The overall design consists of a multithreading system that uses Google's Speech to Text API.
One thread adds the audio files to the thread-safe queue and two consumer threads execute the audio transcription synchronously. Currently supports MP3 audio files of 1 minute of length and 44000 Hz sample rate. All queued items have same priority.
Next steps:
- In order to find the number of API calls to run simultaneously, I will perform a CPU analysis to find the bottleneck of the system when calling all APIs.
- speech-to-text.py: Multi-threaded audio transcription
- threader.py: Multi-threading exercise
- processes.py: Multi-processing exercise