Hi;
I am doing my graduation project using Sphinx4. My project is to prapare a speech recognition API for novice java programmers. In addition to this I want to add speaker identification ability to my API with the help of Sphinx.
I have two questions;
1-) I will let the user read a static sentence (same for everyone) and get the signal array from Sphinx as soon as the speaker reads it. How can I get the recognized signal from Result in order to create HMM parameters of the speaker or is there any way of doing it in Sphinx.
2-) I am using the WSJ_8gau_13dCep_16k_40mel_130Hz_6800Hz acoustic model and cmudict.0.6d dictionary which can be found lib directory but the performance is not so good (40-45% success). How can I improve the performance?
Thanks in advance...
(By the way I can send my whole project)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi;
I am doing my graduation project using Sphinx4. My project is to prapare a speech recognition API for novice java programmers. In addition to this I want to add speaker identification ability to my API with the help of Sphinx.
I have two questions;
1-) I will let the user read a static sentence (same for everyone) and get the signal array from Sphinx as soon as the speaker reads it. How can I get the recognized signal from Result in order to create HMM parameters of the speaker or is there any way of doing it in Sphinx.
2-) I am using the WSJ_8gau_13dCep_16k_40mel_130Hz_6800Hz acoustic model and cmudict.0.6d dictionary which can be found lib directory but the performance is not so good (40-45% success). How can I improve the performance?
Thanks in advance...
(By the way I can send my whole project)