Image processing, highlighting hidden details

[ReneZ]

I added a few more images to the second page: You are not allowed to view links. Register or Login to view..

These are from quire 13 and show a few more interesting artefacts.
I think that overall the variation in colour/darkness is very hard to interpret.

As requested by Tavie, I have now added all ink false-colour images of Q13 and Q20.

For consistency, I renamed the page to:
You are not allowed to view links. Register or Login to view.

[Rafal]

If I may ask, how did you achieve these images with "rainbow" font?
Is it some graphic filter publicly available or self made, dedicated tool?

[ReneZ]

This is indeed self-made, based on some standard tools I created and one dedicated mapping routine.

I experimented with a colour decomposition of which the main axis is along the common hue of ink and parchment. This was not quite as useful as I hoped, but allowed to make these images.

[Rafal]

This is indeed self-made, based on some standard tools I created and one dedicated mapping routine.

Then I would have yet another question.

The technique definitely looks "cool" and seems to indicate places where the scribe changed the pen or made a break and continued the work several hours later.
But does it really? It it possible that algorithm "exaggerates" small, random differences in hue and shows that something important happened at some place where in fact nothing important happened?

Actually I have little experience with writing with the goose feather  When you do it, you put the feather from time to time into ink, right? Does this method detects such moments and colour changes when the scribe put the quill into the ink?

[ReneZ]

The whole purpose of such false-colour images is to magnify (or indeed exaggerate) smaller differences. Whether they are random or not is then a matter of interpretation.

In the above image, the effect is very pronounced, and it can also be seen easily in the original image.