There have been several discussions lately in the ChurchSoundcheck Discussion Group on the topic of measuring sound systems and acoustics. It's been an interesting topic to consider. One issue that has been hard to convey through simple text is the idea of resolution. Hopefully the graphics in this article will help you visualize what people have been talking about.

Resolution determines the ability to see fine details in the measurement. Consider this analogy. If you look at a leaf on a tree several feet away from you, you can appreciate its beauty but you really can't see any detail. You can walk up close to that tree, hold that leaf in your hand and look at it very closely and of course start to see the detail of its shape. Put that same leaf under a microscope and you'll see significantly more detail. Which is better? Each view provides you with information. Much like spiritual gifts, the best detail or resolution to have is the one you need at the moment.

In the case of measuring loudspeaker systems, a "microscope" can often come in handy. That's when we turn to sophistocated analyzers like the TEF, Smaart, MLSSA, and others. But there are also times when a Real-Time Analyzer (RTA) is the right tool to use. We can discuss which measurement tool is the better one to use for a given situation at another time. For now, we simply want to shed some light on this idea of resolution.

All of the measurements presented in this article were prepared with the TEF electro-acoustic analyzer. Its software allows the data to be gathered and presented in a myriad of ways by using TDS, RTA, and FFT measurement techniques. The following graphs are fairly self-explanatory.

The first two graphs compare the resolution of a TDS (Time Delay Spectrometry) measurement with that of a 1/3 octave RTA. The first thing you notice is that the detail in the TDS display is significantly greater than that of the RTA display. In general, the resolution of the TDS measurement is much higher than that of the RTA. There are just thirty data points displayed with the RTA, compared with 2048 in this particular TDS display.

Here's an explanation of the test. The first measurements were taken of a single three-inch loudspeaker (a Radio Shack Minimus 0.3). The frequency response is pretty much what one would expect of an inexpensive little three-inch loudspeaker - not much happening in the low frequencies and not exactly very smooth in the mids or highs. But sound comes out and it does its job. So the first test was simply to compare the resolution of a typical 1/3 octave display with that of the TDS measurement.

The next set of curves shows the interaction of a pair of those RS Minimus 0.3 loudspeakers, purposely misaligned by exactly one inch. One would expect that misalignment to cause phase cancellations due to the constructive and destructive interaction of the sound waves from the two loudspeakers as they arrive at the test mic position (about two feet away). The deep notches and bumps in the resulting frequency response curves clearly show that interaction. Note the difference in resolution between the TDS curve and the RTA display.

The remaining RTA displays show the impact of increasing the resolution of the measurement display from 1 octave to as high as 1/12 octave. Not all RTA's are capable of displaying 1/12 or even 1/6 octave resolution like the TEF can. Take a moment to peruse these differences.

Finally, here's the ultimate loudspeaker. Just check out that response! Wanna buy one!?! Anyone would until they read the fine print. Check the bottom right corner of the display. See that number labeled Smoothing? That feature is there to take the small 'wiggles' out of the curve which no one could hear anyway, and therefore present a more meaningful curve.

Those tiny little wiggles in the response, sometimes refered to as 'grass', are typically the result of reflections from nearby surfaces getting into the direct sound measurement. The TDS measurements I've shown here have been smoothed to 1/6 octave. (RTA measurements do not have this smoothing feature because the resolution isn't high enough to matter.) Some research has shown that 1/3 octave is the smallest resolution that the average listener can hear. Smoothing the display to 1/6 octave allows me to present a more readable graph without hiding any problems in the response.

So would you believe that this is the same exact loudspeaker that I've used in all of the displays here? In fact it is. I've just manipulated the data in this display by smoothing it 1000%, which just goes to show that you can't always believe what you see. Or measure.