After Apple discontinued the Music Memos app that musicians loved to develop song ideas, a new startup called Tape It stepped in, leveraging AI to automatically detect instruments and annotate recordings. I filled that hole with an app that adds . The startup is now taking the next step to improve the audio recording process. This is the introduction of studio-quality automatic noise reduction algorithms, also powered by AI, for all audio, not just voice.
of AI denoiser shipped this week It is available as a free web app, and we plan to license this technology to vendors in the future. The company says this will later be integrated into its flagship Tape It app as well.
Founded in 2020 by musicians and friends, thomas walther and Jean Nash, tape it Initially, we focused on iOS recording apps for musicians. Before creating Tape It, Walther spent three and a half years there after his Spotify acquired his audio detection startup, his Sonalytic. Nash, on the other hand, is a classically trained opera singer, bassist and engineer. The duo was originally inspired to build his Tape It out of their own desire as bandmates to make it as easy to use as Apple’s Music Memos, but even more powerful with AI. It was from.
The original version of the app automatically detected instruments and visually annotated recordings, making them easier to find by looking for colorful icons. Musicians can also add their own markers, notes, and photos to files for later review.
Since then, the app has gained traction and has about 10,000 monthly active users, the company said.
But as Walther told TechCrunch at Tape It’s 2021 debut, the team aimed to expand its use of AI over time.
Image credits: tape
That led to the startup’s latest development: an AI-powered noise remover it’s been building over the past two years. The company explains that the challenge with recording is background noise. To reduce environmental noise and electrical interference, musicians record in studios and utilize complex software. Tape It wants to use AI to provide a more affordable alternative. Their software automatically removes noise such as hum and hiss, as well as spoken words, with the goal of producing studio-quality results in songs, single instrument tracks, and field recordings.
“What we’ve developed is an automated version of the noise reduction software that has been used in professional recording studios for the past 15 years,” explains Walther.
To verify the results, Tape It Presenting academic research with scientific listening tests This shows the quality of the software, which competes with industry-leading noise removal software.
The company explains in the video that while significant advances have been made in audio enhancement systems, they typically only work on audio, distorting or corrupting the music signal. On the other hand, professional noise reduction systems require manual control of complex software by professional users. Tape It’s technology involves connecting a neural network controller to a signal processing-based noise reduction algorithm. This allows automatic denoising of common audio signals, including music. The company plans to present its findings at next week’s AES conference.
“The reason people don’t automate these things is because [professional systems] That’s because traditionally you can’t put them into a neural network…You can’t train a system like that,” Walther said. “In fact, we are the first company to train a system like this, which is why we are so excited about this larger area.”
He said the academic community is probably not interested in the denoising product itself, but in how the denoising product was able to behave this way, as it impacts other applications that automate studio software. He added that he would be interested.
Still, the noise removal software already has interested potential customers, including major studio software vendors and major hardware manufacturers. In such cases, enterprise pricing is available, while smaller startups are offered cheaper plans.
When announcing the news, Walther said, “Everyone is excited about the creativity of AI.” “We’re excited that AI can solve boring problems. We take care of the background noise, so you can focus fully on the creative part and write more songs.” ” he said.
AI technology is not only being used to reduce background noise for musicians, but of course other companies are also leveraging AI to create near-studio-quality sound for podcasters. For example, Podcastle this month released Magic Dust AI, a generative AI tool that removes background noise and enhances dynamic range.
Tape It’s five-person team is based in Berlin, London, Los Angeles, and Stockholm and includes designers and musicians Christian Crusiuspreviously affiliated with design consultancy firm Fjord. obtained Manufactured by Accenture. Most of the work on noise removal software is Christian Steinmetza doctoral researcher in AI and audio.
The company has so far declined offers of funding, but continues to raise funds on its own.
“This is basic research, and we didn’t know how long it would take,” Walther said, explaining why he took this approach. “We thought it was a little risky to get investors who weren’t that patient. [they’d push you to ] Adopt an open source model and move on. But we wanted a greater technological advantage,” he said.
The company is currently considering raising funds given the pace of the AI market and is in discussions, but has not made any formal decisions at this time.