22-03-2026, 09:22 AM
I agree with this and wanted to suggest something similar. It should be possible to determine if the distortion pattern fits better with a Camera Lucida, concave mirror, or other modern projection system. The problem I see is that both digital source images need to be distortion free for such comparison. Are they, especially the VM scans? Largely, I guess.
We'd also need distortion data of projection devices. An improvised optical system with non-parallel, non-flat source- and projection surfaces might create unexpectedly complex distortions. I think I noticed some pincushion distortion but I can't say if this was the optimal solution to reconcile the 2 images.
I also agree that the text inaccuracies might be the result of optical errors in the projection, but I can't imagine why you would choose such approach to begin with instead of directly copying or at least proof-reading the text. Even for someone unfamiliar with 'Voynichese', some errors are obvious. It also looks like the canvas might have been shifted and then re-aligned while copying the text. It however shows that the artist put emphasis on the drawing instead of the text. Which makes zero sense if the copy was created to help decipher the manuscript. So it probably wasn't.
We'd also need distortion data of projection devices. An improvised optical system with non-parallel, non-flat source- and projection surfaces might create unexpectedly complex distortions. I think I noticed some pincushion distortion but I can't say if this was the optimal solution to reconcile the 2 images.
I also agree that the text inaccuracies might be the result of optical errors in the projection, but I can't imagine why you would choose such approach to begin with instead of directly copying or at least proof-reading the text. Even for someone unfamiliar with 'Voynichese', some errors are obvious. It also looks like the canvas might have been shifted and then re-aligned while copying the text. It however shows that the artist put emphasis on the drawing instead of the text. Which makes zero sense if the copy was created to help decipher the manuscript. So it probably wasn't.