Audio Tech Developer DSP Concepts’ Noise Filter Software Qualifies for Alexa Built-In Devices
Amazon has qualified TalkTo, the noise filtering software created by audio technology startup DSP Concepts, for Alexa built-in devices. The qualification streamlines the process for third-party manufacturers integrating the voice assistant into their product to include TalkTo in their software.
“What Amazon wants is people to build products that invoke Alexa,” DSP Concepts CEO Chin Beckmann told Voicebot in an interview. “The more people that can build things, the better. The problem is that you have a lot of product makers all making different things, and not everyone’s going to want the same microphone arrays. With TalkTo, they can filter out and cut through all the noise.”
TalkTo works to cut out extraneous sounds to improve how well a voice assistant hears commands. An Alexa device with TalkTo can listen to someone speaking directly to it even if there’s a lot of other sounds or it’s outside. The sound cancellation adapts to circumstances and can make someone intelligible to a voice assistant in devices with differing numbers of microphones. The software can even adjust to different microphone placements so that designers don’t have to concern themselves as much with the issue of noise when crafting the device. The software is also customizable based on where the device may be used. For instance, it can be adjusted to avoid accidental wake-ups, a frequent complaint made by owners of devices with voice assistants.
“The number one benefit [of TalkTo] is flexibility, and number two is performance,” Beckmann said. “Oftentimes, the issue isn’t the algorithm; it’s the flexibility. [The manufacturers] are restrained by industrial design, so they can’t have a set solution. They have to have something flexible. You know how Alexa can wake up from hearing sounds from a TV? TalkTo can be customized so it doesn’t react to TV. Or we can make it harder to trigger in the middle of the night.
DSP Concepts is best known for its embedded audio software platform Audio Weaver, used in more than 40 million devices, according to the company. DSP raised $14.5 million in February and slightly more than $25 million since Beckmann and her husband Paul Beckmann founded the company in 2003. The rise of Siri, Alexa, Google Assistant, and other voice assistants have led to a lot more interest in the company’s technology and an accelerating rate of adoption, Beckmann explained. It’s also changed what Amazon chooses to qualify for Alexa devices, with the addition of software instead of only hardware.
“Voice interaction is not new; it’s been around for a while,” Beckmann said. “But, audio had always been sleepy until the success of the Alexa product, and then suddenly everyone is becoming an audio product maker.”
Amazon’s decisions on what devices can have Alexa depend partly on how well those devices perform. With the new qualification, those makers using TalkTo can feel more confident that Amazon will approve their device.
“It just cuts through the clutter,” Beckmann said, “A lot of people can claim to have front-end processing, but it hasn’t gone through the rigorous testing and Amazon processing we did.”