In Parts One and Two we examined system EQ goals and interconnect issues. Now we will take our first measurement and interpret it. We will measure a loudspeaker and look carefully at the results.

Check your setup

1. Start Smaart and hit “On”.
2. Generate a test signal.
3. Mute the measurement microphone.
4. Observe the reference signal at the measurement mixer and pan it right. You should see signal on Smaart’s right channel. You should not see signal on the left channel. If so, you have a problem.
5. Mute the reference signal and unmute the microphone.
6. Pan the mic left and observe signal on Smaart’s left channel. Again, you should not see signal on the right channel.

If you see either signal on both channels after panning either direction you very likely you have a mono sound card, assuming the mixer is set up correctly. If this is the case you have a problem.

Let’s assume the setup is working properly and proceed. Set levels for each channel at –12 dB.

Get Your Delay Time

Recall the two signals (test and reference) must be aligned in time in order for this to work. The best way for beginners to accomplish this is to use the “Auto Small” or “Auto Large” features. The button to activate these are found just below the “Delay” window on the display. See Figure 1.

Which to use, AutoSm or AutoLg? It depends on how reverberant your environment is. Remember this: the FFT Time Constant must be larger than the decay time of the space under test. By default AutoSm yields a time constant of about 300 msec, while AutoLg is almost 3 seconds. These values can be changed, but for now leave them at the default value. If you find yourself in a gymnasium use AutoLg.

Let us use AutoSm. Push the AutoSm button and Smaart records the input signals in Windows .wav format and displays the delay time it calculated. See figure 2.

Smaart says the microphone is 2.00 feet away from the source. Here is a key point: every time you get a delay time stop and ask yourself “is this reasonable?” Don’t blindly accept the number Smaart gives you. A variety of circumstances can cause false, and sometimes incredibly wrong results. See the help file for info on troubleshooting delay locator problems.

2.00 feet is very reasonable for this measurement. The loudspeaker is a “hi-fi” loudspeaker and the measurement mic is sitting a short distance away. Also notice the polarity is reported as positive. Be aware with a multi-component loudspeaker with a passive crossover it is not unreasonable to get a negative polarity reading, even if electrical polarity if correct. If we could measure each driver individually without the crossover we could determine the polarity of each. This information is contained in the impulse response, which we will not discuss here..

Now, let us look at the transfer function. Push the “Transfer” button and see figure 3.

Notice the ragged response, with some holes centered near 700 Hz and 4.5 kHz. We can make the picture a little easier to read by using some smoothing on the transfer function, so we’ll apply 5 point smoothing. See the help file for details. See figure 4.

Coherence

We now have a readable transfer function, displaying the difference between the source signal and what is coming from the loudspeaker. One may be tempted to start the equalization process, but we’re not ready yet. This measurement may not be valid. One way to get a handle on this is to look at the coherence of the measurement. From the Smaart help file:

“Coherence is a measure of the linearity between two signals in a transfer function measurement. The Coherence function in SmaartLive basically asks "What are the chances that the signal that went into the system became the signal that came out as a result of any linear process?"

Note that poor coherence is not always an indicator of bad data. In noisy environments, or for measurements taken during a show it is probable you will see poor coherence. But in a controlled environment coherence is a valuable indicator.

For now let us ignore “coherence threshold” and “magnitude threshold”. Push the “Coh” button and a red line is displayed. Red line near the top equals good coherence, red line near the 0 dB mark on the trace equals poor coherence. Let us look for some trends in coherence in order to judge the validity of our measurement. See figure 5.