Processed 116v/17r Multispectral

[npcompl33t]

I'd agree that npcompl33t's processed images are among the clearest I have ever seen, so thanks for that!
But I noticed something odd. Your last b&w image shows a strong discrepancy in the word between what looks like 'te' and 'cere'.



In the false color image above - and all multispectral or true color images I have seen - the word between 'te' and 'cere' displays some odd fused glyphs in the middle like 'c8av', but in your b&w image, these weirdos disappear, leaving a faint but much less ambiguous 'c8v'. What was your workflow for this b&w image?

See also oshfdk's multispectral viewer that easily lets you cycle through wavelengths:
You are not allowed to view links. Register or Login to view.

The issue still remains that we have a lot of highly ambiguous letters that look like the scribe changed his mind half-way or later amended them. Which glyphs are latin and which are voynichese is debatable, but as you said, some look like voynichese 'ch'. Some are completely odd like the o with ^ attached. That's why I'm extremely careful with any proposed readings.

The B&W image here was simply the raw PCA 3 output, using the 21 aligned multispectral captures of f116v, covering visible, UV and infrared imaging. Each capture was background-normalized using a Gaussian estimate (σ = 45 pixels), then standardized before PCA. PCA3 represents the third-largest pattern of variation across those captures and is not a single wavelength or a particular ink. Here is the full PCA3

   

As for why ths is visible specifically in PCA3, at 870 nm, the dense mark is still about 18% darker than nearby parchment, while the adjacent stroke is only about 6% darker. PCA3’s negative infrared weights cancel much of the dense mark’s contrast.

   

PCA3 assigns negative weights to the infrared captures. Those contributions counteract the visible/UV contributions, cancelling the dense mark’s darkness and even making parts slightly lighter than the background.

   


There are multiple things that could explain this: different ink, a heavier ink deposit or different application. If it is helpful I could look into this more, perhaps applying some of the techniques discussed in You are not allowed to view links. Register or Login to view.