Hardware Podcasts

Why Sound Processing Takes Time, Not Just Frequency

EE Times Current · 15 Mar 2024 · 45m

What this episode covers

Dr. Shih-Chii Liu, co-director of the Sensors Group at the Institute of Neuroinformatics (ETH Zurich), explains how her neuromorphic silicon cochlea converts sound into asynchronous spikes to enable low-power voice activity detection. She details the use of delta modulation for faithful signal representation and 'delta networks' that leverage activation sparsity to reduce energy consumption by over 50%. Professor Ralph Etienne-Cummings of Johns Hopkins University discusses the historical context of silicon cochleas, citing early work by Carver Mead and Andre van Schaik, and their commercial applications in hearing aids and sound localization.

Most relevant to

early-stage companies

Topics

Mentioned

Companies
Institute of Neuroinformatics, ETH Zurich, Johns Hopkins University, Fortiss, Caltech, DARPA
Products
ASU
Also
Carver Mead, Andrei van Schaik, Gerd Schindler, Chris Tomasi, Andreas Andreu, Paul Mueller

Related episodes from other shows

Other shows on the same ground, closest first.

  • The Amp Hour · 1h 23m

    An Interview with Jørgen Jakobsen

    Jørgen Jakobsen, an IC designer at Mios Audio, details his career designing hearing aid chips in Denmark and his current work on ultra-low-power Class D amplifiers. He explains how Mios Audio's multi-level output stages eliminate idle losses, achieving ten times lower power consumption at normal listening levels compared to traditional Class D designs. The conversation covers the technical challenges of integrating RF, DSP, and analog audio on small silicon areas, as well as the use of open-source platforms like the ESP32 for prototyping these high-efficiency audio solutions.

    Analog & power electronicsMicrocontrollers & dev platforms
    8 chapters

  • Robot Talk · 28m

    Episode 139: Advanced robot hearing - Christine Evers

    Christine Evers, Director of the Centre for Robotics at the University of Southampton, discusses bio-inspired machine listening that enables robots to understand complex acoustic environments. She explains how embedding human auditory processing into deep learning architectures creates efficient, interpretable models suitable for resource-constrained robots, moving away from massive, energy-hungry internet-scale models. The conversation covers the challenges of ego noise, the need for spatial awareness via microphone arrays, and the concept of 'active audition' where robot movement shapes auditory perception.

    AI & machine learningSensors & measurement in products
    6 chapters

  • Hackaday Podcast · 1h 11m

    Ep 314: It's Pi, but Also PCBs in Living Color and Ultrasonic Everything

    Hosts Elliot Williams and Dan Maloney review a week of hacks, starting with a sound-guessing game won by 'angle grinder'. They discuss Alex Toussaint’s Supercon talk on a 380-microphone sonar array capable of identifying mosquitoes via Doppler shift, and Joe Long’s exploration of full-color PCB silkscreens from JLCPCB and PCBWay for badge art and assembly guides. The episode also covers an automated machine for embedding fiber optics in concrete, a monolithic BMS for prismatic battery packs, and ultrasonic knives for cutting food.

    Sensors & measurement in productsEDA & CAD tools
    7 chapters

  • The Amp Hour · 1h 05m

    Silego, ADCs & Seismic Detection - Dave's Dingo Dystocia

    Dave Jones and Chris Gammell discuss the release of Agilent's 'Dingo' firmware update for the 3000 series oscilloscope, which adds arbitrary waveform generation. They then analyze the technical differences between audio and precision ADCs, using Jones's experience with seismic detection arrays to highlight power and cost trade-offs. The conversation shifts to Silego's ultra-low-cost, one-time programmable micro-FPGAs, noting their proprietary ecosystem and tiny packages. Finally, they touch on the passing of Dennis Ritchie and the University of Minnesota's program for socially awkward engineering students.

    PCB design & layoutBatteries & low-power design
    8 chapters

  • Embedded.fm · 1h 14m

    507: Turn Our Data Into Predators

    Elecia White and Christopher White discuss the Embedded.fm Patreon book club's study of 'Data-Driven Science and Engineering' by Steven L. Brunton and J. Nathan Kutz, focusing on linear algebra, spectral methods, and the Datasaurus dozen. Christopher White details his Dogbotic 'Do-It-Yourself Rhythm Widgets' course, where he builds an analog drum machine using CMOS logic ICs, op-amps, and piezo triggers on breadboards. The hosts also explore the 'I Saved a PNG Image To A Bird' video, discussing AudioMoth ultrasonic recorders and BirdNET-Pi for field audio analysis.

    Embedded firmwareAI & machine learning
    6 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by EE Times Current. The episode itself belongs to its publisher.