2001 April, Vol 49 Number 4 |
CONTENT
PAPERS
Extracting Room Reverberation Time from Speech Using Artificial Neural Networks
T. J. Cox, F. Li, and P. Darlington 219A novel method to extract reverberation time from reverberated speech utterances that are limited to pronounced digits is presented. The reverberation times considered are wide band, within the frequency range of speech utterances. A multilayer feed forward neural network is trained on speech examples with known reverberation times generated by a room simulator. The trained neural networks are able to extract room reverberation times from speech signals picked up in the room to an accuracy of 0.1 seconds. This is an alternative to traditional measurement methods and facilitates occupied measurement of room reverberation times.
Introduction to Head-Related Transfer Functions (HRTFs): Representations of HRTFs in Time, Frequency, and Space
Corey I. Cheng and Gregory H. Wakefield 231This tutorial introduces head-related transfer functions (HRTFs) and describes their role in the synthesis of spatial sound over headphones. HRTFs are defined and their importance in clarifying free-field sound's spatial location clues is shown. Also included are measurement methods and applications to spatialized sound synthesis. Different analytic and signal processing techniques are highlighted by comparing and contrasting representations of HRTFs in time, frequency, and spatial domains.
Perceptual Wavetable Matching for Synthesis of Musical Instrument Tones
Cheuk-Wai Wun and Andrew Horner 250Until recently multiple wavetable synthesis has used a simple relative spectral error formula to measure how accurately the synthetic spectrum matches an original spectrum. A modified error formula takes into account human auditory masking characteristics. It is an improved measure of the perceived quality of the matched spectrum. Listening tests indicate that the perceptual error formula slightly outperforms previous matching methods, especially for instrument tones having several masked partials.
ENGINEERING REPORTS
Signal Mixing for a 5.1-Channel Surround Sound System--Analysis and Experiment
Xie Bosun 263A new signal mixing method for 5.1-channel surround sound is proposed in which the lateral signal becomes stabilized. This method claims a stable 360o sound image within a frequency range up to 800 Hz. It is proposed that the mixing method from the earlier pairwise mixing technique be used to recreate a stable front image while the new mixing method be used to create stable lateral and rear images. The method employed is dependent upon which sound image is optimal for presentation to the listener.
STANDARDS AND INFORMATION DOCUMENTS
AES Standards Committee News 275
Audio metadata; listening tests; intelligibilityCall for Comment on DRAFT AES31-1-xxxx, AES standard for network file transfer of audio -- Audio-file transfer and exchange -- Part 1: Disk format 275
Call for Comment on DRAFT REVISED AES26-xxxx, Second posting of DRAFT REVISED AES recommended practice for professional audio interconnection -- Conservation of the polarity of audio signals 276
Call for Comment on DRAFT REVISED AES10-xxxx, AES recommended practice for di gital audio engineering -- Serial Mutichannel Audio Digital Interface (MADI) 276
Call for Comment on DRAFT AES28-1997 Amendment-1-xxxx to AES standard for audio preservation and restoration -- Method for estimating life expectancy of compact discs (CD-ROM), based on effects of temperature and relative humidity 276
FEATURES
19th Conference Preview, Bavaria, Germany 282
Calendar 284
Program 286
Registration Form 294Updates to Review of Society's Sustaining Members 296
DEPARTMENTS
Review of Acoustical Patents 279
News of the Sections 297
Upcoming Meetings 301
Sound Track 302
Available Literature 303
New Products and Developments 304
Membership Information 306
Advertiser Internet Directory 308
Sections Contacts Directory 314
AES Conventions and Conferences 320
FPO For the cover:
2001 April, Vol 49 Number 4
spine: 2001 April, Vol 49 Number 4