You understand me very well. Of course "synchronization marks" are strictly related with amount of mp3 files. Turning on new file every 10 second probably might cause some delays, what is not desirable.
Quote:
Originally Posted by ixtab
1 mark/ 30 secs = 6600 marks, 5.5 mpp
|
Could be a good factor. However... I think LOTR is not this first ebook which come to my mind when I hit on this an idea.
I think this will be better for shorter texts, even for podcats which provide transcription.
I assume, that original TTS mechanism gets part of text and keeps it in some kind of buffer. The question is how to write program, which could pretend TTS and use this same kind of buffer.