Quote:
Originally Posted by Starson17
Dwanthny,
Are the "SVG glyphs" character glyphs, word glyphs or something else? I don't have a Kindle, but trying to read the links you posted, it appeared that Topaz formats were really scanned books, broken down into small images of words or characters to achieve reflow for the Kindle, with OCR text linked to the word/char images for searching. Does it look like that's what it's doing when you read such a book on the Kindle?
|
Based on what I've read the Topaz format creates a custom library of letter glyphs based on each book. Essentially every instance of the each letter in the same font/caps will look roughly the same, so a custom glyph is created for all matching cases. (However I don't believe the format itself is aware of which letter is which). Topaz doesn't use SVG, but apparently whatever it uses is close enough to svg that the reverse engineers chose to use that. Unfortunately using SVG doesn't work well for ebook readers because of performance/reflow. There were discussions around creating custom ttf files instead in that thread, which would probably scale performance for the readers and may even preserve reflow, but I'm not sure if they're still pursuing this..
I've been reading some Topaz books on kindle for iPhone, and the more I read it the more clearly I can see how it's doing this. In a number of cases I've seen fragments of letters split because of bad scans, hyphens and spaces in the middle of words which were originally split, but reflow has put back together (without realizing the hyphen is no longer neccessary because it's just another glyph in the glyph cache).
People have complained about the quality of Topaz in the past, but what they're really complaining about is bad scans. When you think about what the format itself is doing - taking scans of fixed pages and turning them into a transparently reflowable format - it's pretty amazing.