search

readbeyond / aeneas

aeneas is a Python/C library and a set of tools to automagically synchronize audio and text (aka forced alignment)
http://www.readbeyond.it/aeneas/
GNU Affero General Public License v3.0
2.45k stars 218 forks source link

Put end boundary in middle of silence #252

Closed george-roussos closed 4 years ago

george-roussos commented 4 years ago

Hi,

This tool is very mighty. I am trying to use it with Swedish, and for the most part it works okay. However, the end boundary almost always falls on breath segment, or the beginning of the next fragment, and so there is interference in the alignments. Is there any way to instruct aeneas to only put the boundary on a silent part? For example, right now if there is silence, it is all included in the previous fragment; it would be nice if it could be evenly split.

Thanks!

pettarin commented 4 years ago

https://www.readbeyond.it/aeneas/docs/adjustboundaryalgorithm.html

Nice to see interest from (people from) Storytel...

On 5/31/20 5:15 PM, George Roussos wrote:

Hi,

This tool is very mighty. I am trying to use it with Swedish, and for the most part it works okay. However, the end boundary almost always falls on breath segment, or the beginning of the next fragment, and so there is interference in the alignments. Is there any way to instruct aeneas to only put the boundary on a silent part? For example, right now if there is silence, it is all included in the previous fragment; it would be nice if it could be evenly split.

Thanks!

george-roussos commented 4 years ago

Thanks! And thanks a bunch for this amazing tool 😀