 |
||
 |
Acoustical Information Measured Arrays: SPHEREARRAY2 and UPLARRAY2
| Omni Loudspeaker | Walls
| Sound Field Scaling
MAPP Online provides two measured arrays that can be used to verify the program's accuracy: a spaced pair of sphere loudspeakers, and a spaced parallel array pair of Meyer Sound UPL-1s. In each case, the array data were taken with an actual pair of loudspeakers, measured with the standard procedure used for all MAPP data sets, and the data files present each array as a single object. The sphere loudspeakers are 20-inch diameter spheres, each with a four-inch cone driver installed flush with the sphere's surface. These spheres are of a special type, fabricated to an exact spherical shape with very close tolerances. In the sphere array, their geometric centers are spaced 24 inches apart. In the UPL-1 array, the loudspeakers are spaced with their geometric centers 18 inches apart. You can use these data to see for yourself how accurately MAPP Online models the actual acoustical behavior of real loudspeakers. Simply reconstruct each array in MAPP Online using the single-loudspeaker data, launch a prediction and print the result to a color printer. Then, load the array data and do the same. The Omni source represents an ideal theoretical point source with perfectly omni-directional characteristics and perfectly flat frequency and phase response. The data are scaled to match the procedure used to measure actual loudspeakers. MAPP Online allows you to place omni sources as close as .001 inch apart. The Omni source is represented on-screen as a 6-inch sphere so that it will be visible, but Omnis cannot be selected by clicking on them because the actual source is infinitesimally small. Currently the walls option models only the first reflection from each wall by generating a single image source for each loudspeaker on the other side of each wall. Wall materials are modeled by multiplying the frequency response of each image loudspeaker by a magnitude only frequency response representing the material. The frequency response of each material is interpolated up from octave band absorption values taken from standard acoustical texts. The initials at the end of the material name indicate which book they are from: CH = Cyril Harris, "Handbook of Acoustical Measurements and Noise Control", Third Edition, Acoustical Society of America, 1998. LB = Leo Beranek, "Acoustics", Acoustical Society of America, 1996. KF = Kinsler, L, Frey, A, et al., "Fundamentals of Acoustics", Fourth Edition, John Wiley and Sons, 2000. Sound field predictions are normalized such that the largest value in the sound field is set to 0 dB. The dynamic range of the sound field plot is 42 dB. Positions in the sound field that are more than 42 dB down from the maximum are displayed with the darkest color blue from the colorbar. Frequency Response data is normalized to 0dB at 1000Hz. This is important to keep in mind when comparing two or more frequency responses. Because the Third Octave Band Spectrum is displayed in absolute dB SPL it is often more intuitive to compare Band Spectra. The apparatus used to measure the polar response of each loudspeaker model consists of a Meyer Sound SIM(r) System II analyzer, a B&K 4133 omni-directional measurement microphone, a B&K 2639 preamplifier, and a B&K 2807 power supply. Each loudspeaker is placed in an anechoic chamber on an automated turntable positioner (accurate to within 0.1 degrees rotation), with the microphone placed 4 meters away from the geometric center of the loudspeaker. The turntable is rotated through a full 360-degree arc at 1-degree intervals. This measurement is performed once along the horizontal on-axis plane, and again (separately) along the vertical on-axis plane. A piecewise approximation to a constant-Q transform is utilized in the measurement so that the frequency resolution is consistent across the full frequency range. This transform affords greater than 1/36th octave resolution from 20 Hz to 20 kHz. The microphone to preamp sensitivity is 11.5 mV / pa (or -38.8 dBV @ 94 dB SPL, when expressed logarithmically). This sensitivity becomes part of the measured polar response transfer function data that MAPP Online(TM) uses for each loudspeaker model. The Meyer Sound loudspeakers available in MAPP Online have had their maximum linear frequency weighted, slow time weighted, average SPL measured at 2 meters while being driven with pink noise at the onset of limiting. This is done by driving the loudspeaker with pink noise at a higher and higher level and watching the level of the frequency response in SIM II. Because a frequency response measures the ratio of output to input, if the loudspeaker is functioning linearly, increasing the input will cause a proportional increase in the output and the frequency response will not change. As the input level is increased you eventually reach a point where the speaker can no longer reproduce the highest peaks. Because the input increased, but the output doesn't increase due to limiting, the frequency response decreases. When a loudspeaker's maximum average SPL is measured, the level of the pink noise is increased until the decrease in frequency response just begins to happen. The linear frequency weighted, slow time weighted, average SPL of that speaker is measured for that drive level. The sound level meter is then switched to Peak mode to ensure that peak to average (crest factor) of the output sound is the same as the crest factor of the input pink noise. Because the crest factor of SIM II pink noise is 12.5 dB, the Peak SPL reported by MAPP Online is always 12.5 dB higher than the average. However, this peak SPL computation is invalid when the prediction contains only subwoofers, since subwoofers are band limited to the lower octaves of the audio spectrum. The peak SPL computation for a loudspeaker plus subwoofers is correct. We are working on an improved peak SPL computation algorithm that could recognize band-limited systems (e.g. subwoofer, subwoofer arrays, low-passed loudspeakers, etc.) and correctly predict the peak SPL based on the operating bandwidth. The peak level reported by a loudspeaker in MAPP will always be lower than the peak level reported in the datasheet for that loudspeaker. The peak level reported in the datasheet is measured using a short burst of music. Meyer Sound loudspeakers can reproduce short bursts of high power before the limiters engage. For this reason the peak level for music reported in the datasheet will always be higher than the peak level for pink noise which is reported by MAPP Online. MAPP Online has been calibrated such that a virtual microphone placed on axis to a virtual speaker set to a Relative Level of 0dB will report the same average and peak SPL as the actual loudspeaker when driven with pink noise at the onset of limiting. The omni source and sphere source have been set to produce 94dB at one meter. |
Help Contents |
