seems to be
tcr. In real life, this is because I used calibre, which cannot remove images and probably doesn't do tcr well. If I wanted to, I could have
1. converted all books to zip
2. run this from dos:
for /r %x in (*.zip) do "C:\Program Files\7-Zip\7z.exe" d -r "%x" *.jpg *.png *.jpeg *.gif
3. converted all books from zip to whatever before doing the tests. In that case, epub is about tied with first, but pdf really sucks.
Before you ask, why did I look at gzed books? Because since stacker you could compress files using software.
[Personal link deleted - Please see our rules regarding signatures for new membes - MODERATOR]
P.S. My reason for posting is in case there's a less insane way of removing images that someone seeing this knows about easier than the above method. If so, please inform me; I'm rather ignorant about ebooks.