Breaking Text up (Streaming Synthesis)

[Alicia Z]

When the input text is too long, what methods do we use to break it up into smaller pieces? (Or in other words, where do we put the longer pauses in the text?)

Do we process the text and divide it into pieces by grouping sentences that share the same main idea, or do we cut the whole piece into parts that have (approximately) even lengths?

[Simon]

In the Cereproc system, the pieces are called “spurts” and are either sentences or parts of sentences. We do enough text processing to predict where pauses will be (it might be as simple as at every period and comma). We then assume that these pauses are long enough to prevent any co-articulation spreading across them.

Very few systems (if any) deal with units larger than sentences in any meaningful way.

[Author]

Posts