LinkBack URL
About LinkBacks
Reply With QuoteWhat happens when the music is really loud and you cant scream over it?
What happens if someone singing says something like "Stop"?
My car project has been more software than hardware. I've been spending quite some time developing a music voice control system. There is no need for an LCD as everything loads automatically when I turn on the car. Music is synced via USB Key and the USB port is stowed away in what was the ashtray. Also, the software listens to everything being said in the car so that no buttons need to be pushed. It's able to read both MP3 and WMA tags. I've also programmed in a dynamic playlist feature, so I can listen to "my favorite songs" seperate from regular artist, album, genre, and song requests. All songs are randomized during each album, artist, genre, or favorites request and the system uses TTS to provide me voice feedback. The PC is connected to an Alpine CDE-9845 and the mic is installed in the driver side pillar. I have a few pics and videos at my site. As it is an on-going project, I'll post more as it comes along.
http://www.radiooncommand.com
What happens when the music is really loud and you cant scream over it?
What happens if someone singing says something like "Stop"?
Post Whore's tip:
User CP-> Edit Options-> Number of Posts to Show Per Page-> 100
http://www.frappr.com/mp3car from this thread http://www.mp3car.com/vbulletin/general-mp3car-discussion/65044-frappr.html
Good questions... to answer the first, I have just recently updated the voice recognition engine which has helped this problem. Listening to the music at a moderate level still allows me to pull everything up just fine. I don't typically play my music really loud, so it works for me. Also, if I am playing it loud, I'll just turn it down briefly in order to pull up my next request. A uni-directional mic also helps with this. And, it does a good job not backing up to prevent further recognition. As for the second question, I have to say "Stop Song" in order to trigger the command. I've had it running in my car for about 4 months now (old code) and a few days with the new code and it has only triggered a core command randomly, because it thought that I stated it, less than 10 times.
That's exactly what I plan on doing in my truck in the future, having everything voice controlled. Nice work.
Do you plan on making the software available? If so, do you have an idea of how long it might be? Also, if you need anyone to test out the software before releasing it, I'd be interested.
Take 2 - '03 Audi A4
macmini with an 8" widescreen (in progress)
My 2004 Ford Ranger:
14" LCD w/ custom made aluminum bezel
Completly custom computer case
status: Totaled
I have spent a good amount of time working on the software and plan on making in available for download. As for a time frame, I would have to say my goal would be sometime this summer. That would give me enough time to create a nice installer and setup instructions as well as fix a few bugs, such as the recognition delays after adding wma support. I'll keep you in mind for any testing that might be needed before the release, thanks for offering.
I have fixed the wma delay issue and have been adding a UI, which would make it easier to see the music that has been added. It will be changed to look better, but it's fine for testing right now. The software can still be configured to run in a headless config (my setup).
The picture was taken after the "Play Song" command. The list will change based on what you ask for. For example, by asking for "Play Genre", you will be displayed a list of all Genres related to the stored music. The Mic represents the current state of the Mic, which will automatically turn on/off based on TTS feedback.
Software will be available shortly for download.
The GUI has been updated and yes I know, it's not that great. I only have a few minor bugs to work out and would like to add some bells and whistles to the listbox. Hoping to have the first release done soon. It will be pretty rough, but will get better over time.
One feature that I would like to add is that let say it doesn't pick up what you are asking for, from the listbox you will be able to highlight the item it does not recognize and then there will be a Fix button in order to train the the selected item. This would allow for increased accuracy for future requests.
New Screenshot (After Play Genre Request):
http://www.radiooncommand.com
What language are you developing this in?
Originally Posted by 0l33l
I'm using Visual Basic 2005 Express Edition: FREE . As for the voice recognition, I'm using the Naturally Speaking 8 Standard's Engine. The only piece of software that's not free. I got mine for about $50 with rebate. I started with VB6 and DNS 7, but have been porting over to keep up with what is out there.
http://www.radiooncommand.com
Hey man, this looks and sounds like a really interesting project. If you are looking for beta testers let me know - otherwise, I'll keep an eye out for when you release this.
Keep up the good work.
www.mobile-effects.com
Free file hosting, picture gallery hosting for installs, PM me.
Internet's first Front End Skin browser, featured installs, downloads, links, informative articles - all free to registered users.
Posting Permissions
Bookmarks