Editor's Note: Steve submitted this as his regular "On the Digital Edge" column for the January issue of FOH but, as it is a pretty extensive look at the nuts and bolts of a much-used software program, we are running it as a Road Test instead. Same stuff, different header.
Getting Smaart
The latest revision of Rational Acoustics' Smaart is v7.1 which runs under Mac OSX (10.5 or 10.6) as well as Windows 7, XP, or Vista. Rational Acoustics recommends at least a 2 GHz dual-core processor and Smaart is compatible with CoreAudio, WAV or ASIO audio drivers. I ran Smaart on a MacBook 2 GHz Core 2 Duo/4 GB RAM with Digi 002R and MOTU Traveler interfaces. If you've never used Smaart, it's worthwhile reviewing the resources available from the Rational Acoustics web site, particularly the PowerPoint presentation and basic setup guide. Reading these documents while following along with your system can cut down the learning curve, and since there's a lot to learn about Smaart, this idea is… well… smart.
The Interface
Rational Acoustics incorporated many improvements to the interface of Smaart v7.1, including a "Capture All" command that stores all active measurement traces and an improved trace filing system. A major change is that this version supports simultaneous measurement of multiple channels. Other changes will be discussed below.
Smaart's primary modes (Real Time and Impulse Response) include time and frequency domain measurements, but first you need to set up your hardware in the audio dialogue. When Smaart recognizes your interface, it appears on a menu of available I/Os. You can then set sample rate, bit-depth (16 or 24) and apply your own names to the inputs and outputs (nice for managing multi-channel systems). Smaart played very well with my Digi 002R at sample rates of 44.1-, 48- and 96 kHz, though it does not support 88.2 kHz (which I don't see as a problem). I cannot say the same for my MOTU Traveler, which Smaart did not like very much. Sometimes Smaart would recognize the Traveler, and other times – typically after changing the sample rate – it would not.
Averaging and Weighting
Inputs are organized and added into Groups under the Group Manager, where you'll find parameters including averaging and weighting. Version 7.1 is the first to support multiple channels, enabling simultaneous measurement of, for example, console output, a mic at FOH, another mic in the balcony, etc. Active channels are viewed "overlay" style; clicking on an input in the Control Strip brings its trace to the front of a window. It'd be nice if you could tile the screen into separate windows for each trace (e.g. four windows of RTA, each displaying a channel).
One of the few gripes I have with Smaart is that weighting is neither displayed nor accessible from the Control Strip, though it is indicated in the trace area. Figure 1 shows the Smaart Spectrum (RTA) function. Note that the Control Strip on the right (detailed in figure 2) displays the averaging but not the weighting, which I consider essential [Editor's note: the numeric readout at the top of the Control Strip shows weighting for the dB meter, not the analysis tool]. The Spectrum display can show RTA, Spectrograph or both via split screen (figure 3). All of the screens look great and are easy to read, but one thing I did not like is that when you zoom in or out, the scale of the screen changes, but the resolution of the grid does not (i.e., you can zoom in as far as you want, but the grid is still divided into 6 dB steps). To select an area of a window for zoom, right-click and drag on it or, on a one-button mouse, hold <Control> + <Option> and click and drag.
Other Functions
At the top of the Control Strip is a numeric indicator that shows dBFS, dB SPL (Smaart provides calibration for SPL) or dBLEQ. dBLEQ is capable of long-term SPL monitoring over a user-defined period (we went as far as six hours), with user-defined increments. A "logging" feature creates a text file of these measurements showing minimum and maximum SPL as well as the actual SPL at a given date and time. It's a very useful tool, especially in venues where the neighbors make noise complaints.
Smaart's Transfer Function allows comparison between a reference signal and the post-process version of that signal in an audio system, measurements which reveal interesting traits. An example is shown in figure 4, the Transfer Function of a monitor system in a small control room. This was measured by generating pink noise (from Smaart's signal generator), splitting it and sending it directly into Smaart on the Reference channel and also to the monitors. A measurement mic picked was connected to the Measurement channel. This comparison involves a delay between the two signals (the Reference signal does not travel through the air, and therefore reaches Smaart faster). To maintain accuracy, there must be compensation for the delay. Smaart has an automatic delay finder that calculates delay, even while you are moving the mic around the room. It worked perfectly. In Fig. 4, the lower two windows show Transfer Function. The middle trace (green) shows frequency versus phase difference while the bottom window shows magnitude (green) versus frequency difference between reference and measurement mic. The bottom also shows "Coherence" in red (a discussion of coherence is beyond the scope of this article, but it points toward reliability of accumulated data). In this particular instance we are in Live IR Mode, so Smaart added a third window at the top of the screen with amplitude versus time difference between the two channels. The green spike at approximately 5.6 mS and the smaller bump roughly one mS later provide insight as to why the green trace in the Phase window looks so erratic (phase problems and comb filtering), and also why the red trace is not as coherent as we might like.
Impulse Response
Impulse Response measurement is always tricky, because if measurement parameters are not set carefully, data acquisition is inaccurate. The setting for Time Constant (TC) must be long enough to include the entire decay time of the system under test. When capturing the impulse response of a room with a decay time of 1.5 seconds, a TC of 682 mS yields inaccurate data, because Smaart is not given enough time to "hear" the entire decay. The solution is increased TC, which also increases the FFT (Fast Fourier Transform). Processing time increases with increased TC, so one benefit of using shorter TC is that you see the results faster. Smaart provides TC settings ranging from 2 and 10922 mS, with corresponding FFT sizes and averaging to improve reliability.
There are a couple of minor things that I'd like to see tightened up in Smaart. In addition to the aforementioned difficulty with the Traveler interface, there is an issue with using the MacBook's built-in I/O. Smaart recognizes the I/O, but at times, attempting to use it prompts a message stating "Failed to Start Device." The folks at Rational Acoustics are working on solutions to those issues.
Nit-picks aside, there's no doubt that Smaart is an extremely powerful software tool, in particular for touring sound companies and installers. It takes a bit of time to become familiar with all of the capabilities but should be considered essential for anyone requiring critical evaluation of audio system performance. It's also an excellent teaching tool, providing a means of illustrating a variety of acoustic phenomena. A new license for Smaart runs $895; upgrades from earlier versions range from $450 to $650.
