Call to arms!.....

B

blackneos940

Guest
Okay, so this guy, who runs a restaurant, thinks that a Program that reads input via voice, like Siri, and then puts it on a display, like "Spaghetti, Marinara Sauce, Pepsi, Fish, Roasted Butterfly", so that it would be easier to input things like orders 'n stuff, (ESPECIALLY when a lot of customers are ordering), would be REALLY useful..... :) He said I should make it Copyrighted so people wouldn't steal it, and make money off of it, but then I told him, "If a whole community of people work on this, who will be able to steal it?", or something to that effect..... :3 Anyway, since I'm startin' this project, it is, of course, Free and Open Source..... :D Now, since I suck at Programming in general, and can't fathom how in Linus' name one would take a Programming Language and make it into a Microphone, my question is this: Should we do this in Python, C++, C, or another Language.....? :) Also, it would mark the first ever project on Linux.org...... :D I shall call it: FNU (Food's Not UNIX)..... :D Or perhaps Clambake.....? :3 Now then.......... Let's get this party van rollin'!!..... :D
 


Writing a speech recognition library that could perform voice to text would NOT be a trivial thing to attempt. It would be an extremely complicated and time consuming project. I think there are several existing, free/open source packages available for Linux which can do voice to text. BTW: I've never messed with anything like that, so I'm not familiar with any existing engines, but I'm pretty sure there are a few available. How complete, or effective they are - again, I have no idea!

So I'd say your first step should be to do some research online and try to find an existing library that you can use in your program to convert voice input to text output. That would save you years of development and make your life a whole lot easier! :)

Once you have found a suitable voice to text engine; the next step would be to create a GUI program which can capture voice input from a microphone (as a temporary WAV or PCM file); pass that to the speech recognition library, which will process and translate the voice input and hopefully, pass some sane text back to the application.

As for the programming language used to implement your program - that would depend on a number of things:
- The speech to text engine you are using;
- What language it was originally written in;
- The platform/s you are targeting your application to run on;
- The GUI/widget library you are using (QT, wxWidgets, GTK etc)
- Whether there are any bindings available that will allow you to interact with the library/engine using other programming languages.
etc. etc.

For example, if the library was written in C, but there are no other language bindings; you're stuck with a choice of using C or C++ (even if there are no C++ bindings/wrappers, you could still use C++!). In this situation, if you REALLY wanted to use Python (or any other language), but there are no existing bindings; then you would have to create your own python (or other language) module containing the bindings. Which depending on the amount of functionality available to you; could be a huge job, or it could be relatively simple.

Saying that, if you know exactly which parts of the engines functionality you will need to use in your program, you could just create python (or other language) bindings for the parts of the functionality you will actually be using!

But that's something for further down the line. You can decide on the best implementation language once you have found a voice to text engine that works well enough for your applications needs. That would be my first recommendation!
 
Oh man..... This stuff is WAY over my head....... :( Maybe I should just quit..... I don't even know........
 
I take it the guy who wants this is working in the back of the restaurant? Having been a waitress in the past (way back before computer input and we had to write the tickets) all I can say is "What the heck is he thinking?"! There are way too many variables that would have to be sorted out, like background noise: if someone said "Hey Christi..." while I was inputting an order would be printed on the ticket. The back of the restaurant is way too busy for that. Then you have dialect, other people's accents (ever read a google voice transcript?) to put into consideration.

I am not saying it couldn't be done, it would just be a nightmare in practice to use it.

BUT - I would be happy to join your team if you decide to create it. ;)
 
blackneos940 said:
Okay, so this guy, who runs a restaurant, thinks that a Program that reads input via voice, like Siri, and then puts it on a display, like "Spaghetti, Marinara Sauce, Pepsi, Fish, Roasted Butterfly", so that it would be easier to input things like orders 'n stuff, (ESPECIALLY when a lot of customers are ordering), would be REALLY useful..... :) He said I should make it Copyrighted so people wouldn't steal it, and make money off of it, but then I told him, "If a whole community of people work on this, who will be able to steal it?", or something to that effect..... :3 Anyway, since I'm startin' this project, it is, of course, Free and Open Source..... :D Now, since I suck at Programming in general, and can't fathom how in Linus' name one would take a Programming Language and make it into a Microphone, my question is this: Should we do this in Python, C++, C, or another Language.....? :) Also, it would mark the first ever project on Linux.org...... :D I shall call it: FNU (Food's Not UNIX)..... :D Or perhaps Clambake.....? :3 Now then.......... Let's get this party van rollin'!!..... :D
Click to expand...
I personally would not want to tackle this project. I have to wonder if a product, possibly closed source already exists?

I have used Dragon Naturally Speaking in the past when I was using a Windows computer more than now. If not they have a Developers Kit SDK. Someone somewhere may already be working on a similar project.
 
blackneos940 said:
Yeah, but this guy uses an iPad, and wants it for that...... And I don't even know how to begin to CONVERT it to that..... :(
Click to expand...
Well, you can recompile the sources, but you're going to need an ios development environment. I don't know how different ios is from Android or Linux, or even MacOS. I've done MacOS work, and it is *mostly* compatible with GNU/Linux, with some gotchas. Android may have a Linux kernel, and some familiar libraries, but there's a big difference between GNU/Linux and Android, even when working with the Android NDK.

Unless you've got a lot of experience, I would avoid doing cross-development work...

Also, as was pointed out in another response, background noise can be a big problem. You can filter input somewhat, for example using a noise gate, but in a noisy environment the user has to be pretty close to the microphone for useful pickup. You can't just ramp up gain, as that will also pickup more background noise.
 
Well, get t
ChristiW said:
I take it the guy who wants this is working in the back of the restaurant? Having been a waitress in the past (way back before computer input and we had to write the tickets) all I can say is "What the heck is he thinking?"! There are way too many variables that would have to be sorted out, like background noise: if someone said "Hey Christi..." while I was inputting an order would be printed on the ticket. The back of the restaurant is way too busy for that. Then you have dialect, other people's accents (ever read a google voice transcript?) to put into consideration.

I am not saying it couldn't be done, it would just be a nightmare in practice to use it.

BUT - I would be happy to join your team if you decide to create it. ;)
Click to expand...

Well, get this..... :) He OWNS it..... :D So I guess it's surprising to hear him say that, but, he probably has so few workers, that HE'S the one of the people who would use this sort of thing..... :3 Too many variables, huh.....? :( Well..... I may need more time to think about it, but if I decide, then I'll let you know!..... ^^ Maybe we could even do it in C++ as well!!..... ^^
 
blackneos940 said:
Okay, so this guy, who runs a restaurant, thinks that a Program that reads input via voice, like Siri, and then puts it on a display, like "Spaghetti, Marinara Sauce, Pepsi, Fish, Roasted Butterfly", so that it would be easier to input things like orders 'n stuff, (ESPECIALLY when a lot of customers are ordering), would be REALLY useful..... :) He said I should make it Copyrighted so people wouldn't steal it, and make money off of it, but then I told him, "If a whole community of people work on this, who will be able to steal it?", or something to that effect..... :3 Anyway, since I'm startin' this project, it is, of course, Free and Open Source..... :D Now, since I suck at Programming in general, and can't fathom how in Linus' name one would take a Programming Language and make it into a Microphone, my question is this: Should we do this in Python, C++, C, or another Language.....? :) Also, it would mark the first ever project on Linux.org...... :D I shall call it: FNU (Food's Not UNIX)..... :D Or perhaps Clambake.....? :3 Now then.......... Let's get this party van rollin'!!..... :D
Click to expand...

I am about to release the next version of my Betabot system, which I have completely re-wrote and copyrighted. It is an artificial intelligence chatbot. However, it lacks speech recognition software. Users must type their response. My Betabot system may be a good platform to start with or inspire you.

Betabot version 1.1b will be released within one to three weeks (unless something goes wrong). The software will be free for download at http://dcjtech.info/forum/downloads/betabots/ and will have an open API for the knowledgebase which is open-source under LGPLv3. Betabot v1.1b is 100% pure Python3.4 including the databases which are either Python tuples, dicts, or lists.
 
DevynCJohnson said:
I am about to release the next version of my Betabot system, which I have completely re-wrote and copyrighted. It is an artificial intelligence chatbot. However, it lacks speech recognition software. Users must type their response. My Betabot system may be a good platform to start with or inspire you.

Betabot version 1.1b will be released within one to three weeks (unless something goes wrong). The software will be free for download at http://dcjtech.info/forum/downloads/betabots/ and will have an open API for the knowledgebase which is open-source under LGPLv3. Betabot v1.1b is 100% pure Python3.4 including the databases which are either Python tuples, dicts, or lists.
Click to expand...

Wow..... :3 Devyn to the rescue again!!..... :D I'll take a look at it for inspiration, and if I come up with anything, you bet your sudo I'll put your name in the comments!!..... :3
 
I like Sourceforge....... :3 But, it turns out it uses Windows as it's Server..... Ech..... :< Windows is a Swiss Army Knife..... :) A rusty, cracked, bent, broken, dull, worn-out Swiss Army Knife..... :D
 
DevynCJohnson said:
I am about to release the next version of my Betabot system, which I have completely re-wrote and copyrighted. It is an artificial intelligence chatbot. However, it lacks speech recognition software. Users must type their response. My Betabot system may be a good platform to start with or inspire you.

Betabot version 1.1b will be released within one to three weeks (unless something goes wrong). The software will be free for download at http://dcjtech.info/forum/downloads/betabots/ and will have an open API for the knowledgebase which is open-source under LGPLv3. Betabot v1.1b is 100% pure Python3.4 including the databases which are either Python tuples, dicts, or lists.
Click to expand...
I... I can't find it..... :(
 
That's because it hasn't been released yet. :)
In his previous post, Devyn said:
DevynCJohnson said:
Betabot version 1.1b will be released within one to three weeks (unless something goes wrong). The software will be free for download at http://dcjtech.info/forum/downloads/betabots/ and will have an open API for the knowledgebase which is open-source under LGPLv3. Betabot v1.1b is 100% pure Python3.4 including the databases which are either Python tuples, dicts, or lists.
Click to expand...

I imagine that Devyn will announce its release once it has been uploaded and you can check it out then!
 
JasKinasis said:
That's because it hasn't been released yet. :)
In his previous post, Devyn said:


I imagine that Devyn will announce its release once it has been uploaded and you can check it out then!
Click to expand...

... ... ... ... Oh..... :D I didn't even NOTICE that!!..... :D Sorry, Devyn!!..... :) Also, thanks for pointing it out to me, buddy!..... :) BTW, I'm listenin' to an old UNIX film on me Surface..... :) How can I feel Nostalgic for something I wasn't around in it's heyday for.....? :( Yarrrrr..... :3
 
Tuurns out GPUs are usable for all kinds of parallelizable computations, not just graphics. It's really a generalization of what they're designed to do, after all...
 
JasKinasis said:
That's because it hasn't been released yet. :)
I imagine that Devyn will announce its release once it has been uploaded and you can check it out then!
Click to expand...

Thanks @JasKinasis .;)

I hope to release the next version next week or the week after. I am currently doing the "last minute" tests and debugging. This version is planned to work on Linux, Windows, and Unix systems (past versions were Linux-only). This version is significantly faster than the previous.
 
You must log in or register to reply here.

Staff online

Members online

Total: 1,065 (members: 8, guests: 1,057)

Latest posts

Top