Winning the Battle with Background Noise

[IMGCAP(1)]Microphones and headsets are the good news/bad news joke of the speech recognition industry. The good news is that noise canceling microphones make speech recognition software more effective. The bad news is that many people who express initial interest in using speech are turned off when they find it requires a headset. Andrea Electronics produces anti-noise technology that cancels background noise for clearer communications by users of headsets, handsets, and wireless products. Established in 1934 as the Andrea Radio Corporation, the firm was one of the first television manufacturers in the country. In 1991, the company realized the importance of creating robust audio communications equipment capable of providing reliability and comfort in high noise environments. This year has been one of transition for Andrea as the company moves toward becoming a software solutions provider. They acquired Lamar Signal Processing Ltd., an Israeli software developer, and with it, Lamar’s Digital Super Directional Array (DSDA), which the company believes will be a breakthrough technology for speech recognition. DSDA provides for the cancellation of background noise and the transmission of a clear voice signal, even when the speaker is at a distance from a microphone. DSDA uses an array of microphones, unique system design and DSP processing to control acoustic signals and achieve hands-free operation for speech-centric computing. We recently interviewed Douglas J. Andrea, co-president of Andrea Elec-tronics Corp., on Andrea’s near term goals and the future of the speech industry. Is speech recognition now at a point where it can be called a mainstream product? What we are working toward is the horizontal application of speech recognition. And I think that is still a year or two away. In the near term, we are seeing more acceptance in the vertical markets. A lot of things have to grow together to make speech widely accepted. Microsoft, IBM and Dragon are all doing their part to perfect the PC and the dictation products. The computer industry sees speech as a natural interface for the PC, and one that will improve user acceptance for the PC. Speech helps the PC become a much more enjoyable multimedia experience. What is the most important technological breakthrough in making speech work?
Headsets and microphones have an important place. But for the horizontal market, the far field DSDA technology is generating a lot of excitement. We’re in pre-production with this now and are targeting Auto PC applications first, with speech recognition for desktop PCs to follow. This will mean we’ll be pretty close to the accuracy of a closed headset and users won’t be tied to the machine. In environments where people are at a telephone all day, headsets will grow in popularity. They are a better tool for call center people and others who are on the phone constantly. What does the future hold for speech? Where do you see speech in 3 to 5 years?
I don’t think we will ever really lose the keyboard or the mouse. But we can get to a point where you have full speech capability, so if you want to do everything by speech, you can. You will also see hand held devices that can be controlled by speech and all of them will be run by software. You are also going to see many applications connected with the global position system (GPS) as well as with personal digitial assistants (PDAs). The whole human machine interface will be much more interactive. The computer industry, the box makers and the chip makers, are starting to say speech is ready. Bill Gates is talking about it. Andy Grove sees the chip speed and processing power of the computer increasing. With home automation and information appliances, turning the lights off by voice in your house and that sort of thing will become more common. I don’t see the appliances having microphones in them, but more likely, there will be stations throughout the house, where you can talk to the server and control the utilities. It will create many new applications for speech, probably within two years. In three to five years, this type of technology will be commonplace. The average user does not want the computer to do more and faster computations. He or she wants the convergence of the telephone and the computer. He wants to be able to do video conferencing, and to not have to fumble around with a thousand different buttons. At Andrea, we specialize in the input of voice. So we are the first block in the block diagram that describes speech technology. Andrea is really contributing to the “audio architecture” of the next generation of computers, and becoming a software and technology company. The technology has made the leap into the digital domain. Speech is gaining acceptance. You’re going to see a big growth in 1999 and 2000 and general acceptance among mainstream Americans by 2003.