Sound recognition AI software optimized for Qualcomm Snapdragon platform...
British sound recognition AI experts Audio Analytic have pre-validated and optimized their ai3-nano sound recognition AI software for Qualcomm’s latest Snapdragon platform, giving mobile phones the ability to recognize audio scenes and adjust their audio output accordingly.
Audio Analytic’s sound recognition AI can tell the difference between chaotic, lively, calm and boring acoustic scenes – offering important contextual information which OEMs can use to adapt how phones behave. For example, if the user is in a busy train station, the phone would recognize that as a chaotic scene and boost the volume of notifications so that the user doesn’t miss an important call. When the user returns to a calm environment like an office, the volume of notifications can be reduced or turned off, to avoid annoying their colleagues.
“Sound recognition is the most exciting branch of artificial intelligence right now,” said Chris Mitchell, CEO of Audio Analytic. “As humans, we make sense of the world around us through sound, and by empowering machines with a human-like sense of hearing we’re enabling the next wave of innovation on smartphones.”
Audio Analytic’s ai3-nano sound recognition AI software is pre-validated and optimized to run in always-on, low-power mode on the second generation Qualcomm Sensing Hub, part of the brand-new Qualcomm Snapdragon 888 5G mobile platform. The software takes up less than 40kB on the Snapdragon chipset, so it can run alongside the low-power audio subsystem in the Qualcomm AI Engine, and alongside the Qualcomm Aqstic audio codec, which supports voice assistants.
Audio Analytic also announced several new sound recognition AI apps for smartphones. This includes an app to automatically tag videos with specific sound labels, such as laughter. Another app triggers augmented reality (AR) video filters based on sounds you make for a variety of fun and silly effects. A third app is focused on accessibility, with the AI able to alert a hearing-impaired user to certain sounds in the environment, such as an alarm or a knock at the door.