Posted: Fri 11 Oct 2013, 19:06
Yes, thanks for the reply H4LF82. I think one of the difficulties in getting a voice driven puppy working is that there are clearly many different users, many degrees of vision impairment, and many different intended actions each user wants to perform on their PC.
Much of what you described is well beyond anything I could conceive of - eg: the idea of being able to change system fonts via voice control (and without being able to see the screen) is something I can only imagine will be YEARS away. However - some of the actions a user would want to perform DO seem possible with the current technology - as long as we assemble the building blocks correctly.
I am very interested in getting basic TTS and STT working together (it would be a handy addition to the functionality in Switchpup) and if you have the time / patience to advance an STT project I would welcome your assistance in determining what types of STT are helpful and what is not.
To that end I would like to suggest a very basic project as a way of testing and assembling the building blocks: I would like to make it possible to do the following:
1) Once the Puppy has booted I want it to speak to me: eg "Would you like me to play your music collection randomly?"
2) Another module (sphinx?) would listen to my answer and probably turn it into text (could probably skip this step but I include it anyway as part of the project...)
3) Another module (Simon?) would act on that text and randomly start playing my Music folder (via Peasymp3 or similar)
In this basic project I have not even suggested ncluding any form of STT filemanager to help me hunt for the music folder - that would be an even bigger project and would be a natural follow-on if the basic project works. At first we would just rely on having the Music folder in one place and getting the basics working. Your idea of a verbally controlled volume / mute function would be a great addition too.
I think such a basic project would be a good way to get to first base. Would you be interested in helping test this sort of thing?
Much of what you described is well beyond anything I could conceive of - eg: the idea of being able to change system fonts via voice control (and without being able to see the screen) is something I can only imagine will be YEARS away. However - some of the actions a user would want to perform DO seem possible with the current technology - as long as we assemble the building blocks correctly.
I am very interested in getting basic TTS and STT working together (it would be a handy addition to the functionality in Switchpup) and if you have the time / patience to advance an STT project I would welcome your assistance in determining what types of STT are helpful and what is not.
To that end I would like to suggest a very basic project as a way of testing and assembling the building blocks: I would like to make it possible to do the following:
1) Once the Puppy has booted I want it to speak to me: eg "Would you like me to play your music collection randomly?"
2) Another module (sphinx?) would listen to my answer and probably turn it into text (could probably skip this step but I include it anyway as part of the project...)
3) Another module (Simon?) would act on that text and randomly start playing my Music folder (via Peasymp3 or similar)
In this basic project I have not even suggested ncluding any form of STT filemanager to help me hunt for the music folder - that would be an even bigger project and would be a natural follow-on if the basic project works. At first we would just rely on having the Music folder in one place and getting the basics working. Your idea of a verbally controlled volume / mute function would be a great addition too.
I think such a basic project would be a good way to get to first base. Would you be interested in helping test this sort of thing?