Specific outcomes from the project

Click on a project title to view more detailed information

Signal properties

• Measurement of signal or impulse response
  • The characteristics of the signal have an important effect on the perceived width, and measurement of solely the impulse response does not reflect this.
  • [Mason et al 2002]
• Division of source and environment related aspects based on time or perceptual streaming
  • The division of source and environment related aspects of the signal based upon a single time point does not match perception well. More accurate results can be obtained by dividing the signal based upon perceptual streaming, separating segments perceived to be either a source or reverberation and measuring these individually.
  • [Mason et al 2002]
• The effect of source signal properties on IACC-based measurements
  • Source signals that have different properties can result in differing measured interaural cross-correlation coefficient (IACC) values when they are reproduced in an identical manner in an acoustical environment. Both the duration and the spectral content of the signal can affect the results. This supports the argument that IACC-based measurements made of impulse responses are dissimilar to those made of most musical signals.
  • [Mason et al 2005b]
• Characteristics of musical signals that affect the measured IACC in a room
  • The various properties of musical signals and their relationship with the measured IACC in a room were examined. It was found that the overall duration of the source signal, its onset and offset times, and the magnitude and rate of any spectral fluctuations, have a very strong effect on the IACC.
  • [Neher et al 2006]
• A musically representative test signal that can be used for making practical measurements of auditoria
  • Based on the information shown above, a set of test signals has been created that vary the parameters of musical signals that affect the IACC across a musically relevant range. This allows for convenient and rapid analysis of the range of spatial perception that is possible in a given room.
  • [Neher et al 2006]
• Effect of lateral position within an auditorium on IACC-based measurements
  • Previous research has argued that the IACC is not a good predictor of the perceived auditory source width as it can vary a large amount as the received is moved laterally around the centre of an auditorium, whereas musical signals do not vary a large amount in width in these positions. This is due to the fact that previous measurements were made of impulse responses rather than the musical signals that were auditioned. When the IACC-based measurements are made of the musical signals themselves, the predictions match the perceived effect much more closely.
  • [To be reported]
• Controlled experiment stimuli
  • Stimuli with tightly controlled parameters are required for research into auditory perception; suitable novel stimuli have been created and used successfully in a series of experiments.
  • [Mason et al 2003]

Frequency dependent properties

• Simulation of breakdown of phase-locking required in measurement model
  • The necessity of a simulation of the breakdown of phase-locking in a model of perceived width has been established, and a suitable rectification and filtering process derived.
  • [Mason et al 2005a]
• Effect of IACC on width at high frequencies
  • Previous research has indicated that the IACC of a signal does not affect the perceived width of a signal above a few kHz. However, it was found that the IACC of the signal or the envelope of the signal can affect the perceived width across the whole of the audio range.
  • [Mason et al 2005a]
• Width of diotic stimuli dependent on frequency
  • The frequency dependence of the perceived width of diotic stimuli (signals identical at both ears) has been determined, and an equation derived to model this in the measurement.
  • [Mason et al 2005a]
• Relationship between width and IACC at low frequencies dependent on frequency
  • It has been found that at low frequencies the relationship between width and IACC is dependent on frequency. To model this, an equation has been derived from the expected range of IACC values in natural sound fields at low frequencies.
  • [Mason et al 2005a]

Loudness dependent properties

• Relationship between width and IACC dependent on loudness
  • The loudness dependence of the perceived width of stimuli has been determined, and an equation derived to model this.
  • [To be reported]

Temporal properties

• Optimum measurement window length
  • The optimum window length to use in the IACC measurements has been determined.
  • [Mason et al 2003]
• Temporal resolution of IACC perception
  • Rapid increases in the perceived width of a sound (i.e. decreases in the IACC) can be perceived more easily than rapid decreases in the width of the sound (i.e. increases in the IACC). This has been implemented in the model as a dynamically varying temporal filter.
  • [To be reported]

Display and interpretation factors

• Method to compensate for differences between stimuli
  • Simple IACC-based measurements are not suitable for making comparable measurements of different stimuli, as the frequency content and loudness of the stimuli also affects the width. By deriving methods to compensate for these factors as discussed above, comparable measurements are now possible.
  • [Mason et al 2004a]
• Combination of location and width detection
  • The relationship between the measured source width and the measured location in terms of the interaural time difference has been determined. This allows measurements of perceived location and width to be combined for a more complete spatial analysis.
  • [Mason et al 2004c]
• Effect of IACC on ITD and vice versa
  • The interaction between the interaural time difference (ITD) and the interaural cross-correlation coefficient (IACC) was investigated. It was found that the ITD does not affect the perceived auditory source width unless double peaks occur as discussed below, and that decreasing the IACC causes the sound to move towards the median plane. Methods were derived to model this.
  • [Mason et al 2004c]
• Effect of double peaks in the IACC measurement on the perceived auditory source width
  • For narrow-band signals of around 500 to 1000 Hz with a fairly large ITD or higher frequency signals with any ITD, it is possible for two peaks to be detected in the IACC analysis. It was found that this can affect the perceived auditory source width, and a method was derived to take this into account in the measurement.
  • [Mason et al 2004c]
• Conversion of measured results to perceptually-relevant scale
  • Once the measured source width and location are on a common scale (using the conversion discussed above), it is possible to display the results on a more perceptually relevant scale, such as perceived angle. This is more intuitive and useful than existing IACC-based metrics that provide results as an abstract coefficient.
  • [Mason et al 2004d]
• Integration of measured results across frequency
  • Preliminary research has been conducted into the integration of the measured results across frequency. Further research is required before a suitable model can be developed.
  • [To be reported]
• Creation of display techniques
  • A novel method of displaying the results of the predicted auditory source width over time and frequency has been derived. This allows the user to select the frequency and time segments for analysis in an interactive and intuative manner.
  • [To be reported]

Application and evaluation

• Practical concert hall evaluation method
  • Based on the investigation into suitable musically relevant test signals and the development of the IACC-based binaural hearing model, a practical method for conducting measurements of the perceived auditory source and environment width of auditoria has been derived.
  • [To be reported]
• Practical 2-channel stereo reproduction system measurement technique
  • Based on the results shown above, a practical technique for evaluating the spatial performance of 2-channel reproduction systems has been developed. This includes synthetic test signals as well as an implementation of the hearing model with convenient display options.
  • [To be reported]
• Practical low bit-rate codec measurement technique
  • The method described above to analyse the performance of reproduction systems can also be modified to analyse the performance of various low bit-rate codecs.
  • [To be reported]
• Evaluation of frequency and loudness compensations
  • The techniques to compensate for differences in frequency and loudness between stimuli discussed above were evaluated. This indicated that these methods improved the accuracy of the perceived width measurements by a large amount.
  • [Mason et al 2004d]
• Evaluation of modified measurement model with musical stimuli
  • The modifications that were made to the IACC-based binaural hearing model were derived using highly controlled artificial stimuli. The applicability of these results to more externally valid musical stimuli was tested by evaluating the resulting measurement model. This showed that the model accurately predicted the perceived width of relatively simple single musical stimuli.
  • [Mason et al 2004d]
• Evaluation of derived test signals
  • A set of test signals for use in measurements of concert hall acoustics have been derived as discussed above. Evaluations were conducted to determine how well IACC-based measurements made of these matched equivalent measurements of musical signals. It was found that they provided a reasonable match to a range of musical signals in a range of acoustical environments, and the match was much better than measurements made of impulse responses.
  • [Neher et al 2006]
• Evaluation of display and analysis methods
  • The display and analysis methods developed above were evaluated through use by the experimenters and through evaluation by the industrial collaborators.
  • [To be reported]
• Evaluation of binaural capture techniques
  • It was considered that the requirement for the binaural hearing model to use signals captured using a dummy head would be restrictive to its widespread implementation. It was found that small omnidirectional microphones mounted on the sides of a human head, or a pair of spaced cardioid microphones could be used as an alternative, at least at low frequencies.
  • [To be reported]

Other work

• Perception of interaurally phase inverted signals
  • Existing IACC-based measurements usually take the maximum absolute value across the range of tau of +- 1 ms. An experiment was conducted that examined the perceived effect of signals that were interaurally phase-inverted (i.e. had an IACC of -1). This indicated that the perceived auditory source width of signals with negative IACC values is different to those with positive IACC values.
  • [Mason & Brookes 2006]