08-11-2026, 05:24 PM
Comparing the results for JPEGs and TIFFS, Claude was not happy because the two graphs converge (i.e. move in opposite directions) throughout the Herbal section.
It failed to notice that overall though, the shapes are often similar. So I had it calculate the relative changes at the local level. These are usually consistent between both scans. So if a page is lighter than the pages around it in the JPEGs, it is also lighter in the TIFFs.
This means that some observations remain intact, most importantly that Herbal B pages are lighter than the Herbal A pages in their proximity. ("On versos both captures give about 1.2 L* (p<0.02).") As you can see in the graph, the Q20 pattern also holds. Both captures also show the first and last folio as much darker, but the JPEGs give much darker values for the folios right after f1.
It failed to notice that overall though, the shapes are often similar. So I had it calculate the relative changes at the local level. These are usually consistent between both scans. So if a page is lighter than the pages around it in the JPEGs, it is also lighter in the TIFFs.
This means that some observations remain intact, most importantly that Herbal B pages are lighter than the Herbal A pages in their proximity. ("On versos both captures give about 1.2 L* (p<0.02).") As you can see in the graph, the Q20 pattern also holds. Both captures also show the first and last folio as much darker, but the JPEGs give much darker values for the folios right after f1.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)