Speech to text conversion for non-english language

问题

I am trying to implement naive speech to text conversion for non-english language. where user speaks in other language and text is also in the same language .Is that possible ? If yes can anyone help with the idea of how to implement it ? Any Help would be greatly appreciated.Thanks in advance.

回答1:

It is unlikely any commercial speech recognition solution will support Sanskrit, so the only choice you have is to add support for Sanskrit into open source engine like CMUSphinx.

This is pretty straightforward, you actually just need to follow the documentation and you can get to the point. You also need to have a knowledge of the scripting language which will help you to cut manual work on some steps.

Read Introduction - CMUSphinx Tutorial introduction to become familiar with concepts of speech recognition - features, acoustic models, language models, etc.
Try CMUSphinx with US English model to understand how things work. Try to train with sample US English AN4 database.
Read about your language in Wikipedia
Collect a set of transcribed recordings for your language - interviews, audiobooks or record them yourself.
Based on the data you collected, create a list of words and a phonetic dictionary. Most phonetic dictionaries could be created with a simple rules with a small script in your favorite scripting language like Python. See Generating a dictionary for details.
Segment the audio to short sentences manually or with sphinx4 aligner, create a database with required files as described in training tutorial Training Acoustic Model For CMUSphinx
Integrate new model into your application and design a data collection to improve your model.

If you have questions, feel free to ask. This paper also might be of help for you.