|
I think if a document is to complex, or has too many cross-links (which I am guessing that this document might have) then I think that it may be beyond Calibre's ability to covert without running out of memory. Note that it is complexity that appears to cause the memory problem - not simpy document size. Calibre has been optimised to handle the vast majority of conversions that are of far simpler documents, but the consequnces are that it can fail on very complex ones. You may therefore simply be fighting a losing battle in trying to convert this particular book with Calibre.
As an aside you mentioned increasing the split value. If you do not keep it below 300KB then it is likely to fail on the vast majority of reading devices even if it did appear to successfully convert.
|