4 ms·
Compositor: A database of eighteenth-century printers’ ornaments
- traverseda 5y agoMost of these are low enough resolution that I certainly couldn't use them to produce anything. I guess they're mainly intended for book/fragment identification? Which is too bad. They're presumably old enough that they could be public domain, although the university of Birmingham doesn't seem to have licensed them permissibly. It would be nice to make some stamps out of these. Libraries of design elements like this can be really nice for making all kinds of stuff, but the quality and the licensing both mean this will probably only be usable for reference.
- kwinkunks 5y agoIt really is such a waste of time and resources. Like you said, the database seems to be 'all rights reserved', even though the original printed artifacts are obviously out of copyright. Lots of institutions do this with scans of old books, paintings, etc. I don't understand why institutions limit their own impact like that.
- rektide 5y agohate to say it but could definitely use some machine learning to radically upscale most of them
- mistrial9 5y agothey appear to be using some process already to find the ornaments. As noted here, they may be more interested in improving their own software and their own publicity than providing anything useful for others
- thechao 5y agoThere's actual garbage in the DB; clearly, they've used some sort of automated method for thresholding images, "finding" ornaments and pulling them out. As such, I don't think the ornaments in the DB are particularly interesting. The raw data (unthresholded) is, but I bet that's not copyrightable.