Weekly Report 2024-11-01

[Jim-Hutchinson]

Removed popup message functionality as it is not helpful for a visually impaired user and cluttered the rest of the code.

Began implementation of Piper TTS for AI voice synthesis instead. The model is used by NVDA, and was presented for use in a very similar application to my own, which is how I found it. Image captioning for the visually impaired

Support for AMD hardware, which I use, is not currently in the main version of Piper TTS. However, a fork does exist, so I will be proceeding with that version.

[Jim-Hutchinson]

I'm having issues installing Piper TTS. I may need to change python versions to 3.10.12. If that does not work, I will find a different test to speech model to use instead.

[Jim-Hutchinson]

I ended up setting up Silero TTS with the v3_en model from here. The model offers multiple speakers and is able to run on my hardware with reasonable speed, even on my laptop. When hosted on a more powerful machine, I hope to make the speech near real-time.