Artifacts Of Scanning

My comeuppance: never having been terribly interested in the details of printing stamps (eg rotary vs flat press), I am driven to nail-down the different methods of printing postcards. Go figure.

Perhaps the greatest variety of printing techniques are found in the halftones. One reason is that there were so many publisher active during The Golden Age, when halftone techniques dominated the postcard industry. But another reason is that there are just so many different ways to generate so many different patterns of so many different dots.

Recently, I've noticed a number of cards with a very peculiar pattern in the image:



Q/ Why in the world would someone create a tartan pattern of dense & sparse dots? Does the tartan pattern subliminally smooth the image?

I typically scan postcards cards at 200 dpi but, for smaller areas of interest (the stamp, cds, publisher imprint/logo, or a detail of the image on the 'art' side of the card) I scan at 600 dpi.

And here's what two parts of that same area of interest look like at 600 dpi:





As you can see, the tartan pattern is diminished, revealing uniformly-spaced dots.

I won't bother embarrassing myself by trying to do the math to find an exact answer, but I can promise you those dots are not spaced at 200 dpi, or a harmonic thereof.

And, so, a cautionary hint: even if you scanned it yourself, you may not be looking at what you were looking at.

Cheers,

/s/ ikeyPikey

edited to correct 'checkerboard' to 'tartan'

PS: none of these images were run thru any image processing algorithms

Optical illusion..?

An optical illusion is something our brains produce from something our eyes see.

The tartan pattern is a real (not illusory) artifact of scanning.

If the tartan pattern led you to see a Scot holding bagpipes ...

Cheers,

/s/ ikeyPikey

Editing to add: I apologize, Wert, for not being more directly responsive to your point. The way to prove that the tartaning is not, itself, an optical illusion is to enlarge the tartan'd image.

Tartanizing and moiré are problems not only of scanning. Preparing printing plates as well!

In the Brasil examples the line structure of the design is interreacting with the cross screen necessarily in preparing a photogravure cylinder. The gauge of that cross screen determines the "size" of the squares! When this series was first issued in 1954, the Casa da Moeda used a completely different type of screen that guaranteed NO moiré! But after a few years they forgot about that and introduced the cross screen with all fatal consequences.....



The traditional method of preparing a photogravure cylinder was to take the original design that was often 4 times as large as the eventual stamp size, and copy it photographically 2, 4 or 5 times and then used that strip of block in a step-and-repeat machine to make up a whole multipositive film of 100 stamps. That film was proceeded and got a  screen projected on it before it was wrapped around the photogravure cylinder; then the chemical  etching could start, etc....

All characteristics of that original 2,4, 5 or 10 stamps would be reproduced 50, 25, 20 or 10 times and were available in a counter sheet.

als these characteristics were already in the multipositve film, a consecutive new cylinder - with a different screen or not - would have the same characteristics!

In the two blocks above several of such characteristics are indicated!  They are NOT plate flaws! They are systematical characteristics  occurring a lot of times per sheet, whereas a plate flaw would occur ONLY ONCE per sheet!

Furthermore it is obvious from this scan that the blocks [xadrez grande] that were caused by moiré are different each time depending on the position in the sheet - theoretically it would be possible to tell the 100 positions apart!  :)

Wonderful information in this thread.

Galeoptix, Greetings: It is possible to plate this issue, and you have not done it yet? ;)

Here is another example of moire patterning (THANK YOU, GALEOPTIX!) that is an artifact of the digital scanning of a half-tone image on a postcard.

Happily, it is a particularly lovely & informative postcard, to boot.

http://en.wikipedia.org/wiki/Moire_pattern ... if you'd like to catch-up ;)

Cheers,

/s/ ikeyPikey

First, let's get the informative part out of the way:



Here's the pretty part (at 200 dpi); note the moire throughout the sky:



But this clip (at 600 dpi) shows that the moire is but an artifact of digitization:



While postcard collecting is mostly by subject, I always set aside anything by Tuck:



And this card, as a special bonus, bears the crest of which Tuck's were proud:



Leaving no stone unturned, this card even has a decorative stamp box:

Ikey,

I started with Brasil in September 2014 [!] and joined a Brasil web forum. They mentioned several shapes of squares but apparently had no idea what caused it and why some stamps have small and why other have large squares...

Their division in 3 types - liso (i.e. no squares), grande and pequeno is a simplification!







you need to specify the gauge of the screen and also the type of screen!  Furthermore, the direction of printing is important! The ink flowing downwards [10cr peq] or upwards [10cr gra]...

Screen 60 [lines/cm] is responsible for the smallest squares, screen 70 and 80 for the somewhat larger ones....

As I mentioned above you can - provided you have enough material - plate the cylinders as each position has a different arrangement of the squares. But so far I am happy to find most of the version of the 1954 Bisneta Series. The 0.20 is supposed to have a large square version but I haven't found it yet....

I admire your timing!

anytime you have a series of dots (pixels of your screen) trying to reproduce other dots(halftone dots in the scan) at certain magnifications a moire pattern can occur. These patterns will appear different and at different magnifications on other monitors.

When preparing scans of already halftoned images for print, moires are reduced/eliminated by using a certain halftone screen pattern and or rotating the screen. This was back when printers were still preparing plates from film. Not sure what they do now with direct-to-plate printing.