With over twenty years’ experience in digital audio signal processing and speech recognition, Richard has designed and coded algorithms that power the audio experience across multiple audio categories from stadium concert sound reinforcement, through bass guitar effects pedals to consumer electronics devices and business teleconferencing. He has 17 granted and 15 pending patent families across a broad range of audio signal processing topics.
Richard is currently Director, Speech Analytics at Dolby Laboratories and is based in Dolby’s Sydney R&D lab, where he leads a team of researchers and engineers that melds classical signal processing techniques with machine learning.
The Technology and Challenges of Audio AI
- A tour of current speech recognition and other audio AI technology and how it is changing the world of audio.
- AI technology is more accessible now than ever before but traps await those who apply the technology unwarily.
o A couple of specific examples are presented of audio AI systems that don’t behave quite as you might expect.
- With machine learning approaches to audio processing, careful design is still required. This lesson is perhaps more evident in audio processing than in other fields. Instead of designing algorithms we must now design:
o appropriate cost functions and benchmarking schemes,
o training and development sets that will generalise well to the people and places in which your technology will be deployed, and
o domain-specific data normalisation and augmentation schemes