AudioTelligence Raises $8.5M for Speech Recognition Software
Audio technology startup AudioTelligence has raised $8.5 million in a Series A funding round led by Octopus Ventures. The British startup plans to use the new capital to extend the reach of its speech recognition software, which is designed to help voice assistants identify and understand what a human is saying when surrounded by background noise.
Loud Speech Detection
AudioTelligence licenses software that acts as an “autofocus for sound,” according to the company. When there is a mix of sounds, the software can pick out a human speaking to a voice assistant and filter out the irrelevant audio. According to the company, a voice assistant’s ability to understand what someone is saying in a noisy environment rises enormously when its technology is applied. Voice assistant comprehension rises from 22% to 94% accuracy without any training of the algorithm.
AudioTelligence spun out of Cambridge University’s CEDAR Audio in 2017 in order to further develop its technology. The new round of funding comes a year and a half after the company closed a $3.1 million seed round of funding from Cambridge Innovation Capital. CEDAR Audio and Cambridge Innovation Capital both participated in the latest round as well.
“Voice command systems work reasonably well when the audio scene is quiet, but performance deteriorates rapidly once you have multiple people talking or when there’s background music,” AudioTelligence CEO Ken Roberts said in a statement. “The number of applications where our technology is needed is enormous and still growing every day. We’ve already seen some great results from real-world testing and this investment will fund further product development to ensure we can all communicate clearly with the next generation of smart consumer devices and each other.”
Noise Filter Funding
AudioTelligence’s software-based approach to noise filtering is a potentially huge enhancement to standard voice assistant technology, which often relies more on microphone calibration. Its licensing system also means developers don’t need to get new hardware every time there’s an improvement. That’s especially useful as voice technology proliferates beyond smart speakers and into televisions, cars, and other products.
The funding is a concrete marker of the rising interest in this technology too. The announcement comes just a week after DSP Concepts, an audio software startup in California, raised $14.5 million in a funding round. DSP’s Audio Weaver platform offers a growing number of software modules for audio processing, with noise filtering a notable part of its appeal. Like Audio Weaver, AudioTelligence could be applied to voice assistants to help reduce accidental wakeups from a speech on television or the radio in the fast-growing automobile voice assistant space.