Add STTMuteFilter to un/mute the STT

[markbackman]

Please describe the changes in your PR. If it is addressing an issue, please reference that as well.

This adds a new processor called STTMuteFilter, which can mute the STT service while the bot is speaking. In this context muting means:

• Don't send the audio frame to the STT to be transcribed
• Prevent the interruption frame from passing through the pipeline

This processor is intended to be added to the pipeline before the STT service.

To help accomplish this task, I've added a new frame type, STTMuteFrame, which is emitted by the STTMuteFilter based on the processor settings.

The processor can be configure to:

• Never mute when the bot speaks (default)
• Mute only the first time the bot speaks
• Mute every time the bot speaks
• Provide custom logic for when the bot should speak based on a callback

I've included foundational example 24-stt-mute-processor.py to show how to use this.

[markbackman]

@aconchillo I'm not sure how to classify this processor. I stuck it with filters but that doesn't feel right.

[aconchillo]

This is awesome! Just added a few comments.

[markbackman]

Ok! All changes are made. Do you think this belongs in pipecat.processors.filters or somewhere else?

[aconchillo]

Ok! All changes are made. Do you think this belongs in pipecat.processors.filters or somewhere else?

Thanks! That's an always hard question. If it was a filter we would call it stt_mute_filter.py in which case the class would be STTMuteFilter. The concept of a filter is to stop or passthrough certain frames, and that's what this processor is doing, so it might make sense to have it there, so we should rename it STTMuteFilter probably as all the rest.