Thread: Google's updated voice app for Android makes cars jealous

Looks great. Will need a data plan to take full advantage, but still nice nonetheless.

Would be great if Voice Actions could be interfaced with our carPC's. Possibly a RideRunner, openMobile or Centrafuse plugin where the driver (or passengers) would use their Android phone to control their frontend (i.e. next song, play video, pause, play Bruce Springsteen, etc). Then our Android phones would be the a new input device.

I've still yet to get android voice working in-car. It works great at home, or anything outside of a car... but you introduce road and engine noise... and it just doesn't work :/

Originally Posted by malcom2073

I've still yet to get android voice working in-car. It works great at home, or anything outside of a car... but you introduce road and engine noise... and it just doesn't work :/

Yea that's definitely a problem. Having the bluetooth headset/mic near your mouth would hopefully help, but I'm sure there will be some problems. I haven't tried it personally.

If I understand it correctly, Google translates the voice on their servers so it all depends on how their speech recognition handles the ambient noise. I realize your in-car experience has not been great, but the video indicates that Google wants you to use Voice Actions in the car so hopefully they can filter out all the noise and take this into consideration.

If voice didn't work, I would be just as happy with a regular android app that works over bluetooth. There's one for XBMC that works (over Wifi) very well:

http://www.youtube.com/watch?v=gY-r-POY3yY

I'm missing the point...

Why would you want to use a phone as a voice recognition device, have that voice sent all the way to google's servers, translated, and then sent all the way back? A carPC can do everything on that list locally using a microphone designed for in-car use.

Originally Posted by justchat_1

I'm missing the point...

Why would you want to use a phone as a voice recognition device, have that voice sent all the way to google's servers, translated, and then sent all the way back? A carPC can do everything on that list locally using a microphone designed for in-car use.

I agree, it seems a bit redundant. But I think there are still applications for it. In my earlier comments, my initial thoughts was that Voice Actions for Android could compliment a persons carPC which is already equipped with voice recognition (I still remember your openMobile voice videos btw). So if someone has an Android phone then they could still use it to control their carPC OR use their standard microphone while in the car. For me personally, if I were to setup voice-recognition for my carPC, I would choose to buy/use an Android plugin rather than buy an Andrea mic. It would be a smaller investment to gain the same functionality (if it could be replicated). But that's just my situation.

But other applications may still exist. A person could pass his/her phone to the backseat passenger and let them control the carPC. It's possible that these backseat passengers will not have an Andrea Mic close by to listen to their commands. Same would apply for passengers in the third row seats as well. Plus, it could still have applications if you're not even in the car, but still within BlueTooth range.

Again, I haven't given this a lot of thought, but I think the carPC community could take advantage of it in some way.

Originally Posted by justchat_1

I'm missing the point...

Why would you want to use a phone as a voice recognition device, have that voice sent all the way to google's servers, translated, and then sent all the way back? A carPC can do everything on that list locally using a microphone designed for in-car use.

Bcause you can use your phone in more places than your car pc and Google has the strength and motive to put some very sophisticated recognition algorithms behind their translate service.

Originally Posted by matt11601

I agree, it seems a bit redundant. But I think there are still applications for it. In my earlier comments, my initial thoughts was that Voice Actions for Android could compliment a persons carPC which is already equipped with voice recognition (I still remember your openMobile voice videos btw). So if someone has an Android phone then they could still use it to control their carPC OR use their standard microphone while in the car. For me personally, if I were to setup voice-recognition for my carPC, I would choose to buy/use an Android plugin rather than buy an Andrea mic. It would be a smaller investment to gain the same functionality (if it could be replicated). But that's just my situation.

But other applications may still exist. A person could pass his/her phone to the backseat passenger and let them control the carPC. It's possible that these backseat passengers will not have an Andrea Mic close by to listen to their commands. Same would apply for passengers in the third row seats as well. Plus, it could still have applications if you're not even in the car, but still within BlueTooth range.

Again, I haven't given this a lot of thought, but I think the carPC community could take advantage of it in some way.

A directional array microphone will always have better noise reduction properties then a cell phone but that "pass the mic" idea is a cool one.

Originally Posted by Bugbyte

Bcause you can use your phone in more places than your car pc and Google has the strength and motive to put some very sophisticated recognition algorithms behind their translate service.

Apparently you've never used google voice lol... They use nuance for their voice servers which isn't exactly cutting edge just industry standard. Their homebrew stuff (used for google voice) is horrific in its translation quality, probably quite a few years before its good enough to even be in the same league as its competitors.

Anyway, my point wasn't why is this significant to cell phones (which it certainly is) it was why is this significant to cars.

Originally Posted by justchat_1

A directional array microphone will always have better noise reduction properties then a cell phone but that "pass the mic" idea is a cool one.

Apparently you've never used google voice lol... They use nuance for their voice servers which isn't exactly cutting edge just industry standard. Their homebrew stuff (used for google voice) is horrific in its translation quality, probably quite a few years before its good enough to even be in the same league as its competitors.

Anyway, my point wasn't why is this significant to cell phones (which it certainly is) it was why is this significant to cars.

What about a phone with directional array mics and built-in noise-cancellation technology? Like the Droid X. I have no problem using voice commands with my x whether i'm in the car or not. Matter of fact its the only way i use my phone in the car. Its by far easier to compose a text message by voice than picking up the phone, and its easier to say "Call wife" than go through front end menus to press a contact button.

Originally Posted by Sonicxtacy02

What about a phone with directional array mics and built-in noise-cancellation technology? Like the Droid X.

Cell phones including the droidX have omni-directional mics not directional mics. They may have 2 (iphone4) or 3 (droidx) but they still have to cancel 180 degrees of background noise which even the best algorithms cant do in quite a few cases.