2023/09

With the thriving era of digital communication, the public's demand for video and audio quality is steadily on the rise. According to a 2020 survey of the podcast industry, conducted with over 7,000 respondents, recording quality was identified as one of the most influential factors affecting audience listening experiences. This growing demand aligns perfectly with the proactive development of noise reduction by audio AI startup "DeepWave."

At the annual benchmark event in photography, the Taipei Photography and Video Device Exhibition 2023, DeepWave showcased "Noise Eraser" as the "AI savior of audio in visual productions." This software combines innovative features such as noise reduction and vocal separation through AI, effectively resolving the persistent problem of environmental noise during audio recording.

Post-production handled by AI assistant for top-notch content at minimal cost 

Noise Eraser employs cutting-edge artificial intelligence technology that accurately distinguishes between vocals and background sounds in various audio files, achieving noise reduction to enhance audio recording quality. Furthermore, since AI learning reduces the need for professional manpower and associated costs, it significantly lowers the threshold for noise reduction technology, opening up new possibilities for digital content and new media creators.

As visual quality continues to improve, the enhancement of audio quality becomes increasingly crucial. Integrating AI audio recognition technology with photography equipment is becoming an irreversible trend. Whether using GoPro for extreme and adventure shooting, the ZV-E10 for Vlogging, or the Canon EOS R10 for fast-focus shooting of vehicles, Noise Eraser effortlessly complements noise reduction, ensuring seamless integration of visuals and sound. This enables the easy creation of high-quality audio, offering creators in various fields a wider range of possibilities.

Noise Eraser accommodates material needs from a variety of devices, with the creator’s work demonstrated by YouTube travel adventure channel [Rosie Gabrielle].

Innovative applications of audio processing technology 

Beyond the realm of self-media, DeepWave's AI technology finds widespread application in multiple domains. In the field of education, it assists in language training and separates clearer audio for teachers, reducing the time pressure for preparing teaching materials and allowing for the evaluation of students' pronunciation and skills.

DeepWave's Chief Operating Officer, Sung Yeh, noted that optimizing speech has traditionally required a considerable investment in professional manpower, which is often prohibitively expensive in terms of both time and cost. As a result, individual YouTubers, Podcasters, social media influencers, or online educators may not be able to afford such expenses. They may rely on noise-resistant recording equipment for initial capture but encounter difficulties hiring professionals for post-production. Therefore, Noise Eraser addresses the noise reduction needs of post-production.

Looking ahead: The future of audio recognition technology 

DeepWave specializes in the field of AI audio recognition and believes that in the fiercely competitive AI market, acoustic recognition possesses significant untapped potential. In comparison to the saturated field of image recognition, acoustic recognition offers greater prospects. Chief Technology Officer Ken Yeh pointed out that while large language models like ChatGPT have garnered widespread attention, the development of large-scale models for audio processing has not reached the same level of maturity as image or text recognition, and there are significant gaps in related databases.

Therefore, one of DeepWave's future goals is to leverage its unique technical capabilities in acoustic recognition and its accumulated audio database to contribute to the development of large-scale language models. DeepWave firmly believes that through ongoing innovation and breakthroughs in sound processing technology, it can enrich the sound processing capabilities of large models and bring broader development prospects to the field of AI.