Text to speech using azure cognitive services

edgeimpulse / notebooks

Notebooks using the Edge Impulse libraries

4 stars 1 forks source link

Text to speech using azure cognitive services #37

Closed danuw closed 1 year ago

danuw commented 1 year ago

This is to add alternative to get audio samples for keyword spotting generated using Text-to-Speech Azure Cognitive Service.

It mirrors the same structure as the existing version for Google TTS.

FYI: The voices option default to available voices with different styles as combination of language, pitch gender returned many audio that seemed exactly the same.

Hope that helps some and maybe allows for event more sample data (combining Google and Azure TTS services)

dagshub[bot] commented 1 year ago

Join the discussion on DagsHub!

ShawnHymel commented 1 year ago

@danuw Thank you for the contribution! Looks like one of the tests is failing. Please:

Remove the "03-" prefix from the filename (as we removed those from all files)
Fix the issue with the repository not passing the tests. From this, I believe you need to clear all cell outputs and commit again.

danuw commented 1 year ago

Hi workflow is awaiting approval.

(FYI I tried to clean manually but ended up undoing some of my changes and using

jupyter nbconvert --ClearOutputPreprocessor.enabled=True --inplace notebooks/generate-keyword-spotting-dataset-azure.ipynb

from https://saturncloud.io/blog/how-to-clear-jupyter-notebooks-output-in-all-cells-from-the-linux-terminal/#clearing-jupyter-notebooks-output-from-the-linux-terminal (provided to me by https://github.com/Willmish

Notebook was also renamed. )

Thank you @ShawnHymel (btw love you videos :) )