The Voynich Ninja

In another thread You are not allowed to view links. Register or Login to view. a few other locations to look for potential features in MSIs, including the top right corner of f1r. It's mostly empty in the processed images of the models I trained on the margin of f1r, but this doesn't exclude possible writing there, provided that something about this writing (the writing implement, the condition of the vellum, the ink, the way the inscription faded, an offset image) is substantially different from the writing on the right margin.

I haven't yet trained any models, but here is the set of all MSI images of this area and a few potential new features to investigate. The cropped images are from the original 16-bit TIFFs mapped to 8 bit by applying gamma 3.0 and then running a min-max filter for this area in each individual TIFF (setting the darkest pixel to black and the brightest to white and scaling the rest proportionally).

[attachment=9238]

The full resolution image: You are not allowed to view links. Register or Login to view.

Without running another linear regression model it's really hard for me to assess the actual presence and shapes of these features (and the feature Searcher mentioned), since different parts of them are visible in different wavelengths.

I find it tricky, especially here. This corner must have been handled a lot by grimy fingers over the centuries.

oshfdk wrote:

Quote:Gamma1 (dark and bright parts are equally represented): You are not allowed to view links. Register or Login to view.
Gamma2 (more details in the dark parts of the image): You are not allowed to view links. Register or Login to view.
Gamma3 (even more details in the dark parts of the image): You are not allowed to view links. Register or Login to view. 

[url=https://github.com/oshfdk/f67r2planets/blob/main/images/1rtoprightcomb900x1500x1200x700g3.0b1.0w99.0.jpg][/url]

I'm not sure about the rest notes in the right top corner on f1r, but your approaches above quite clearly show the letters "AR" in that corner.
[attachment=9241]
I can only suspect two another symbols, one of which is bird-symbol again and the second one reminds me the symbol of Pisces. The notes that look like left-declined text of three letters is more or less visible in general, but questionable in its interpretation or even meaningfulness. For me, it looks like "pic", meaning "pictus" (picture), but all this are suppositions. It would be pretty useful to see the picture with the minimal subjective human intervention.

(23-09-2024, 10:51 AM)Searcher Wrote: You are not allowed to view links. Register or Login to view.I'm not sure about the rest notes in the right top corner on f1r, but your approaches above quite clearly show the letters "AR" in that corner.

I was trying run a test on this AR shape, linear regression cannot separate it well from the rest of the page. However, I noticed that the model lit up the folio number very distinctly. So, given the folio number is very clear, I tried the other way around and ran linear regression with the following mask (black areas positive, white areas negative, grey areas are not used in the training).

[attachment=9243]

Then it shows a pretty clear straight stroke (maybe an older folio/quire number?) roughly coinciding with R in AR.

[attachment=9244]

This is the whole page processed with the same weights. There are some faint forms in the bottom right corner and on the top margin, center, but so far it's hard to tell what they are.

[attachment=9245]

This doesn't give much in terms of AR shape, but shows one more time how easy it is to separate the right margin writing and the signature at the bottom even using a very small and clear sample of writing (the folio number).

This is not the end of my investigation for the AR pattern, I will further refine the search using other clearly visible shapes. But as I see it, there is really something resembling an AR-like shape on the page, but either it was a single shape created in some way, or this is just a combination of different features on the page that together resemble an AR shape. It's also important to consider what other elements on the page are enhanced by the same linear transformation of color channels, this might give some clues as to what a shape is made of.

Training on the right leg of AR only (note that when I say "left of AR", "leg of R", etc., I'm only specifying the locations of page, I'm not sure whether this is a single feature or some features mingled together). The diagonal stroke is highlighted as well, also the big weirdos, including the smudge over the second weirdo on the left margin. The texture of vellum is highlighted in many places. Possible feature 2 ("big/Sig") is also visible.

Adding the diagonal stroke and the leg together. More texture and ink more pronounced overall on the page.

Adding the diagonal stroke, the leg and the folio number together. Text overall is more visible.

Without the folio number (I just kept a bit of the background to the left of it in the samples), but adding a clear piece of A crossbar.

[attachment=9247]

Full resolution (recommended): You are not allowed to view links. Register or Login to view.

So far I failed to reproduce a clear shape there. It looks like the diagonal stroke is more associated with the text ink, while the leg of R is more associated with the weirdos ink, I'll try the regression on the weirdos by themselves to maybe clarify this.

(23-09-2024, 10:51 AM)Searcher Wrote: You are not allowed to view links. Register or Login to view.I'm not sure about the rest notes in the right top corner on f1r, but your approaches above quite clearly show the letters "AR" in that corner.

Note that so far I've tried to uncover this shape using only clearly visible elements on the scans, because otherwise there is risk of making the model overfit and creatively combine layers to show just what you want. The risk of this especially with the simple linear regression is extremely low, but I prefer to play it safe. I will certainly try directly using the points you provided as the training shape after I exhaust other possibilities.

I wasn't able to detect AR feature as a single shape, but there is some new information about some of the elements.

Given that the model for the right leg of R highlights red paint in weirdos, I checked what would red paint model highlight. Among other things it displayed a very clear inverted V shape consisting of the right leg of R and the bottom part of the diagonal stroke. It is possible then, that R is actually a combination of a diagonal stroke in text ink and an inverted V mark of some other kind.

I trained another model on this inverted V shape, it highlighted a rectangular feature at the bottom right of the page. This could be something like an offset of a rectangular label on the inside of the front cover, or maybe something else entirely. Maybe the inverted V feature is an offset as well, I can't tell really.

I made then a visible text ink model off the main text, which highlighted the diagonal stroke of the R feature and the folio number, but nothing else in that corner.

As the last resort I put the proposed shape of the AR feature over the scan, using the holes in vellum as guides, and recreated the shape for training. Nothing new was detected this way. Various elements are highlighted on the page, consistent with the idea that AR shape can be a combination of different kinds of marks.

[attachment=9248]

Full resolution: You are not allowed to view links. Register or Login to view.

I also placed the flipped version of 1v over 1r to see if any of these shapes could be explained by anything that happens on the other side of the folio, but couldn't find anything.

So, that's what I found in the upper right corner:

[attachment=9249]

Hope this helps!

Please note that there was an earlier thread about this and other hidden/faded notes: You are not allowed to view links. Register or Login to view..

Highlighting these hidden notes was done by 'eye and hand', in addition (perhaps) to image processing.

I could not quickly find the exact location - it may have been another thread - but Searcher confirmed this aspect of the interpretation.

In particular, I remember that there was a suggestion of another mention of Tepenec/Sinapius in the margin of one page, though I don't remember which version of his name it was. This turned out not to be real in the end.

I think I'll be adding some bits and pieces related to You are not allowed to view links. Register or Login to view. MSI's here. I do occasionally run linear regressions on this or that feature, just to understand what it could be.

There is a tiny o-like loop (or part of a loop) in a EVA p (which belongs to EVA 130 composite) roughly at the middle of the page. From the visible light TIFF it's hard to tell whether this is just a spot or an actual pen stroke. I've checked the following registry of extended characters: You are not allowed to view links. Register or Login to view. and EVA 130 doesn't have this loop there.
To me the loop looks a bit more yellow than the normal ink and appears similar to the color of various spots on the page.

[attachment=9498]

The verso of the folio has a painted leaf at the same location, so this doesn't look like a see-through of some text from the back of the page.

I've extracted the cropped portion of this area from individual MSI images, here's the result:

[attachment=9499]

This doesn't seem to be normal ink, at least some of the images that clearly show the ink part have no traces of this small feature. On the other hand, it does look like a clearly defined feature. I tried to identify other parts of the folio that have similar spectrum (similar presence/absence profile across the whole set of images). The following image has all areas with similar profile highlighted in red.

[attachment=9500]

It doesn't look like any other red areas show anything of interest. There is a red circle above EVA 'rod' to the bottom right, corresponding to a spot on the page, however it appears incidental. On the individual MSI images this is just a blurred square-like blob:

[attachment=9501]

Considering all this, it does look possible that someone put a tiny circle or semicircle on that EVA p, maybe using a different kind of ink or paint.