Regarding Acoustic Model Merging

[pramod]

Hi ,

I am currently working for hindi speech recognition , I followed all the steps mentioned in the CMU Sphinx

Building the dictionary
Building the language model
Building the acoustic model

i have done above steps and tested with the below comand
**pocketsphinx_continuous -hmm en-us-adapt -lm en-us.lm.bin -dict cmudict-en-us.dict -infile input_audiofile > simple.txt
**
I am getting the output as am expected .

Now I have Doubt regarding the Acoustic model

i have built the acoustic model for hindi and english acoustic model i took from the CMU Sphinx

Is there any possibility to merge the both acoustic models , if that merging is done we can say the model as multi language acoustic model .

As i found language model can be merged and i did that .

Please let me know if there is anyway to merge the acoustic models

Thanks in advance

Regards
V.Pramod

[Nickolay V. Shmyrev]

Is there any possibility to merge the both acoustic models , if that merging is done we can say the model as multi language acoustic model .

No, there is no such tool.

Please let me know if there is anyway to merge the acoustic models

You have to implement this thing yourself or you can train multilingual model from both Hindi and English data.