What are the differences between the English-US acoustic model that comes with PocketSphinx, and the model in the file "cmusphinx-en-us-ptm-5.2.tar.gz" ? Both are PTM, both recorded in 16k, but tha later is much bigger. In which situations should I use it? And how - simply tell pocketsphinx_continues.exe the relevent dir?
I'm using PocketSphinx on a desktop computer, in grammar mode and very small dictionary.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
What are the differences between the English-US acoustic model that comes with PocketSphinx, and the model in the file "cmusphinx-en-us-ptm-5.2.tar.gz" .
This is the same model, the one that comes in pocketsphinx is simply a compression version for mobile, the one in downloads contains additional files.
1) Mdef file is converted to binary format with pocketsphinx_mdef_convert
2) Mixture_weights file removed, it is only required for adaptation
And how - simply tell pocketsphinx_continues.exe the relevent dir?
pocketsphinx_continuous -hmm <acoustic_model_dir>
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
What are the differences between the English-US acoustic model that comes with PocketSphinx, and the model in the file "cmusphinx-en-us-ptm-5.2.tar.gz" ? Both are PTM, both recorded in 16k, but tha later is much bigger. In which situations should I use it? And how - simply tell pocketsphinx_continues.exe the relevent dir?
I'm using PocketSphinx on a desktop computer, in grammar mode and very small dictionary.
This is the same model, the one that comes in pocketsphinx is simply a compression version for mobile, the one in downloads contains additional files.
1) Mdef file is converted to binary format with pocketsphinx_mdef_convert
2) Mixture_weights file removed, it is only required for adaptation
Thanks.