ASR optimization

Hello,

I am quite new to speech recognition.
I have some questions related to recognition tuning and optimization
techniques.

Where could I find an introduction to the following techniques:
- Accoustic model retraining | adaptation
- Hand-tuned pronunciations
- Utterance transcription
- Audio caching & streaming

Is there any other recognition optimization techniques?

What is the improvement usually measured after optimization?

Thanks.
Best regards.

Laurent