We're pleased that you want to share this information. To send a link to this page, fill in the information below. The e-mail will show you as the sender and will show your return address.
* indicates required information
* Recipient E-mail:
Each address entered must be valid.
(Separate multiple email addresses with commas. Limited to 5 addresses.)
* Your Name Here:
Enter your name.
* Your E-mail:
Enter a valid address.
send a copy to me
Note: You are not adding recipient email addresses to any promotional email list.
Multimedia consists of content with both video and audio information. While there has been no shortage of effort in analyzing the visual aspects of multimedia in all its forms, the question of what can be done with sound is beginning to gain momentum. Sensing, capture, and analysis of sound in its many forms are becoming important for a variety of applications.
Kodak is working on finding ways to take advantage of sound. If we capture sound at a quality on par with that of today's video-enabled digital cameras, many interesting and useful applications become possible. A capture device with appropriately arranged sensors (microphones) provides audio content that can be analyzed for speech-centric applications and audio classification. By analyzing the captured sound, we gain an understanding of the acoustic environment as well as information about sound sources in the environment, forming a basis from which semantic inferences can be made.
The work being done in this field should provide audio capture capabilities that will provide rich raw data. From this high-quality sound, advanced signal processing and statistical analysis techniques can be used to improve speech recognition, reduce noise, and automatically adjust sound or image capture parameters.
Through speech-to-text or natural language processing, semantic audio information can be captured to provide meaningful tags or captions of the associated image or video clip. Audio classification can also provide contextual information about the location or environment of the captured scene or event.