Multiple Microphone Inputs?

[commasense]

Thanks for these answers. We’re getting a clearer picture. So you have a robot that’s controlled by a separate computer? Have you considered putting mikes on the robot and using the tricks Cerowyn mentioned?

@Digital Stimulus

@commasense

Could he/she press a button when giving commands, or does this have to be a hands-free operation?

Well, much of my research concerns fault-tolerance (although this particular part is a narrow part of it); so, having backup options is OK. In other words, multiple forms of input (e.g., speech isn’t working, use the keyboard) are an option. Of course, it’s incredibly frustrating to have to switch modes, and it would be nice to avoid it whenever possible.

You may have misunderstood where I was going with my question. I was thinking of a fixed mike with a momentary “on” button that the user would press when he was addressing the system. The rest of the time it would be off, avoiding the problem of extraneous sounds being picked up. If the user is only occasionally addressing the system, this might be a workable solution.

Would you happen to be able to recommend something (book, website, etc.) on acoustic design?

I’m sure a little Googling would turn up plenty, but all of my books on acoustics are quite old (remember I said I’m a “former” semi-pro sound engineer) and predate the tremendous advances made since the introduction of personal computers. Also, I get the sense that your studying acoustics at this point would be like studying watchmaking because you wanted to find out what time it is.

Maybe sticking to the KISS principle should be my guide, as it’s looking like I might be biting off more than I can chew at this point.

Exactly. All of your answers confirm my view that a wireless headset mike is your best option. The advantages are tremendous and the downside is minimal. As Musicat has said, the farther the mike is from the sound source, the more problems you’ll have with speech recognition. After getting the system to work with the headset mike, you could try playing with multiple fixed mikes, mixers, gates, acoustic treatments, etc. But be prepared for reduced intelligibility and a higher error rate at greater cost and inconvenience.

Here are a bunch of pro audio wireless headset mikes, ranging in price from about $100 to $800. Although I wouldn’t recommend the cheapest ones on that page, I think most of the units in the $150-350 range would probably work great for you. (I have no connection with that vendor, am not endorsing it, and haven’t shopped around to see if those are the best possible prices.)

[Musicat]

@Digital Stimulus

While I’ll google terms later, is there a site you can supply that explains these terms (perhaps even with some recommended equipment)? And I am on a limited budget, being a non-resident student (thus sacrificing my stipend) while finishing up.

Here’s a Wiki entry and the blurb for a commercial product. It may be more than you wanted to know:

Noise gate

bayviewproaudio.com

As far as recommended equipment, I have been out of the field for long enough that I don’t know what’s currently on the market or what’s best, sorry. If you can grab an active audio engineer, I’m sure he could tell you.

I do know you can get new, 2-channel and professional gate/companders for $100.

[Digital_Stimulus]

@Cerowyn

We did a robot in Japan that used three built-in microphones connected to a cheap controller with a Windows driver. The robot did voice recognition; there’s a video somewhere on the 'Net of me demonstrating it during a cable TV show. Despite all of the background noise in the studio and my bad Japanese, the robot worked well.

When you say “controller”, to what are you referring? That is, a PCI card? Hub-type (mixer) hardware? Also, might you dig up a link to that video (and/or other related items)?

We used basic algorithms to have the robot turn towards the speaker, and if necessary approach, to improve voice recognition. It was a useful trick that meant that we could focus S-R accuracy tuning to a smaller subset of audio conditions.

One huge question: assuming the robot was autonomous, how did you localize on the sound source? It seems to be almost as difficult a problem in a dynamic environment (i.e., non-fixed, where there’s no guarantee where a person is located). Although the point about restricted audio conditions is well-taken. One issue is that I’m only going to have a fairly cheap robot available, which makes mounting a laptop on it relatively difficult (due to both power/battery consumption and additional weight).

I’ll dig out the audio controller information tonight, and see if they have Linux drivers.

Any information would be much appreciated. Thanks.