News Article

Additional Algorithms Patented for SysTune

2015-08-06 14:14

AFMG has been granted another patent for the algorithms used in our real time audio measurement software SysTune. (Patent No.: US 9,060,222 B2)

AFMG is proud to announce that another patent has been granted for the sophisticated algorithms we developed for SysTune, our professional audio measurement software.

Method for determining an averaged frequency-dependent
transmission function for a disturbed linear time-invariant system,
evaluation device and computer program product.


We simply call it the SSA Filter.


When is the SSA Filter Useful?

With the SSA Filter, AFMG has come up with a unique tool to help users achieve reliable, stable measurements even in noisy environments and when using a measurement signal like speech or music which does not continuously provide sufficient signal content on all frequencies. Usually, this is the case during soundcheck while others keep working and making noise or during live performances with frequent announcements or loud audiences.

What Makes the SSA Filter so Helpful?

The best thing about the SSA Filter is that it will make sure that no irrelevant information (e.g. noises from truss hammering, a starting fork lift or a screaming audience) will contaminate the measurement. The user can simply continue measuring while SysTune will analyze each measurement cycle according to three different criteria to distinguish bad data from good, valid measurements. Only the good data will be used for further processing.

How Does the SSA Filter Work?

The SSA Filter consists of three filter stages. In general, SysTune measures the transfer function of a system continuously. Each time new measurement data is available, the current estimate of the transfer function is updated. When the SSA Filter is active, only those frequencies are updated where the new data passes through all three filter stages. The frequencies of the new data which do not pass the filters are considered dominated by noise and are rejected. This way, some measurement cycles may not yield any new valid data but, most importantly, existing older data is not spoiled by invalid new data.

The first filter stage is the so called Signal Threshold Filter. It filters out all frequencies where the power density of the excitation signal is below some user-determined threshold. All measured signal content on frequencies which are not significantly excited is likely due to some external noise and is therefore ignored.

The second filter stage is the Excursion Filter. This filter assumes that very fast changes to the transfer function of the system are unlikely. It rejects the new data for all frequencies that would cause a sudden, large change in the transfer function. Reasons for measurements that show sudden differences to the data collected before might be a slamming door, a falling stack of chairs, or a spontaneous cheer of the crowd.

The third filter is called the Coherence Filter. The coherence is a parameter to indicate how closely the measured signal correlates to the signal used for excitation. The Coherence Filter rejects new data for all frequencies where the coherence is below a user-defined threshold. For these frequencies, the measurement results can be expected to be strongly contaminated by some noise uncorrelated to the excitation signal. Consequently, the frequencies will be disregarded and not enter the averaging with the previous measurements.

Together, these three filter stages form a valuable tool to help sound engineers do an efficient and reliable job. Continuous measurements do not need to be re-done when disturbing noises appear, invalid data will automatically be rejected whilst at the same time not the full measurement cycle will be lost. Good data will still accumulate on those frequencies that pass the filter. The combined filter algorithm was therefore named Spectrally Selective Accumulation (SSA) Filter.

(Patent No.: US 9,060,222 B2)

Go back