Questions about scene breaks in conversion

[lunixer]

Hello. I've noticed some weird behavior with the scene break heuristic processing and had some questions.

1) Why does Calibre not save the preferences in "replace soft scene break" if, for example, I use * * *? Or anything else. No matter what I put in there that field is always blank when I go to convert something.

2) If I leave the "replace soft scene break" field blank and my book uses blank lines for soft scene breaks, will Calibre leave those blank lines? Note that under look and feel I use both "removing spacing between paragraphs" and "add blank line between paragraphs"

3) What heuristics does Calibre use to detect a soft scene break? I just read a book where I replaced soft scene breaks with * * *, to see how it would work, and it only detected about half of the scene breaks. The rest were completely unmarked. The book used a blank line between scenes.

Thanks!

[theducks]

Originally Posted by lunixer

Hello. I've noticed some weird behavior with the scene break heuristic processing and had some questions.

1) Why does Calibre not save the preferences in "replace soft scene break" if, for example, I use * * *? Or anything else. No matter what I put in there that field is always blank when I go to convert something.

2) If I leave the "replace soft scene break" field blank and my book uses blank lines for soft scene breaks, will Calibre leave those blank lines? Note that under look and feel I use both "removing spacing between paragraphs" and "add blank line between paragraphs"

3) What heuristics does Calibre use to detect a soft scene break? I just read a book where I replaced soft scene breaks with * * *, to see how it would work, and it only detected about half of the scene breaks. The rest were completely unmarked. The book used a blank line between scenes.

Thanks!

If you recently set that in preferences AND you have previously converted that book, you are using the books saved settings.
Click restore defaults on that books conversion page.

[lunixer]

Originally Posted by theducks

If you recently set that in preferences AND you have previously converted that book, you are using the books saved settings.
Click restore defaults on that books conversion page.

Already tried this. It didn't work. When I go *back* into the preferences to look, the heuristic processing thing still is blank. I have made sure to click apply. The preference is definitely not being saved.
Vitals: HP dv7, Ubuntu 10.04, Calibre 8.13 and 8.14 (behavior is in both)

[theducks]

Originally Posted by lunixer

Already tried this. It didn't work. When I go *back* into the preferences to look, the heuristic processing thing still is blank. I have made sure to click apply. The preference is definitely not being saved.
Vitals: HP dv7, Ubuntu 10.04, Calibre 8.13 and 8.14 (behavior is in both)

Confirming your observation: Natty(32), Win XP, Saved (applied) value not showing on re-visit.

[user_none]

Please open a bug for 1 at https://bugs.launchpad.net/calibre .

[theducks]

Bug #826038 submitted

[lunixer]

Originally Posted by theducks

Bug #826038 submitted

Thanks. I've subscribed to the bug and linked back here. Do you have any comments on the other questions that I asked in the OP?


2) If I leave the "replace soft scene break" field blank and my book uses blank lines for soft scene breaks, will Calibre leave those blank lines? Under look and feel I use both "removing spacing between paragraphs" and "add blank line between paragraphs"

3) What heuristics does Calibre use to detect a soft scene break? I just read a book where I replaced soft scene breaks with * * *, to see how it would work, and it only detected about half of the scene breaks. The rest were completely unmarked. The book used a blank line between scenes.

[ldolse]

Originally Posted by lunixer

2) If I leave the "replace soft scene break" field blank and my book uses blank lines for soft scene breaks, will Calibre leave those blank lines? Under look and feel I use both "removing spacing between paragraphs" and "add blank line between paragraphs"

Yes - Calibre will retain those blank lines, even if those other features are enabled.

Originally Posted by lunixer

3) What heuristics does Calibre use to detect a soft scene break? I just read a book where I replaced soft scene breaks with * * *, to see how it would work, and it only detected about half of the scene breaks. The rest were completely unmarked. The book used a blank line between scenes.

Basically Calibre looks for any blank lines in the document - it then determines which are 'whitespace' by looking at the paragrahs before and after them - to see if they're placed around headings, frontmatter, etc. After that it assumes the rest are scene breaks. Basically to be a scene break there needs to be a full, punctuated paragraph both before and after the scene break. Scene breaks which are purely defined in css aren't detected. You'd have to open a bug with the book to allow this to be investigated further.

I did just fix some bugs in this logic which haven't been submitted to trunk, but they only cover the scenario where the document has a mix of hard and soft scene breaks.

[user_none]

Also, TXT and PML input for instance require two plus blank lines to be retained as a scene break.