[dev] Re: What would a well-designed voice assistant program look like?

From: LM <lmemsm_AT_gmail.com>
Date: Thu, 22 Mar 2018 14:11:59 -0400

Laslo Hunhold wrote:
>When reflecting about a problem, it should be a priority to think about
>how it can be split up. In this case, we can talk about two problems,
>which combined, solve the "voice assistant" problem.
>
> 1) voice recognition
> 2) assistant

Personally, I'm more interested in the speech to text and text to
speech parts of the project. I've talked to several people who want
an assistant that can copy the behavior of the commercial products and
let users add their own functionality to extend it. I happen to like
the idea of an expert system with the use case constrained to a
particular problem over a program that tries to do everything. One
thing I was considering that would probably work fine offline is to be
able to ask for the meaning of a word (rather than typing the word
into a GUI interface) and have the program speak (as well as display)
the definition. Might also be helpful for people with accessibility
issues. Could probably use a stardict compatible backend or some
other dictionary format and library to look up the word definitions.

Are there any Open Source voice recognition or text to speech
libraries someone on the list would recommend? At the moment I'm
investigating kaldi and pocketsphinx. I also read that Mozilla has
come out with Deepspeech based on Tensorflow.

Thanks.
Received on Thu Mar 22 2018 - 19:11:59 CET