VoxForge
VoxForge
Email from Andrew:
Hi,
I'm having difficulty segmenting an audiobook with htksegment.pl.
I've successfully completed the prior steps from the "Automated Segmentstion Using Forced Alignment (Draft)" document.
The error I'm getting from sox is: "sox formats: bad input format for data file [filename] data encoding was not specified"
I've tried using u1 file format, but it resulted in static-filled wav files.
Please help!
--- (Edited on 5/13/2009 1:27 pm [GMT-0400] by kmaclean) ---
HI Andrew,
Have you looked at the Sox manual? I believe that the scripts assume
WAV or RAW format files. You might try converting your audio to one
of these formats before running the script.
Can I post this thread on the Website so that others may benefit from
this discussion?
thanks,
Ken
--- (Edited on 5/13/2009 1:27 pm [GMT-0400] by kmaclean) ---
Email from Andrew:
I ended up using an old version of sox and I got past this problem, only to have festival seg fault when I try to use the voice based on the segmented audio.
--- (Edited on 5/13/2009 1:28 pm [GMT-0400] by kmaclean) ---