Categories :

Amplitude a/b Testing: Explained

Amplitude a/b testing is a method used to determine the deviation from the analytical curve of a product in an acoustical environment. An amplitude a/b test is a value obtained for a system in which the product has been biased against time and its phase. 

In this kind of testing the deviation from the analytic curve can be detected using a low-pass filter with high bandwidth. The low-pass filter will remove all sounds of interest, thus only the amplitude a/b changes can be observed. The result of amplitude a/b testing is the amplitude response envelope, which is compared to the amplitude response envelopes of reference products. 

What is a/b testing used for?

Amplitude a/b testing is a very convenient and economical technique when comparing two or more signals or waveforms. The method is ideal for finding the deviation from a prescribed function at the origin. In addition, amplitude a/b testing can also detect and measure any changes that occur after a signal passes through a filter. For example, if a signal passes through a low-pass filter with low bandwidth, the change in amplitude a/b (the “pulse width modulation”) will be a measure of the product’s phase shift.

Amplitude experiment analysis

It is important to know that a/b testing can also detect and measure any changes that occur after a signal passes through a low-pass filter. For instance, an alarm clock that goes off at 2 a.m. can generate a sinusoidal signal that can decay exponentially. To find out if the product has any phase shift, amplitude a/b testing can be used. The test can determine if the signal decaying exponentially is shifted out of the band’s range by a factor of ten.

For amplitude a/b testing to be successful, there are a few conditions that must be met: 

  1. First, the system should be operating at the analytical level. Most systems will meet this condition because the amplitude a/b produced is directly proportional to the input signal. 
  2. Second, a known time frame for measuring the result should be chosen. There are many time frames to choose from, ranging from milliseconds to microseconds. 
  3. A third option is to use a moving average, which is a period following a trend that can be determined by plotting a series of moving averages.

Loudspeaker Measurements Testing

Amplitude a/b testing can be used to validate impulse responses for loudspeaker measurements. A low-pass filter with high bandwidth is derived from the amplitude a/b test. Amplitude match deviations are determined by the use of statistical analysis methods such as t-test or F-tests. Compliance with standards can be verified using amplitude a/b testing according to IEC standards 61672 and ISO 3382.

Amplitude a/b testing can be used to validate loudspeaker impulse responses by comparing amplitude deviations of the amplitude response with amplitude deviations of an amplitude reference. The key is:

  • that amplitude reference and real-time response are derived independently of each other
  • this prevents errors due to time delays in amplitude responses or amplitude references.

This is quite important for obtaining correct matching results because some echoes are not present in the amplitude reference but are part of the original signal.

How to successfully use a/b testing

To use the amplitude a/b test to your advantage, there are a few simple steps to take. Firstly, the high and low sides of the sample waveform should be measured. Then, the amplitude of the high side of the waveform should be compared to the amplitude of the low side of the waveform. The sign comparison should be done between different waveforms, as it is possible to make some significant deviations from the true signal due to an inaccurately measured high or low side. For best results, test signals from at least three different channels.

How to successfully use a/b testing

The amplitude a/b test results should be plotted onto a graph and the square of the amplitude against the y axis of the display should be plotted as a negative slope. If the slope of this straight line is negative, the control signal is above (past) the amplitude a/b curve. The test signal should also be plotted against the slope of another line, called the zero-point curve. If the zero-point curve is flat, the control signal is below (past) the amplitude a/b curve.

A good technique for amplitude a/b testing is called the drawn-out plot. This technique involves drawing a line from the highest peak to the lowest peak of the amplitude a/b curve. This allows a more clear see of the change in amplitude over time. To do this, the range of the plotted point should be plotted on a separate axis. It can then be shifted left or right to get a better view of the change in amplitude.