The Redeye Complex

Wednesday, August 13, 2008
Music Please

My home server/jukebox project has taken an interesting twist recently.

My goal is to have something hooked up to the stereo which can play anything from my digital music collection. What that could mean is a computer hooked up to the stereo with a L/R RCA cable. Or maybe use some kind of special device made especially for music.

Now playing music is easy. It’s the selecting music task and navigating through the music library that is the difficult thing to do. My goal is to navigate and select music without have to use a computer monitor, keyboard, or mouse. The solution should not more difficult than operating a CD player. Just three or four buttons to hear music.

There are some neat devices like the Sony VAIO WA Wireless Digital Music Streamer (which the Sony Store as of last week is selling for $149!!) that does a good job of easily playing back music and podcasts. Just navigate to a network shared folder, select item, press play. Quite easy and nice to use. The tricky part is navigating through a large library of music. That small screen means plenty of button pressing.

HP has a new interesting form factor, a touch screen PC that doesn’t work too bad. Price is alright at $1399 considering you get a full function PC with a large touch screen. HP even has developed a new interface into media applications. I tried the music app and was impressed. (You can try one out at BB and FuSho). Navigating on the big screen seemed efficient. Just drag items to a playlist and press play. Not bad at all.

A while ago, I had this idea that since I already had a feedback device, my stereo speakers, could I use some kind of navigation interface that would talk back to me as I browse a directory. Turns out there such an interface on Rockbox.

Rockbox is an open source project that replaces the interface on music playing devices. One of its features is a audio navigation system sort of like what I was thinking about. The number of devices that Rockbox runs on is limited and I need at least 300 GB of storage to access. I thought that there must be a computer based solution that does something similar.

Enter the Windows Speech Recognition Macro from Microsoft.

I ran across a MS blog entry on speech recognition with the opening line…
Have you ever wanted to control your media player with your voice? Well now you can! With today’s Macro of the Day, you can say things like “Play Hotel California”, or “Play The Eagles”, or, “Play Genre Rock”, and even “Play something by The Eagles”.
Well, yes I would.

Turns out that MS has speech recognition built into almost all flavours of Vista. You just have to turn it on.

The new component is the Speech Recognition Macro which is available as a free download.

From the Overview …
The Windows Speech Recognition Macros tool – or WSR Macros for short – extends the usefulness of the speech recognition capabilities in Windows Vista. Users can create powerful macros that are triggered by spoken commands which can perform a series of tasks from as simple as inserting your mailing address to as complex as providing a completely different speech interaction with applications.
Very interesting.

I tried the special media player macro on the blog but it didn’t load properly. The macro would generate an error likely because it could not process a large music library. Could be some wonky characters or bad tags somewhere in a file. So I started with an empty library and added just a few albums. Throw in a cheap microphone and voila – the macro worked. Music would play based on what I tell it to play – literally.

Now this is very interesting.

Instead of debugging the macro, I started playing with the speech recognition macro kit, creating my own macros to play a playlist, usually an album, when I say a phrase. For an album, the macro phrase trigger would be “Artist Name Album Name”. “Alpha Stargazing Special Edition” for example. This works since you can instruct Windows Media Player to play a playlist from the command line.

Created 10 macros for 10 albums and every one of them started playing when I spoke their phrase trigger. I expanded the test to one hundred, two hundred, and now up to 300 albums. Different artists and different albums from the same artist. It didn’t matter as long as the phrases were unique. I don’t know what the limit of speech recognition phrases there are in Windows. So far so good. I even created macros to pause/resume playing (Music Please), stop playing (Music Stop), previous track (Music Previous), next track (Music Next). They all work.

There are some gotchas of course. If the microphone is next to the speakers that are playing music, the computer can’t hear your commands. So its good to have a mute switch for your speakers handy. Its even good to have a mute switch for your microphone so the computer doesn’t grind away trying to hear a phrase in the music that is playing.

The amazing thing about this speech recognition project is that it has cost me nothing so far. No new software or hardware. Just some time.

Well almost nothing.

I decided to get a better microphone. Something wireless perhaps. After some research into microphones I picked up a Microsoft LifeChat ZX-6000 for about $80 which is based on the Xbox wireless microphone. Installing the software wasn’t easy. And getting it to work on the computer was slightly challenging. But the results are worth it. From about 20 feet away from the wireless receiver (and about 30 feet away from my computer) I can tell my computer what music to play. I can interrupt playback, select something else, advance tracks all without having to navigate any menus. It’s not perfect. Sometimes the windows get gummed up and I do have to access a keyboard. I haven’t figured out how to add a playlist to an existing NOW PLAYING playlist. And I have to remember what those darn album names are. Its bad enough trying to remember an artist name like “Do Make Say Think” with one of their albums “You You’re a History in Rust”. Also have to ever tried pronouncing “Djivan Gasparayan”?

So time on home server/jukebox project is spent fine tuning the speech recognition aspect and fine tuning music playback through a computer. Like I've posted before, music playback on a computer is not as satisfying as you might think. That however will be a post for another day.

Until then …

“Music Please”


Newer Posts Older Posts