jitsi / gsoc-ideas

Google Summer of Code ideas
60 stars 29 forks source link

GSOC -22, Speech to Text (ISSUE) #20

Closed shivamsingh124 closed 2 years ago

shivamsingh124 commented 2 years ago

Hi Jitsi team, I have used Artyom Speech to text API in past which can be very well integrated in this project, I am attaching a video demo of the working prototype as well.

Why I am using artyom

Working prototype:-

https://drive.google.com/file/d/15lmaRUiFTYdGFd-yCguBwRnsAr0I9J2d/view?usp=sharing

shivamsingh124 commented 2 years ago

Link to issue :- https://github.com/jitsi/gsoc-ideas/blob/master/2022/speech-to-text.md

Please assign this issue to me so that I can start working and making PR for the same.

nikvaessen commented 2 years ago

Applications for GSoC should be made via the program website. Also Artyom.js seems to be a javascript wrapper around the browser's webkitSpeechRecognition API. Note that this is not 'free' [1] as in libre software [2]:

On some browsers, like Chrome, using Speech Recognition on a web page involves a
server-based recognition engine. 
Your audio is sent to a web service for recognition processing, so it won't work offline.

[1] = https://developer.mozilla.org/en-US/docs/Web/API/SpeechRecognition [2] = https://www.gnu.org/philosophy/free-sw.en.html

shivamsingh124 commented 2 years ago

@nikvaessen Got it!. So I should find an alternate that can work on offline mode also?

nikvaessen commented 2 years ago

No, it does not need to work offline, we want a free solution. See the project description:

| The current speech-to-text implementation uses the paid, proprietary Google speech-to-text API. It would be nice to offer a free, privacy-friendly alternative.

shivamsingh124 commented 2 years ago

@nikvaessen Got it, I thought of working on our own ML model but it is not the case, I will look for a free alternative and will let you know at the earliest