Browse Prior Art Database

Method for sound classification and "noise" reduction in an augmented reality environment Disclosure Number: IPCOM000199422D
Publication Date: 2010-Sep-02
Document File: 1 page(s) / 42K

Publishing Venue

The Prior Art Database


The solution disclosed detects that something is making a noise and uses pattern matching to identify what that noise might be. It then merges this information with additional contextual information in order to determine what object might be making the noise. Eg it might be determined that the noise sounds like a piano and a camera feed might indicate that there is a piano in the direction the sound originates from. Additional contextual/environmental information can then be used to identify the importance of a sound emitting object. For example, if the user is by a busy road, then you might not want to indicate to the user that the cars are all making a noise, you might however want to make them aware of a passing ambulance with blaring siren. Equally, it may not be important to make the user aware of the humming noise being emitted from the air conditioning unit. Because you can now determine the identity and state (both visually and aurally) of the object, you can determine whether it is important for a deaf user to know about that noise. This solution has particular use for indicating auditory information to a deaf person, who can see objects but not hear the sounds they emit.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 56% of the total text.

Page 1 of 1

Method for sound classification and "noise" reduction in an augmented reality environment

Hearing impaired people lack auditory information from their environment which other people take for granted. An augmented reality system can be used to provide extra information to these users by representing auditory information in the system as visual or haptic feedback.

    Augmented reality systems exist which can display the presence of sound sources, however these systems are subject to problems where information is overly plentiful (ie. "noisy"). For example, the system cannot distinguish between important and unimportant sources, and cannot provide information about the source. This limits the applicability of the system to controlled environments, and is not so helpful to hearing impaired users in the real world.

    Partial solutions to this problem, such as restricting presentation to high-volume auditory information, are of limited application as information is lost.

    This improves over the current state of the art by allowing the full range of auditory sources to be represented to the user. In addition, the information is classified and filtered to prevent representation of unimportant information.

Assume the presence of an existing system which can isolate discrete sound sources and provide their relative directions/locations.

The disclosed system works in the following manner:

Uses pattern matching to distinguish the sound type and source (music,



ackhammer, car, emergency vehicle's siren, etc.). This information is added to the AR system's metadata for the sound source and may influence the AR system's display method for the sound.

Matches the sou...