I've managed to find some work-arounds for my issues with Balabolka. As mentioned, if I have a multi-sentence recording to make, if I perform a pitch change one sentence, as soon as I try to make a pitch change in a following sentence, the adjusted word would be followed by an un-commanded period of silence, before proceeding to the next word. At first, I simply booted it up in ocenaudio and deleted the offending silence. Then I began to notice that not only was there a silence, but the subsequent word would have an un-commanded pitch change applied to it, as though it was beginning a new sentence, so just deleting the space would result in a very odd-sounding sentence. For some time, my solution was to break up my vocal dialogue into chunks, one per sentence, or longer if no pitch changes were implemented. I would then stitch them back together in Ocenaudio... a time-consuming process. For a time I had a third option, which was to use a feature to manually insert audio breaks, which I used in place of periods, fooling it into thinking it was one long sentence. This worked most of the time, but I would occasionally encounter another break during a long message. Now, I've discovered that all I need to do is perform a hard return after ever sentence, as though each was a separate paragraph. This appears to work, though the paragraph break seems to create a break delay longer than your standard period does, so I need to trim each one slightly.