scanner can use as a calibrator), and produce a colour report for each setting?

Using non-default backgrounds may affect the scanners assessment of the overall scene, in a similar way that a camera responds to a portrait against different background shades. The scanner shouldn't do, but it does, because it should take into consideration the light source it uses to light the bed of the scanner - whose colour profile will alter subtley with age.

Time consuming, yes, but necessary to engage in moving this thread forward in an attempt to move away from subjectivity to objectivity (which is what people do when they use a colour guide or consult a catalogue).

There needs to be some quantification process which is superior to the Mk I eyeball, which we all know, can be heavily susceptible to variation.

[midas42]

Why not put each of those 4 dpi settings that have no background (which the scanner can use as a calibrator), and produce a colour report for each setting?

Using non-default backgrounds may affect the scanners assessment of the overall scene, in a similar way that a camera responds to a portrait against different background shades. The scanner shouldn't do, but it does, because it should take into consideration the light source it uses to light the bed of the scanner - whose colour profile will alter subtley with age.

Time consuming, yes, but necessary to engage in moving this thread forward in an attempt to move away from subjectivity to objectivity (which is what people do when they use a colour guide or consult a catalogue).

There needs to be some quantification process which is superior to the Mk I eyeball, which we all know, can be heavily susceptible to variation.

I have been doing a lot research into colour recognition and image recognition etc. I have produced colour histograms of a few stamps and tried to find a consistent method that would give a true refection of the stamp's colour.

I have to admit that I have failed so far to make any inroads and have realised that this problem is far beyond me. Having been around the block a few times I have realised that this is a task for the scientists; mathematicians and physicists and not a layman like me.

I will have to rely on my guesswork and some help from the forum when I run into colour recognition problems. The SG colour guide is really no help to me as their lavender is quite often my lilac, or vice versa! Like chaddy, my wife has suggested she would be a better judge of colour, but having been married for close on 54 years I don't at this stage want to get into any arguments that might compromise what I get dished up for supper.

[Global Administrator]

.
Over our 13 years here we have had dozens of threads Wailing At The Moon, seeking a magic way to sort stamp colours online from endless members.

As Allanswood has just shown perfectly above - it is a Fool's Errand. CANNOT be done online with any accuracy at all.

Like it or not, this system below works BEST for Commonwealth stamps - want to know what MYRTLE GREEN looks like - place this guide over the stamp in question - simple -

“Stanley Gibbons Stamp Colour Guide/Key”

[Allanswood]

I always look at rare shades in the simple understanding that a rare shade is a visually very different colour from those that aren't rare... and that 999/1000 I won't have that shade.

And that many collectors don't realise for years until someone else points it out to them, that they are partially colour blind to certain colours.


The quest for "50 shades of Lavender" is often a wild goose chase made easier by simply soaking a few for various lengths of time and you'll get your 50 shades.

But the variables of software, scanner, age, paper make it impossible to determine an individual colour for a stamp.
It's been discussed on SB for over a decade. It cannot even be attempted without a single, calibrated regularly, model of scanner that we would all have to buy and a certified test card of shades used to constantly re-calibrate (if it cold be calibrated that way), because even now if you have the same model, same age scanner as me - you can be sure that it will not scan the same stamp with the same result due to the variables.

That's why researchers have looked for other ways to confirm what colour a stamp was originally by not looking at its actual colour.

[satsuma]

The trouble is that the SG colour key isn't the basis for colour names that, for example, the ACSC uses.

Got a Cyprus Green on that chart?
Got a Lime Yellow on that chart?
Got an Intensely Dark (Black) Brown on that chart?

For that matter, did the Australian printers colour match the SG key to proof their output?

It's a useful tool, no doubt but it is not a panacea for assessing shades.

It seems to me that the online discussion is about attempting to help people identify their stamp shades with enough certainty that they don't sell rarities for a pittance through ignorance, and to help them avoid buying the cheapies as rarities.

This kind of scan can be far more helpful than a colour key. The OS stamp is the 88D Intense Deep Green

[Allanswood]

The ACSC make up their own "names" to differentiate between other shades of the same stamp. I thought the SG colour key was designed for use with the stamps produced for the Commonwealth (by UK Printers), showing examples of the shades that were ordered for stamps throughout the Commonwealth because so many countries ordered and use the same colours.

You can call something "Cypress Green" but thats just a made up name to show a difference with the other shades for the same stamp and not what the ink manufacturer or the printer at the time called it. I could have called it "Cootamundra Wattle Green" and write my own catalogue.

[Global Administrator]

The trouble is that the SG colour key isn't the basis for colour names that, for example, the ACSC uses.

Naturally - who said it was?? They make up their own names. Entirely their decision.

It is for the SG listed British Commonwealth stamps. And has worked well for a Century.

The OP was complaining he could not sort the Queen Victoria he showed - too bad for him - it is very simple using the SG Guide, for those with normal eyesight.

Collect only 1840 1d Black stamps if any other colours are too hard to decipher!
.

[castores]

Colour blindness aside, who says I see a colour shade the same as you.

It seems the only way is to have examples of the stamp (shades) and place your stamp against each to get 'best fit'

I think I'm quite bad at choosing shades. I recently was looking at a stamp with four shades and, as usual, went in circles for quite sometime.

Taking into account all that I've read in these forums on colour. I started again, ignoring what was written and just trying to place the stamps into 'like' piles of four. Then it was a matter of fitting my piles to the shades mentioned.

Not a perfect fit but the majority are now correctly listed.

Without having many shades to work with then the SG colour guide would have to give 'best fit'.

[Rigs]

It is for the SG listed British Commonwealth stamps. And has worked well for a Century.
.

There needs to be a serious review of expectations in this regard.

For example, how is that stamps printed over a century ago can be compared to modern day colour guides?

I have worked with chromalin proofs for thirty years.

Even the most recent pms color guides fade after a few years. I can guarantee they are a lot more accurate that many SG colourways.

After a hundred years no stamp looks like it was printed yesterday. Anyone who entertains such a delusion should have their head examined.

And then you want to distinguish subtle shades with an SG colour guide printed last year?

Seriously? ...

Perhaps the real reason is that if SG can perpetuate the difference in prices for so-called ‘shade varieties’ then they can keep on selling catalogs ...

[Global Administrator]

After a hundred years no stamp looks like it was printed yesterday. Anyone who entertains such a delusion should have their head examined.

Another candidate for collecting only 1d Blacks it seems.

For the other 99.9%, the time proven system works well.

[castores]

I don't have an SG colour guide, though I have taken images from the internet (along with other colour palettes).

I like to compare the 'same' shades and those that are close.

But when it comes to my stamps, if I have enough, (as per above post) I try to block them into the same number of shades listed before even looking at which might be which.

"If I have enough"

If anyone has mountains of worthless Australian 1d KG V side-face reds they don't mind throwing my way, I'd have something to play with...

[Allanswood]

SG colour guides are not for subtle shades of a colour. The guide only has about 140 named colours.

Subtle and small variations of shade I can't really fathom the need for . But if lillac is worth $1 and purple $100 then that's something. But trying to show 50 shades in between that were never in the printing records and suggesting or looking for a pot of gold rarity is just chasing the wind.

Many are those that describe a new expensive rarity and are the only ones with stocks of them and then yell to have it included in a catalogue.

[KGVIStamps]

One of the biggest problems I see with collectors and color shades is that they are trying to take a group of factors and just look at one part of the problem.

Generally speaking - for British Colonies who had their stamps printed by the Crown Agents and specifically for those from 1900 until 1950 when you see color shades listed in the catalogs they are almost always printings.

If you use either the Commonwealth Five Reigns Catalogue or the Stanley Gibbons Catalogue, you will see the initial listing which is the stamp as it was originally released. In many cases that is the only color you will see. But where there are multiple colors, notice that many of them have dates.

This would indicate that they were later printings and sometimes the date can help you identify the stamp even when the color seems confusing. Good examples of this are the King George V issues from about 1912 - 1920 where the paper colors changed due to supplier issues, and where the ink colors changed as well.

Look at other Colonies stamps from the same time period printed by the same printer and you might find yourself able to identify more of them.

The King George VI issues are an even better example. There were a lot of printings of many Colonies' issues and the stamps were printed by one of four printers (Bradbury, Wilkinson; De La Rue; Harrison; and Waterlow).

There are significant differences in the paper and gum from these printers over time that can be used in conjunction with the color data to make a decision about a specific stamp if you know the printing date.

Generally speaking the 1938 era issues tend to be on paper that will tend to have a yellowish gum; the time period around 1940 tends to be whiter than the 1938 paper and relatively thick looking (not very transparent)

The 1943 era paper tends to be the thinnest with a very transparent appearance and off-white gum; the papers after the war tend to be thicker appearing than the 1943 paper and still off-white; and the 1950 era paper tends to be a brighter white. I use black paper as a background and place the stamp upside down to view this effect.

If you can isolate the time period it is a lot easier to sort by colors. I do use the Gibbons Color Guide, but mostly to help me decide between a range of colors. I have never found an exact match, but it is useful in deciding on the various shades.

I have scans of my album pages on my website showing the listings of printings by Potter & Shelson and Frank Saunders along with various other people who have studied these stamps. Please feel free to visit if you are intersted in them.

I have attached four images that I put together as I studied the King George VI issues. Hopefully they will help you see what I was describing about the paper and gum characteristics.

One final tip - accumulate a large number of stamps when you are comparing color shades. Two or three is not enough to make a decision. I realize that can be difficult these days, but sometimes a dealer will allow you to view their stock.
.

[Rigs]

But the variables of software, scanner, age, paper make it impossible to determine an individual colour for a stamp.

That about sums it up ... before you even get to last year’s SG colour guide, or the sheriff’s smiling sun fade gif.

[Global Administrator]

Let's stick to the topic please folks.

SG colour guides are not for subtle shades of a colour. The guide only has about 140 named colours.

Subtle and small variations of shade I can't really fathom the need for . But if lilac is worth $1 and purple $100 then that's something. But trying to show 50 shades in between that were never in the printing records and suggesting or looking for a pot of gold rarity is just chasing the wind.

Many are those that describe a new expensive rarity and are the only ones with stocks of them and then yell to have it included in a catalogue.

Yes, SPECIALIST catalogues are for the tiny colour nuances - the ACSC lists and prices about 100 different "reds" for KGV ''1d Reds'' and my thoughts on the total lunacy of that, especially for USED copes are well known and well documented.

A mug's game IMHO. SG lists just a few. And the SG Colour Key ID's those nicely.

Glen

[MPSStampCollect]

I don't know if this was mentioned. The Stanley Gibbons Stamp Color Key has 200 shades (25 x 8). I entered all of the chip names into a spreadsheet so that I can quickly search for particular colors.

[castores]

I don't know if this was mentioned. The Stanley Gibbons Stamp Color Key has 200 shades (25 x 8). I entered all of the chip names into a spreadsheet so that I can quickly search for particular colors.

Cool, I'm considering getting said SG colour key.

[Global Administrator]

I'm considering getting said SG colour key.

I sell about 50 a year.

For anyone collecting COMMONWEALTH stamps per SG .. they are truly essential, they really are.

Glen

[castores]

I will be including in my next 'bulk' buy!

[satsuma]

I came across this quote a while ago when I was researching this topic.

So does anyone know whether SG colour matched the De La Rue or Bradbury Wilkins output for their chart.
Or were the differences too insignificant to bother with despite this opinion from the publishers Smith Press in 1949?