What is the SyncSource™ Transfer Function?

A hallmark of the transfer function measurement process embodied in Smaart is the ability to measure a system’s response using whatever random signal is passing through it, commonly known as "Source Independent Measurement". This allows engineers to measure their systems under actual use conditions, with signals like speech and music, allowing for simultaneous listening evaluation. Moreover, it allows engineers to measure systems without the need to control the input signals, often a practical restriction.

However, transfer function measurement with random signals does involve certain compromises, most notably the need to impose a data window on the measurement signals to make them suitable for the FFT algorithm. The FFT math assumes that the input signal is the entire input signal, or alternatively, a single, complete cycle of an infinitely repeating signal.

Raw random signals do not meet this criterion, resulting in the introduction of a significant amount of mathematical noise and instability into the output data. Applying a data window greatly reduces the severity of both issues by conditioning the incoming signals for the FFT – essentially a “fade in” and “fade out” that makes the input signals appear more periodic. This windowing comes at a cost of a slightly raised measurement noise floor and some spectrum spreading, neither of which is particularly detrimental to the accuracy of our TF measurement, particularly once adequate averaging is applied.

The new SyncSource™ transfer function measurement process in Smaart v9 avoids this issue altogether by using a pseudorandom excitation source (Pink Noise or Sweep) that has a cycle length matched to the length of the FFT being used. Also referred to as a "Source Dependent Measurement”, the SyncSource™ transfer function measurement avoids the use of a data window by using a truly periodic signal. As such, synchronized measurements require the use of Smaart's built in signal generator to provide the excitation source. 

This creates a more deterministic measurement with improved measurement stability and better correlation (higher Coherence). One key reason for the higher coherence is that the reverberant energy influencing the measurement becomes functionally correlated with the reference signal and no longer decreases the coherence values.

Example of Random Source TF Measurement (top) versus SyncSource™ TF Measurement (bottom)

SyncSource™ TF measurements are available in Suite and RT editions of Smaart. To enable synced-source measurements, go to the “Advanced” tab of the Options > Preferences dialog and check the box labeled “SyncSource™ TF” in the Transfer Functions Settings area:

When a SyncSource™ measurement is enabled, a field is added in the transfer function control bar for the user to specify their Stimulus type - either Random or Synchronized. If “Synchronized” is chosen, a second measurement parameter field labeled “Length” is added for the user to set the FFT length that will be used for both FFT size and the Generator source.

 
Example of Transfer Function Control Bar with SyncSource™ disabled (left) and enabled (right)

Choosing Synchronized as the Stimulus will:

• Limit the Signal Generator choices to only Pseudo-Random noise or Sweep
• Set the FFT size / Period Length of both the transfer function measurement engine(s) and the Generator source to the length specified in the Control Bar Length setting. These controls become inaccessible in Transfer Function measurement settings and Generator Settings
• Limit the available settings for Transfer Function FFT size to 16k and larger
• Disable any Transfer Function engines configured as exempt from Global FFT size setting, and pop an error message stating the same
• Disable the transfer function measurement data window

MTW is still calculated simultaneously, using the pseudorandom noise source and FFT data windows.

SyncSource™ TF provides excellent measurement stability and improved coherence, even in the presence of excessive external noise, without the need to increase the measurement level significantly above the noise floor. However, synchronized measurements do require the use of Smaart's signal generator to be used.