Sound source separation at low-latency requires that each incoming frame of audio data be processed at very low delay, and outputted as soon as possible. For practical purposes involving human ...listeners, a 20 ms algorithmic delay is the uppermost limit which is comfortable to the listener. In this paper, we propose a low-latency (algorithmic delay <; 20 ms) deep neural network (DNN) based source separation method. The proposed method takes advantage of an extended past context, outputting soft time-frequency masking filters which are then applied to incoming audio frames to give better separation performance as compared to NMF baseline. Acoustic mixtures from five pairs of speakers from CMU Arctic database 1 were used for the experiments. At least 1 dB average improvement in source to distortion ratios (SDR) was observed in our DNN-based system over a low-latency NMF baseline for different processing and analysis frame lengths. The effect of incorporating previous temporal context into DNN inputs yielded significant improvements in SDR for short processing frame lengths.
Inferring User Intents from Motion in Hearing Healthcare Johansen, Benjamin; Korzepa, Maciej Jan; Petersen, Michael Kai ...
Proceedings of the 2018 ACM International Joint Conference and 2018 International Symposium on Pervasive and Ubiquitous Computing and Wearable Computers,
10/2018
Conference Proceeding
Odprti dostop
Sensors in our phones and wearables, leave digital traces of our activities. With active user participation, these devices serve as personal sensing devices, giving insights to human behavior, ...thoughts, intents and personalities. We discuss how acoustical environment data from hearing aids, coupled with motion and location data from smartphones, may provide new insights to physical and mental health. We outline an approach to model soundscape and context data to learn preferences for personalized hearing healthcare. Using Bayesian statistical inference we investigate how physical motion and acoustical features may interact to capture behavioral patterns. Finally, we discuss how such insights may offer a foundation for designing new types of participatory healthcare solutions, as preventive measures against cognitive decline, and physical health.
Modern hearing aids (HAs) are not simple passive sound enhancers, but rather complex devices that can log (via smart-phones) multivariate real-time data from the acoustic environment of a user. In ...the EVOTION project (www.h2020evotion.eu) such hearing aids are integrated with a Big Data analytics (BDA) platform to bring about ecologically valid evidence for policy-making within the hearing healthcare sector. Here, we present the background of the BDA platform and a concrete case study of how longitudinally sampled data from HAs can 1) support hypotheses about HA usage prognosis, and 2) bring new knowledge of how HAs are used across a typical day. In five participants, we found that the hourly HA usage was negatively associated with both the mean and the variance of the signal-to-noise ratio, and that increases in the daily total HA usage were associated with higher and more diverse sound levels.
Provider: - Institution: - Data provided by Europeana Collections- All metadata published by Europeana are available free of restriction under the Creative Commons CC0 1.0 Universal Public Domain ...Dedication. However, Europeana requests that you actively acknowledge and give attribution to all metadata sources including Europeana
Provider: - Institution: - Data provided by Europeana Collections- All metadata published by Europeana are available free of restriction under the Creative Commons CC0 1.0 Universal Public Domain ...Dedication. However, Europeana requests that you actively acknowledge and give attribution to all metadata sources including Europeana
The recent introduction of Internet connected hearing instruments offers a paradigm shift in hearing instrument fitting. Potentially this makes it possible for devices to adapt their settings to a ...changing context, inferred from user interactions. In a pilot study we enabled hearing instrument users to remotely enhance auditory focus and attenuate background noise to improve speech intelligibility. N=5, participants changed program settings and adjusted volume on their hearing instruments using their smartphones. We found that individual behavioral patterns affected the usage of the devices. A significant difference between program usage, and weekdays versus weekends, were found. Users not only changed programs to modify aspects of directionality and noise reduction, but also continuously adjusted the volume. Rethinking hearing instruments as devices that adaptively learn behavioral patterns based on user interaction, might provide a degree of personalization that has not been feasible due to lack of audiological resources.
Despite the technological advancement of modern hearing aids, many users leave their devices unused due to little perceived benefit. This problem arises from the limitations of the current fitting ...procedure that rarely takes into account 1) the perceptual differences between users not explained by measurable hearing loss characteristics and 2) the variation in context-specific preferences within individuals. However, the recent emergence of smartphone-connected hearings aids opens the door to a new level of context awareness that can facilitate dynamic adaptation of settings to users' changing needs. In this position paper, we discuss how user auditory intents could be modeled as context collected via mobile devices and suggest what kinds of contextual information are relevant when learning situation-specific intents and the corresponding preferences of hearing impaired users. Finally, we illustrate our ideas with several examples of real-life situations experienced by subjects from our study.
Provider: - Institution: - Data provided by Europeana Collections- Tegnet af Joakim Skovgaard og med tekst af Henrik Pontoppidan.- All metadata published by Europeana are available free of ...restriction under the Creative Commons CC0 1.0 Universal Public Domain Dedication. However, Europeana requests that you actively acknowledge and give attribution to all metadata sources including Europeana
Provider: - Institution: - Data provided by Europeana Collections- Europeana Collections 1914-1918 Tegnet af Joakim Skovgaard og med tekst af Henrik Pontoppidan.- All metadata published by Europeana ...are available free of restriction under the Creative Commons CC0 1.0 Universal Public Domain Dedication. However, Europeana requests that you actively acknowledge and give attribution to all metadata sources including Europeana
PDF
