The file ./model/en-us/en-us-phone.lm.bin is a binary file. Where does it come from ?
Also, where is strored the list of phonems in the source code ?
I mean when the recogniser loop through the list of phonems ?
Last edit: Esteban Lorantino 2018-10-23
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank, your right. I see them at the top of the file: AA^@AE^@AH^@AO^@AW^@AY^@B^@CH^@D^@DH^@EH^@ER^@EY^@F^@G^@HH^@IH^@IY^@JH^@K^@L^@M^@N^@NG^@OW^@OY^@P^@R^@S^@SH^@SIL^@T^@TH^@UH^@UW^@V^@W^@Y^@Z^@ZH.
I found that the invalid ones (AI, OE, etc...) that I have are from a french dictionariry from "LIUM" but I have no idea from where is comming this "LIUM dictionary" ...
Do you have any idea about it and where is the source (URL) ????
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The file ./model/en-us/en-us-phone.lm.bin is a binary file. Where does it come from ?
Also, where is strored the list of phonems in the source code ?
I mean when the recogniser loop through the list of phonems ?
Last edit: Esteban Lorantino 2018-10-23
See 'Training phonetic language model for decoding'
https://cmusphinx.github.io/wiki/phonemerecognition/
What will happen if my dictionary file contains phonem that are not in the CMUBet list, such as AI and OE ?
Last edit: Esteban Lorantino 2018-10-23
You will see an error about a phone missing in acoustic model.
Can I extract the list of phones keywords from the acoustic model ?
You can look inside mdef file. If mdef file is binary you can convert it to text and look inside.
Thank, your right. I see them at the top of the file: AA^@AE^@AH^@AO^@AW^@AY^@B^@CH^@D^@DH^@EH^@ER^@EY^@F^@G^@HH^@IH^@IY^@JH^@K^@L^@M^@N^@NG^@OW^@OY^@P^@R^@S^@SH^@SIL^@T^@TH^@UH^@UW^@V^@W^@Y^@Z^@ZH.
I found that the invalid ones (AI, OE, etc...) that I have are from a french dictionariry from "LIUM" but I have no idea from where is comming this "LIUM dictionary" ...
Do you have any idea about it and where is the source (URL) ????
What is a "phoneset map" ?
(source: https://cmusphinx.github.io/wiki/tutorialadapt )
Last edit: Esteban Lorantino 2018-10-24
https://sourceforge.net/projects/cmusphinx/files/Acoustic%20and%20Language%20Models/French/
Mapping is a function that maps phones of one phoneset to another one.