Subtle Computing’s acoustic isolation models help computers understand you in noisy environments

California based startup Microcomputing Google is tackling the problem of picking up people’s voices in noisy environments with its own sound isolation models — a technology that could benefit voice-based AI products and services.

Consumer applications using voice AI today are experiencing exponential growth. AI bloggers meeting like Granola, Fireflies, to understandand Read Artificial Intelligence It has received the attention of both users and investors. Existing companies such as OpenAI, Clickand an idea It has integrated audio transcription solutions. App makers like Wispr Flow and Willow are working on voice dictation. Then there are hardware companies like Anthem and Sandbar Which uses hardware as a way to transcribe your voice, then uses artificial intelligence to generate vision and interaction.

One challenge these companies face is capturing users’ voices in any environment, such as cafes or noisy offices.

To address this problem, Subtle Computing has developed a comprehensive sound isolation model that can understand what you say even in noisy environments. Chen said there are a lot of companies working on understanding voice. Sometimes, hardware manufacturers send audio to the cloud to get clean output, but that’s not efficient, he noted.

The startup trains specific models to fit a specific device’s acoustics and adapt to the user’s voice rather than training a single model that works across devices.

“What we found is that when we maintain the acoustic properties of the device, we get much better performance than generic solutions. This also means that we can offer customized solutions to the user,” Chen said.

The company was founded by Tyler Chen, David Harrison, Savannah Cofer, and Jackie Yang, who met at Stanford University. Chen, Harrison and Yang were pursuing their doctorates while Cofer was studying for his MBA. They came together in Steve Blank’s Lean Launchpad course, where they worked on alternative interfaces to computing and began building microcomputing.

“As we interact more with artificial intelligence, we are moving toward a future in which we talk to our devices,” Chen said. “But the obvious question is how well our devices understand us, the users, in all the environments we work in every day. Whether it’s in a very noisy café or a shared office where there are other people around you, and you might be talking about something private — audio doesn’t work that way today,” he added.

The startup said it can only run the model to isolate audio on some devices, which are only a few megabytes in size and have a latency of up to 100 milliseconds. The company can also run a different model of transcribing audio and giving text output to other devices. Chen said that thanks to its isolation model, the company’s copy model can better understand users, thus creating more accurate text.

Subtle Computing said Qualcomm has selected the startup as a member of its audio and music extension program. This means that the startup’s technology will be compatible with Qualcomm chipsets and will be available on devices produced by OEMs.

The company has raised $6 million in seed funding led by Entrada Ventures, with participation from Amplify Partners, Abstract Ventures, and angel investors, including founders such as Twitter’s Biz Stone, Pinterest’s Evan Sharp, and Perplexity’s Johnny Ho.

Voice AI is a noisy space, and although interactions are increasing through this medium, the overall voice experience isn’t great, noted Karen Rutter Davis, managing partner at Entrada Ventures and former early project manager at X (Alphabet). She believes the startup’s focus on sound insulation brings a different perspective to the market.

“While you can debate whether AI will increase or decrease the time you spend on a daily basis, we can all agree that advances in computing power and machine learning/AI provide opportunities for breakthroughs in the voice interface – if implemented correctly,” Davis said. She added: “Microcomputing technology is about meeting people where they are with audio interfaces that survive extreme noise and extreme quiet, providing a reliable, easy and enjoyable audio experience. It is a game changer.”

The company said it has also partnered with a consumer appliance brand and an automotive brand — without naming them — to deploy its solutions. But Subtle Computing doesn’t want to be just a model supplier for other companies.

The startup also said it plans to announce a consumer product including hardware and software next year, without providing details.