Dealing with noisy audio in recordings can be a challenge for audio engineers. However, a German startup, AI-coustics, is revolutionizing the field by leveraging generative AI to enhance the quality of voices in videos.
Recently, AI-coustics secured €1.9 million in funding as it emerged from stealth mode. CEO Fabian Seipel explains that their technology surpasses standard noise suppression methods to improve voice clarity across various devices and speakers.
Seipel, who has a background in audio engineering, co-founded AI-coustics with Corvin Jaedicke, a machine learning lecturer at the Technical University of Berlin, in 2021. Their shared experience with poor audio quality during online courses at TU Berlin fueled their mission to address issues of speech intelligibility.
While the market for AI-powered noise reduction software is competitive, AI-coustics distinguishes itself through its innovative approach to training AI models for noise reduction.
AI-coustics uses a model trained on speech samples recorded in their Berlin studio. By simulating various audio issues during training, such as noise, reverberation, and distortion, AI-coustics prepares its model for real-world applications.
To ensure inclusivity and mitigate biases, AI-coustics focuses on recruiting a diverse range of speech sample contributors. Seipel emphasizes the importance of diversity in eliminating bias and enabling the technology to work effectively across languages, accents, and demographics.
In a test using three video clips, AI-coustics effectively improved voice clarity by reducing ambient background noise. This demonstrates the platform’s potential to enhance audio quality in real-time and recorded scenarios.
AI-coustics envisions integrating its technology into devices like soundbars and smartphones to automatically enhance voice clarity. Currently, the startup offers a web app, API for post-processing audio and video, and an SDK for seamless integration into existing workflows and hardware.
With a revenue model based on subscriptions, on-demand pricing, and licensing, AI-coustics has already secured multiple enterprise customers and 20,000 users. The company plans to expand its team and enhance its speech-enhancing model in the coming months.
Despite concerns about job displacement in the audio mastering field, Seipel sees AI-coustics as a tool that can streamline audio production processes. By automating tasks and maintaining high speech quality, the technology offers efficiency gains for content creation and broadcasting.
The funding for AI-coustics, provided by Connect Ventures, Inovia Capital, FOV Ventures, and Jan Bohl, signifies a step forward in advancing AI technologies for audio enhancement and noise reduction.