Thursday, March 29, 2012

Voice-Activated Devices

My disability has made it difficult, or nearly impossible, to push the simplest or softest of buttons. 10 years ago I could feed myself, pick away on a standard keyboard, play my PlayStation 2, push my cell phone keypad buttons, turn on the radio, and use a well designed remote control. Today, I can't do any of those things because Spinal Muscular Atrophy is a cruel bitch.

But, technology to the rescue, at least for some of those things. The fewer things I can do with my fingers, the more things I seem to be able to do with my voice. When I first started seeing these gadgets on the market, I made it a point to try as many of them as I could afford. Please note: I'm going to provide links to all the devices I mention, but these are not paid or sponsored in anyway… I just like to be helpful.

Today, I'm going to focus on cell phones because I found it the most useful of voice-controllable devices. Outside of a computer, a cell phone will give the most bang for the buck when it comes to freedom and independence… as long as it can be used independently without help from anyone. In future postings, I will be sharing about remote controls, radios, alarm clocks, lamps, and tons of other devices. Thankfully, the list is expanding each and every year.

In my search for voice-controllable cell phones, I quickly discovered a major problem. Sure, while some interesting gadgets were starting to appear in the consumer marketplace, almost all of them required some type of physical activity on the users part to put the phone into "listen" mode. To this day I don't understand why someone would want to push a button on their cell phone just to then ask their cell phone to call Steve. Why not just push the button on the phone that calls Steve directly? This type of "voice control" doesn't interest me and can't help me because if I can push one button, I can probably push all of them.

But sometime in the year 2000 I found it. A fancy new Qualcomm QCP-2760 cell phone through Sprint. The phone itself was unremarkable by itself and the buttons were really difficult to push, but Kyocera sold a car speakerphone kit specifically for that model that interacted with it in some magical way that promised total dialing control just by talking to it. My dad & brother helped me wire it to my 12-volt wheelchair batteries, and after clipping the "visor" microphone to my shirt lapel I turned it on. After programming keywords, contacts, numbers, and other stuff it asked me to say, I tried it out. It truly was "always listening" for a trigger word and it was quite responsive with a quality sound on both ends. It could dial all of my contacts, and it could dial any number I needed just by me saying the numbers out loud. Incoming calls were not announced by name or number, unfortunately, and I discovered quickly that the voice "answer/don't answer" options didn't really work. I was forced to put my phone in auto-answer mode and it just answered every call no matter where I was. Occasionally, it would lock up and quit responding but a simple on/off always took care of it. Looking back, a lot of thought was put into its design and though it has its quirks and problems, it has not been topped yet feature-wise in my opinion. Firstly, the trigger word was required to be heard twice within 2 seconds of each other… what an innovative idea. Today I'm using a Moshi Bluetooth Speakerphone (more on that later) and it will activate several times per hour just by the TV or by me sighing (I still can't figure out how a sigh sounds like "Hello Moshi", but I digress). For activation by voice, the Kyocera speakerphone would beep when it heard the "magic" word/phrase that you chose when you first programmed it… then, if it heard it again within a very short time it would start the voice dialing prompts. There were a few false calls here or there I made accidentally, but otherwise I felt this was a very good system. The obvious drawback (besides being speakerphone-only) is how bulky the entire kit was, and the difficulty for most people to wire it to their wheelchairs. And, while using it, you are stuck with the Sprint Network and that 1 particular phone. It served me well, however, for 4-5 years but I knew I had to keep looking for its inevitable replacement.

The day finally arrived, I believe sometime around 2005… I found a cell phone and speakerphone option that was much simpler in its design and implementation. The Sony-Ericsson T637 through Cingular (now AT&T) and the DriveBlue Speakerphone made by the French company Parrot. Without any wires to the phone, the DriveBlue could dial and receive calls entirely by voice. A wired microphone was still used, but the speakerphone itself was a 1-piece 12-volt plug-in gadget that was amazing in its simplicity compared to the Kyocera kit. And, since it didn't need wired to the cell phone it seemed too good to be true. Bluetooth was in its infancy, and the companies that included it as an option on their cell phones all seemed to have a different idea on what protocols and handshaking modes to implement. Because of that, the DriveBlue had to be paired with just the right phone to take advantage of everything it could do. The Sony-Ericsson T6xx series along with 1-2 others were the only cell phones that the DriveBlue could interact fully with. Having one of these phones paired with it allowed it to activate voice dialing by using the "magic word" instead of pressing a button.

Now, the Sony-Ericsson T637 was special in itself. To this day, it is only one of a handful of phones that had a built-in "magic word" voice dialing option. In a pinch, with a headset, you can dial the T637 without any other kit or speakerphone. Unfortunately, the programming is poorly done and you only get a weird series of beeps, sounds, and tones to indicate what your voice is being interpreted as. There is no confirmation on who you are dialing or where you are in the prompting modes. Answering is only possible using auto-answer or pressing a button. But, voice dialing with a "magic word" is there inside my phone, so I still felt good about it :-)

However, I was much more satisfied using it in combo with the DriveBlue. Bluetooth pairing was simple enough, and it quickly copied over my phonebook. A DriveBlue application appeared in my Bluetooth menu on my phone, and going in there allows complete configuration and programming of the keywords and names of all your contacts. Unlike the Kyocera kit, this one announces incoming calls by name (if they are in your programmed contacts), but it still won't announce the number for unknown callers. The "magic word" was needed to be heard only once for the voice prompts to start, so I had to make it a very weird complicated 3-tone phrase to stop it from activating all of the time. It still did, but that was the best I could come up with. The voice dialing prompts were basic, but effective, and allowed different numbers for each contact to be dialed (home, office, and cellular), but unlike the Kyocera kit it cannot dial a non-programmed number. It would rarely lock up, and accuracy was pretty spot-on. One great feature not repeated with any other device was the ability to hang-up using only your voice while on a call. Every other device required me to wait until the other side disconnected (and that can be several minutes if the call went to voicemail). During the call, the DriveBlue listened for the "hangup" magic word and would respond immediately. Unfortunately, I did lose more than a few calls due to inaccurate interpretation, but I love that feature and miss it every day. Sound quality was, however, poor compared to the Kyocera as it sounded very tinny and crackly with the T637, and the people I called said it didn't sound as clear either. The build quality was suspect as well and the microphone wouldn't work unless it was pulled out of the socket ever-so-slightly. Again, to make it work on a wheelchair you still need to wire a 12-volt cigarette lighter socket to the batteries, but this is a much simpler modification for a friend to do. Parrot quickly released version 2 of the DriveBlue and called it the DriveBlue + and they promised better sound, better compatibility, and a better microphone. I upgraded when I could, and found the build quality was better with the microphone connection but not with the microphone itself. Sound was only slightly improved for me, but this one seemed to lock up several times per week.

I bought several original kits, and several + kits and they lasted me until almost all of the microphones broke. Replacements are difficult to find, and they quit making the DriveBlue sometime in 2008. You still see the occasional original kit available on eBay, but they are getting rather rare. Parrot later produced one additional car kit that included the "magic word" option, but it was a complete fail. With the EasyDrive, Parrot not only did away with the external microphone, but they made the voice prompting much more complicated and not as responsive. I didn't feel like mounting something so big so close to my mouth anyway, so I quickly sold it on eBay before it lost its resale value.

Well, now I'm using the Moshi Bluetooth Speakerphone. Moshi is a company that sells various items that are entirely voice-controlled, and for a little while in 2009-2010 they made a cell phone speakerphone meant for a car visor. It is completely self-contained, charges with USB, and is always listening. It works out of the box with most Bluetooth phones, and does a fair job doing what it was designed to do. Unfortunately, it seems a little too simple and/or dumbed down for the general population and doesn't allow any customizing or programming your own voice for better recognition. The microphone isn't sensitive enough for my weak voice, and everyone complains they can't hear me very well. On my end, the sound is a massive improvement over the DriveBlue, and quality wise Moshi is pretty solid and has survived several large falls to the floor. To activate voice dialing, you must say "Hello Moshi" then one of several pre-programmed key words. You can't directly program your contacts and dial them by voice, but a simplified version of "call favorite 3" or "redial" is used. Incoming calls at least announce a text to speech version of the contacts name, or the number (finally), and the incoming options of "answer/ignore" work flawlessly. It does lock up sometimes, but no worse than anything else. One thing Moshi can do is to interact directly with the voice-control/command options on your cell phone. By first saying "Hello Moshi", then saying "Phone Command", you are then able to interact directly with your particular cell phone's built-in voice options (play a track, call Steve at work, etc.). Moshi has been a good intermediate step, but the major drawback for me, besides the microphone quality, is the physical size/shape and lack of an external microphone jack. Since I speak quietly, to have any chance of anyone hearing me I wear it on my shirt collar. It looks a little goofy, I've been told.

The future awaits, but my next attempt will be the BlueAnt S4. It is the only "magic word" enabled cell phone option (that I am aware of) that hasn't been discontinued. It is very similar in all respects to the Moshi, except for 3 very specific differences. #1: there are no favorites it can directly dial (Moshi at least had 6); you must use the "phone command" option to call anyone, #2: there is an android app that interacts directly with the BlueAnt that supposedly allows you to listen to and respond to text messages completely by voice, and #3: it can stream music from your phone. I just bought one, and once I have the chance to try it out, I'll post the results. But, I may just skip android and try it with the iPhone 4s and Siri to see what I can do just by talking :-)