What is the difference between continous and semi continous model

[Chandrashekhar Kotekar]

Hi,

I have very basic(stupid) question about acoustic models. I would like to know what is the difference between continous and semi continous model? and how to decide which model to use?

Thanks,
Chandrashekhar

[Nickolay V. Shmyrev]

I would like to know what is the difference between continous and semi continous model?

Semi-continuous models tie model parameters together so that the number of parameters is smaller. Semi-continuous models are less precise to describe the variability of the language.

Due to smaller number of parameters describing acoustics of the speech semi-continuous models are more efficient in decoding (faster) but have less accuracy with unlimited training data. With limited training data semi-continuous models could be more accurate due to smaller number of parameters to train.

Sphinx4 doesn't support semi-continuous models, only pocketsphinx supports them.

and how to decide which model to use?

It depends on what you are going to do, you need to provide more details in order to get answer on your question.