Discrete-frequency axial-flow fan noise reduction using active noise control is described. The unique aspect of the current research is the use of the fan itself as the antinoise source in the active noise control scheme. This is achieved by driving the entire fan unit axially with an electrodynamic shaker which mechanically couples the solid surfaces of the fan to the acoustic medium. The fan unit is thus transformed into a crude loudspeaker. A near-field microphone serves as an error sensor, where transfer function measurements between the electrical input to the shaker and the electrical output of the microphone are found to be reasonably free of phase distortions and linear. A feedforward algorithm utilizing the output of a tachometer as a reference signal is used. The experimental apparatus is composed of a baffled fan unit in a free field. A small cylindrical flow obstruction is placed on the inlet side of the fan to enhance noise emissions at the blade-pass frequency and harmonics. The experiment successfully demonstrates the concept of active control of tonal fan noise using a shaken fan as the cancellation source. For the fan operating in a planar baffle, the fundamental blade-passage frequency sound-pressure level at the location of the error sensor is reduced by 20 dB, while the second and third harmonic levels are reduced by 15 and 8 dB, respectively. Placing a cabinet enclosure over the baffled fan did not affect these results significantly, and free-field sound power measurements indicate similar level reductions with the active control in operation.

High-level spontaneous otoacoustic emissions (SOAEs) were measured from 16 ears using both spectral and time averaging. The purpose was to determine the source of an upward shift in frequencies of synchronized SOAEs (SSOAEs) observed while using a subroutine of the ILO88 system of Otodynamics Ltd. An HP3561A signal analyzer performed spectral averaging to extract SOAEs with no external stimulation applied to the ear canal. Synchronized SOAEs were derived using the ILO88 system performing time averaging following click stimulation. The frequencies of all SSOAEs were shifted upwards by 6 to 21 Hz when compared to corresponding SOAE frequencies determined with spectral averaging. Additional measurements of signals in a cavity and of click-evoked otoacoustic emissions in selected ears indicated that the frequency shift is the result of an error in the ILO88 software. Incorrect cursor readouts in the program cause an apparent upward shift in frequency of 12.2 Hz. This error was confirmed by the manufacturer.Spontaneous otoacoustic emissions ͑SOAEs͒ represent narrow-band signals that can be recorded in the outer ear canal when no external acoustic stimulation is presented ͑see Probst et al., 1991 for a review͒. In general, two methods have been used to record them. In the first, the sound-pressure level in the ear canal is measured by a low-noise microphone with no stimulation applied. The microphone signal is averaged in the frequency domain ͑e.g., Whitehead et al., 1993͒. The second method consists of recording SOAEs synchronized by acoustic stimuli, for example clicks, using averaging in the time domain. This enables the detection of long lasting oscillations following click-evoked otoacoustic emissions ͑CEOAEs͒. It has been shown that for an ear with strong SOAEs, a CEOAE spectrum exhibits peaks corresponding to SOAE frequencies ͑Probst et al., 1986; Gobsch and Tietze, 1993͒. Software of a widely used commercially available instrument for measuring OAEs, the ILO88 ͑Otodynamics Ltd., Hatfield, UK͒, includes a subroutine for measuring synchronized spontaneous otoacoustic emissions ͑SSOAEs͒. Several recent studies have reported SSOAE data collected with the ILO88 system ͑Wable and Collet, 1994; Kulawiec and Orlando, 1995; Prieve and Falter, 1995͒. As part of an ongoing study of otoacoustic emissions in normal-hearing humans in our laboratory, we have measured SOAEs using both spectral averaging and the synchronization technique of the ILO88. In comparing the two results from the same ear, we have observed a slight but consistent difference in the frequencies of SSOAEs and SOAEs. Therefore, we sought to characterize this discrepancy further and to determine its source. Because of the widespread use of the ILO system, we believe that it is important that our findings be reported.Both ears of eight subjects from our laboratory pool who had known SOAEs that were at least 10 dB above the noise floor of the instrumentation were tested with two methods. In the first method, the sound-pressure level in the ear canal was me...

Far-field radiation of a baffled rectangular piston is described using a frequency-domain Rayleigh integral solution. The piston model is driven at a single frequency at two opposing edges, where these individual edge sources are independent of one another in both amplitude and phase. Directivity plots of the analytical solution are used to graphically observe directional effects caused by changing the phase and amplitude of these edge sources. Interesting and unexpected beamforming characteristics are shown to be produced by certain amplitude and phase combinations. [Work supported by IBM through the Shared University Research program and the Applied Research Laboratory at Penn State Univ. University.]

Many methods of auralization convolve a source signal (e.g., cello recorded in an anechoic room) with a room’s impulse response (which has been computed using method of images, ray tracing, etc.). Many instruments are finite-sized sources because they produce music having frequencies where the product of the wavenumber and the instrument’s characteristic length is not small. Sound produced by a finite-sized source in the presence of boundaries can include scattering and diffraction, resulting from the presence of the source in its own field. These effects are not accounted for by the auralization types mentioned above. A geometrically simple example of a finite-sized pulsating sphere in the presence of a rigid infinite boundary is solved using the translational addition theorem for spherical wave functions (TATSWF). Using TATSWF, the original problem is solved by replacing the rigid infinite wall with an image of the finite-sized sphere. This is a surprisingly complicated problem to solve, given the simple geometry, and serves to illustrate how a source can perturb its field when near a boundary. Examples are presented for which significant changes in the pressure magnitude occur. [Work supported by the Applied Research Laboratory, Penn State.]

In an effort to develop a model for impulse responses in reverberant spaces, techniques for source radiation modeling were explored. One method of decomposing a source field involves an analysis based on spherical harmonic expansions of the field. Spherical harmonic decomposition is a powerful tool, but it is a mathematical approach which is completely independent of the type of source that creates the field. As a consequence, this technique cannot, in general, be used to gain insight into the underlying physical nature of the source. Moreover, because this is a brute force technique, the analysis can be extremely inefficient. In this presentation, modifications to the spherical harmonic analysis and synthesis equations are made to overcome these problems. These are based on the (theoretical) source types associated with the individual spherical harmonic expansion coefficients, and determination of the location of the primary components of the (physical) source. These modifications can greatly improve the computational efficiency and allow the source to be described by a set of coefficients which is more directly related to actual source types present. Illustrative examples will be shown. [Work supported by the Applied Research Laboratory at Penn State University under the Educational and Foundational Research Program.]