Feature Request: Support for Readable Streams for Synthesizing Text-to-Speech

### Is your feature request related to a problem? Please describe.

Modern streaming APIs such as Open AI's latest ChatGPT API require the ability to generate text-to-speech as the text is still being fed into the Readable Stream. Currently, the Google Cloud Text-to-Speech API does not support Readable Streams for text-to-speech synthesis.

### Describe the solution you'd like

I would like to request support for Readable Streams to be added to the Google Cloud Text-to-Speech API for synthesizing text-to-speech. With this feature, users can generate text-to-speech as they are still feeding text data into a Readable Stream.

### Describe alternatives you've considered

As an alternative solution, users can wait until all of the text data is fed into the Readable Stream before generating the text-to-speech. However, this is not optimal for streaming applications where real-time text-to-speech generation is needed.

### Additional context

Adding support for Readable Streams for text-to-speech synthesis is essential for interfacing with modern streaming APIs like Open AI's latest ChatGPT API. This will allow for real-time text-to-speech generation as the data is being streamed, enabling more seamless integration with streaming applications.

It would also be useful if the text-to-speech synthesis for streams could indicate which word was being spoken at the time of the audio. This would allow for more accurate audio transcription and for the applications to indicate visually (e.g. underline, highlight, etc.) which words are being spoken for narration and better accessibility support.