This can be done without special software. An important step when splicing in aspiration noise is to *replace* equivalently long portions of the voiced onset of the /b/ sound. For example, copy the first 30 ms of aspiration from /p/ and paste it at the 30ms mark of the vowel after the /b/ burst. Otherwise, you get the /p/ aspiration and then the complete /b/ voiced onset (the rising F1 transition). If this is what you've been doing, then I'm not sure why it wouldn't sound good.
Another thing- since you are creating sounds from the original voiced utterance, you’ll get a pitch contour that is too low for an aspirated sound. The pitch manipulation function in Praat can solve this issue as well.
I am looking for a good a way to manipulate voice onset time in real speech, for example to create a continuum from "bear" to "pear".
So far, I spliced in aspiration noise between fricative and voice onset using PRAAT, but the result is not really satisfying.
Can anybody give advice?
Thank you very much!
Daniela SAMMLER, Ph.D.
Max Planck Institute for
Human Cognitive and Brain Sciences
phone: +49 341 9940 2679
fax: +49 341 9940 2260