Hearing Seminars

CCRMA hosts a weekly Hearing seminar. All areas related to perception are discussed, but the group emphasizes topics that will help us understand how the auditory system works. Speakers are drawn from the group and visitors to the Stanford area. Most attendees are graduate students, faculty, or local researchers interested in psychology, music, engineering, neurophysiology, and linguistics. Meetings are usually from 11AM to 12:30 (or so, depending on questions) on Friday mornings in the CCRMA Seminar Room.

The current schedule is announced via a mailing list. To be added to the mailing list, send email to hearing-seminar-request@ccrma.stanford.edu.  If you have any questions, please contact Malcolm Slaney at hearing-seminar-admin@ccrma.stanford.edu.

Recent Hearing Seminars

  • Alex Brandmeyer on High-frequency Neural oscillations for Auditory Perception

    Date: 
    Fri, 05/22/2015 - 11:30am - 1:00pm
    Location: 
    CCRMA Seminar Room (Top Floor of the Knoll)
    Event Type: 
    Hearing Seminar
    Our transparent everyday experience of sound belies the complex perceptual processes which transform acoustic energy in the environment into rich representations of sound-emitting objects. Since the publication of Bregman’s seminal work on ‘Auditory Scene Analysis’, much progress has been made in understanding the neural mechanisms that support the formation of auditory streams and object representations. This talk will focus on the functional role of neural oscillations in audition, and the interplay between top-down (i.e. attentional and executive processes) and bottom-up (sensory) processes during the formation of perceptual representations.
    FREE
    Open to the Public
  • Compositional models for audio processing

    Date: 
    Fri, 05/15/2015 - 11:30am - 1:00pm
    Location: 
    CCRMA Seminar Room (Top Floor of the Knoll)
    Event Type: 
    Hearing Seminar
    Compositional models of audio, including those based on non-negative matrix factorisation (NMF), explicitly consider the fact that sound components combine largely constructively in the composition of more complex sounds. The use of compositional models has yielded state-of-the-art results in many audio processing tasks, such as sound source separation and music content analysis. In this presentation, I will give an overview of compositional model approaches to noise-robust automatic speech recognition (ASR) and discuss methods to build a dictionary of sound components. I will then show how compositional models can also be used to learn recurrent acoustic patterns representing words or phrases, with as application the construction of a language-independent vocal interface.
    FREE
    Open to the Public
  • Making Speech Content Sound Better

    Date: 
    Fri, 05/08/2015 - 11:30am - 1:00pm
    Location: 
    CCRMA Seminar Room
    Event Type: 
    Hearing Seminar
    CCRMA alumnus and Adobe Researcher, Gautham Mysore will lead Friday’s Hearing Seminar with a discussion about how we can improve the quality of audio. Professional quality audio is not trivial to create. Can we build tools that make it easier?

    Last week at the CCRMA Hearing Seminar we talked about how to measure the quality of a speech signal. This week we talk about how to make it sound better.

    Who: Gautham Mysore
    What: Making Speech Content Better
    When: Friday May 8 at 11:30AM
    Where: CCRMA Seminar Room
    Why: Because better audio quality is always better

    Title:
    FREE
    Open to the Public
  • Naomi Harte - ViSQOL, An objective measure for speech quality

    Date: 
    Fri, 05/01/2015 - 11:30am - 1:00pm
    Location: 
    CCRMA Seminar Room (Top Floor of the Knoll)
    Event Type: 
    Hearing Seminar
    How do humans judge perceptual quality of a signal? It’s easy to measure mean-squared error, but that is not how the brain measures quality—that would be too easy. Instead, we want to look inside the brain to see how sound is encoded. This problem is important as we build more devices that process sound and we need to assess whether they are doing a good job. Naomi’s work addresses the problem when we have access to the original and the modified sound. It’s an improvement over the PESQ and POLQA work, if you are familiar with the area.

    FREE
    Open to the Public
  • Gabriella Musacchia on Brain Plasticity with Music and Auditory Enrichment

    Date: 
    Fri, 04/24/2015 - 3:00pm - 4:00pm
    Location: 
    TBD: CCRMA Seminar Room or the classroom
    Event Type: 
    Hearing Seminar
    Brain plasticity with music and auditory enrichment: Implications for early music education. Note special time!!!!! 3PM Seminar!!!
    FREE
    Open to the Public
  • Sunil Puria on Developing and Testing a Photonic Contact Hearing Aid

    Date: 
    Fri, 04/17/2015 - 11:30am - 1:00pm
    Location: 
    CCRMA Seminar Room (Top Floor of the Knoll)
    Event Type: 
    Hearing Seminar
    The EarLens Contact Hearing Aid (CHA) is a non-surgical investigational hearing device consisting of two components: a tympanic membrane transducer (TMT), which is a light-activated balanced-armature transducer that drives the middle ear through direct contact with the umbo; and a behind-the-ear unit (BTE) that encodes amplified sound into pulses of light that are emitted via a light source in the ear canal to wirelessly drive and power the TMT. In comparison to conventional acoustic hearing aids whose output is greatly reduced above 4-5 kHz, this approach is designed to provide output levels of up to about 120 dB SPL and functional gain of up to 40 dB for frequencies up to 10 kHz.
    FREE
    Open to the Public
  • John Woodruff on Machine Listening: What do they hear and why?

    Date: 
    Fri, 04/10/2015 - 11:30am - 12:45pm
    Location: 
    CCRMA Seminar Room (Top Floor of the Knoll)
    Event Type: 
    Hearing Seminar
    Whenever our devices capture sound, audio systems are there to recognize what’s been said or turn down all that noise. But how do our phones, tablets, remotes, headphones, hearing aids and thermostats know what to listen to? Most systems use one or both of two assumptions – 1) I’m listening for speech, 2) the sound I want came from that direction. Robust speech recognition systems are perhaps the most ubiquitous realizations of the first assumption. Large-scale training on noisy speech embeds the capability to “listen for speech”, but such systems are fundamentally limited when there are competing talkers.
    FREE
    Open to the Public
  • Michael Mandel on Auditory bubbles: Estimating time frequency importance functions

    Date: 
    Fri, 03/20/2015 - 11:00am - 12:30pm
    Location: 
    CCRMA Seminar Room
    Event Type: 
    Hearing Seminar
    Listeners can reliably identify speech in noisy conditions, but it is not well understood which specific features of the speech they use to do this.  This talk presents a data-driven framework for identifying these features.  By analyzing listening-test results involving the same speech utterance mixed with many different "bubble" noise instances, the framework is able to compute the importance of each time-frequency point in the utterance to its intelligibility, which we call the time-frequency importance function.  These results can be seen as a quantification of a listener's strategy for understanding a 
    FREE
    Open to the Public
  • Sound Classification by Prof. Dan Ellis (Columbia)

    Date: 
    Fri, 03/13/2015 - 11:00am - 12:30pm
    Location: 
    CCRMA Seminar Room
    Event Type: 
    Hearing Seminar

    I’m happy to welcome Prof. Dan Ellis (from Columbia, and on sabbatical at Google) to Stanford CCRMA to talk about recognizing environmental sounds. Recognizing speech and music are relatively common applications of machine learning. But what about the rest of the world? Speech and music are only a small fraction of the sounds that we hear throughout our day.

    Dan Ellis has been at the intersection of speech recognition, audio analysis and music processing research throughout his career. He brings an unusual range of interests and skills to all three problems, and I highly recommend his work.

    FREE
    Open to the Public
  • Moving Auditory EEGs out of the Lab

    Date: 
    Fri, 02/13/2015 - 11:00am - 12:30pm
    Location: 
    CCRMA Seminar Room
    Event Type: 
    Hearing Seminar

    It's now time to take EEG out of the lab and into the real world. Our speaker on February 13 will talk about his efforts to build a mobile EEG recording device and validate its performance with an auditory task.  Maarten De Vos received his PhD from KU Leuven (Belgium) and is visiting Stanford for a few months, before heading to a faculty position at Oxford.

    His validation experiments are interesting because he describes them as a means of decoding attention.  He's using P300 attention-modulated correlates to measure what people are attending to.  Very interesting.

      Who: Maarten De Vos (University of Oldenberg, Oxford and visiting Stanford)

    FREE
    Open to the Public