hello,everyone.
I use my acoustic model trained audio data. He can recognize my voice.
I now have a few questions.
1. I just need to identify the (0 ~ 9, next, previous), but my acoustic model size 2.7M. These are not related to the size and audio data?
2. I'm running android demo, in the recognition process, the recognition results have been changing. (example: I said 2, demo the results of a first identification, later turned into a 2), this can not control, do not Showing 1 to demo?
3. I need to train people to use the model. then I Training Model of audio data, is not needed more than the audio data, or can find a common voice characteristics of the audio data to train the model.
thanks
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
hello,
There is also a problem. When I do not talk. If there is some noise, you can
also identify a number of random results. This problem is very serious.
I changed some parameters, according to here:http://cmusphinx.sourceforge.net
/wiki/pocketsphinxhandhelds
Never change.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
[url]http://cmusphinx.sourceforge.net/wiki/tutorial[/url]Tuning the performanceApplication notesTroubleshootingYou've followed a link to a topic that doesn't exist yet. If permissions allow, you may create it by using the Create this page button.
Has solved two problems. A problem is left. Where is the answer to this
question?
I'm running android demo, in the recognition process, the recognition results have been changing. (example: I said 2, demo the results of a first identification, later turned into a 2), this can not control, do not Showing 1 to demo?
Does have some difficulty reading English.
Thanks ,
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
My question is not in a language model? I use to create the language model
network services. And this is a statistical model. Grammar language model
should I use?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Okay. I'm trying to learn English, which some shame.
What I mean is: I use the android voice recognition. In the recognition
process. The result is change. For example. Should be 2. but it will be from 1
to 2. This question is not relevant and the language model
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
hello,everyone.
I use my acoustic model trained audio data. He can recognize my voice.
I now have a few questions.
1. I just need to identify the (0 ~ 9, next, previous), but my acoustic model size 2.7M. These are not related to the size and audio data?
2. I'm running android demo, in the recognition process, the recognition results have been changing. (example: I said 2, demo the results of a first identification, later turned into a 2), this can not control, do not Showing 1 to demo?
3. I need to train people to use the model. then I Training Model of audio data, is not needed more than the audio data, or can find a common voice characteristics of the audio data to train the model.
thanks
hello,
There is also a problem. When I do not talk. If there is some noise, you can
also identify a number of random results. This problem is very serious.
I changed some parameters, according to here:http://cmusphinx.sourceforge.net
/wiki/pocketsphinxhandhelds
Never change.
You can find all the answers on your questions in wiki
http://cmusphinx.sourceforge.net/wiki/tutorial
hello.
Has solved two problems. A problem is left. Where is the answer to this
question?
Does have some difficulty reading English.
Thanks ,
hello ,Nickolay!
My question is not in a language model? I use to create the language model
network services. And this is a statistical model. Grammar language model
should I use?
Sorry, I do not understand what you are asking.
Maybe you can write in Mandarin instead.
Okay. I'm trying to learn English, which some shame.
What I mean is: I use the android voice recognition. In the recognition
process. The result is change. For example. Should be 2. but it will be from 1
to 2. This question is not relevant and the language model
I still don't understand