Running StreetDeck under Windows Vista

**terrygold** · 10-19-2006, 08:44 PM

The StreetDeck 1.0.5.7 BETA release makes it possible to run StreetDeck under windows Vista RC2. After installing StreetDeck I realized that Yahoo! Music Jukebox checked the operating system version and refused to install under Vista. There's an easy way to get it to install.

Save the Jukebox install executable on your computer. I just did it and the name for the executable was ymjsetup. Right click the application and select Properties. Under Compatibility Mode, select "Run this program in compatibility mode." Use the default of Windows XP (Service Pack 2)

Double click ymjsetup and the application should go through the normal installation process.

Before starting StreetDeck, you'll need to restart your computer and start Yahoo! Jukebox and login for the first time. That should get you going.

**debugger** · 10-21-2006, 05:37 PM

mms://wm.microsoft.com/ms/inetpub/ch...ecognition.wmv

Do you think its possible to use "start listening" feature in SD instead of having to do gesture an "A" or something....

**god_of_cpu** · 10-21-2006, 06:45 PM

Originally Posted by debugger

mms://wm.microsoft.com/ms/inetpub/ch...ecognition.wmv

Do you think its possible to use "start listening" feature in SD instead of having to do gesture an "A" or something....

No, you need a trigger event so it mutes your playing media and because there is a ton of other ambient noise while driving.

**debugger** · 10-21-2006, 08:31 PM

Originally Posted by god_of_cpu

No, you need a trigger event so it mutes your playing media and because there is a ton of other ambient noise while driving.

How about a plugin for a mic that will substract music and background noise from the mic input. I've been working on noise cancellation software for a while now for incar environment.
The biggest problem is the processing power and the use of multiple microphones.

I think the best option is an external button on the steering wheel using Grinder pl.

**terrygold** · 10-21-2006, 09:00 PM

On my pioneer AVIC-Z1, I had a button on my steering wheel that would trigger the "start listening" of the unit. Not as nice as just saying the "start listening" phrase but it worked OK. (Which was about all that did with the Pioneer speech recognition.)

**god_of_cpu** · 10-22-2006, 12:02 PM

Originally Posted by debugger

How about a plugin for a mic that will substract music and background noise from the mic input. I've been working on noise cancellation software for a while now for incar environment.
The biggest problem is the processing power and the use of multiple microphones.

I think the best option is an external button on the steering wheel using Grinder pl.

The DA-350 is an array microphone that already subtracts background noise. http://store.mp3car.com/ProductDetai...Code=COM%2D017 Its not perfect, nothing is, if your music is loud enough its going to pick it up too, but it's still by far the best microphone I've ever used (and damn well better be for $300) it works about 10x better then any other microphone I've used in the car and if your going to do voice recognition is the only thing I'd bother recommending and the only thing that is reasonably usable with the windows down or even up for that matter.

I don't think you can do noise cancellation any better then the andrea mic without tuning it specifically to only one car and one audio system.

You can already start voice reconition by a button. Just bind an event from your girder plugin to eFunctionToggleVoiceMode.

**debugger** · 10-22-2006, 06:37 PM

Originally Posted by god_of_cpu

You can already start voice reconition by a button. Just bind an event from your girder plugin to eFunctionToggleVoiceMode.

Yeah, thats the best way to do it. That mic you are talking about is awesome! As the matter of fact two of those would work great for my setup. I'm gonna try to make a trigger even to eFunctionToggleMode when I say: "Start listening". We going to use one-sided autocorrelation linear predictive coding and robust similarity measuring techniques. Any human voice is in the certain frequency range. Anyways like you said it will only work for a particular vehicle and a person. It will take a long time to train the system and is almost as hard as coding an AI engine.

One other solution I was thinking of is to simply replace "start listening" to a human whistle. It would be less stress on the cpu and the mic will only listen for a certain frequency and amplitude.