AP - The Recognized Standard In Audio Test

Speaker testing is challenging because extraneous noises and reflections from surrounding walls and floors interfer with the acoustic measurements. The ideal testing setup is a dedicated acoustic tester in an anechoic chamber (or a remote, windless location such as a desert), where there are no reflections or other disturbances. However, the cost and time for this kind of setup is impossible to justify for the vast majority of engineers, especially if the need is only occasional.

Most engineers who want to make acoustic measurements wish to do so in their lab, or even on a noisy production line. For these situations, two main techniques have been developed to negate the effect of reflections: MLS and time gated log chirp (continuous sweep).

MLS (Maximum Length Sequence) technique works by deriving frequency response from impulse response before reflections have had a chance to bounce back from the nearest surface to the measurment mic. However, a key drawback is that MLS isn't able to measure distortion, and in fact a driver's distortion can end up corrupting its frequency response measurements.

Time gated chirp (continuous sweep) is a better approach. Again, the measurements are derived from an impulse response, however, continous sweep returns a larger set of measurements including level, frequency response, phase, distortion, and group delay. Signal to noise ratio is also better because the crest factor of a sine wave is lower than the pink noise used in MLS. With the APx's implementation of continuous sweep, 14 graphed measurements are returned within six seconds. In addition, the measurement provides synchronous averaging and Nth-octave smoothing of results.

Setting up an acoustic test

Setting up an acoustic test is the biggest challenge in measuring loudspeakers because it's hard to determine exactly when the first reflection hits the measurement microphone (and thus pollutes all subsequent signals). The latest APx software uses drag and drop cursors on an Energy Time Graph to define where the time gate should start. APx also automatically calculates distances to double check the measurement correctly. Measurement results are recalculated based on the new time window and regraphed in real time.

Common speaker tests include frequency response, sensitivity, maximum SPL, phase, group delay, and impedance. The APx Series and 2700 Series both feature extensive programming capabilities for automated production line testing, including advanced reporting, pass/fail limit setting, and control of turntables and other external equipment.

Common Challenges

• Even with chirp or MLS, the ability to test low frequencies is directly related to room size: by the time low frequencies are reaching the microphone, higher (faster) frequencies have had the opportunity to reach a wall and reflect. For example to measure a 200 Hz response accurately, the nearest surface (including the floor!) must be at least 30 feet from the microphone.
• Speaker and test microphone placement are critical. Depending on the proximity of surfaces, and the distance between the speaker and test microphone, a variation of only an inch can change the results.
• Real world conditions: While eliminating reflections produces predictable results that can reveal driver deficiencies, it doesn't reproduce the acoustic environment in the living room or concert hall where the speaker will actually be listened to.

Tips for Optimum Testing

• Compare to a "known good unit" or "Golden Unit." While you can't always have a perfect test environment, on a production line you can compare unknown units to a known good unit, to identify defective samples.
• Use an acoustic coupler to test headphones. While the response curve may be an approximation of what a real human would hear, it's still a good way, using the Golden Unit method above, to check for a defective or damaged driver.