September 2001

The Audio Toolbox™ Plus
Review by Dick Olsher
Click here to e-mail reviewer

  If you've ever entertained the idea of purchasing an SPL meter, a Real Time Analyzer (RTA), signal generator, or a distortion meter, but were intimidated by the cost or implied technicalities of such an investment, then take note. Your prayers have been answered by TerraSonde's Audio Toolbox™ (ATB). In one affordable package you get all of the above plus a whole lot more. Through the power of digital signal processing, the ATB combines the functionality of multiple test and measurement tools in one easy to use hand-held device. A built-in rechargeable battery system allows you complete freedom of movement. In addition, a PC and Mac computer interface is included for printing. This is not a consumer toy! Forget about that Radio Shack SPL meter you may have lying around the house, the ATB conforms to Type 1 precision instrumentation standards. It is packaged in an aluminum and steel chassis and features an external mic, a full complement (¼", XLR, and RCA) of inputs and outputs, BNC microphone connector and external cable, a 3.5" x 2.5" easy to read LCD display, 40 non-volatile memories, and an internal speaker for signal monitoring. The Plus version of the ATB, reviewed here, also comes fully loaded with the "Contractor's Software" suite of applications and tools.

There are four main menus: Acoustic Analysis, System Tools, Test Functions, and Utilities. Menus are navigated by turning the encoder knob on the front panel. Individual submenus or selections are made by clicking the knob. My strategy during this review was to focus on those features and capabilities that would be of the greatest interest to serious audiophiles and audio Do-It-Yourselfers. This seemed to be the only sensible approach, as a complete discussion would take us far away from the hobbyist arena.

Sound Level Meter (SLM)

A SLM should be standard issue for any serious audiophile. The ability to measure the loudness of an ambient sound level in units of standardized sound pressure level (SPL) is essential in optimizing the setup of a listening room. The SLM function is found under System Tools. Scroll through the top of the SLM display to select desired settings. Clicking on the first field (upper left) allows you to choose one of five available ANSI-standard averaging modes: Slow, Fast, Impulse, Peak, and LEQ or time-average sound level. I find the Impulse setting (35 ms exponential decay time) most useful when measuring the SPL of music, while the Slow setting is quite handy when mapping out sound distribution in a room using a test signal.

The next field gives 12 choices of filtering modes from flat (un-weighted), to A, B, or C weighted, to any of nine ANSI Class 1 filter octave bands - 31, 63, 125, 250, 500, 1000, 2000, 4000, and 8000 Hz. The A filter approximates a typical speech spectrum and is rolled off below 1000 Hz. The C filter is the flattest of the three, however, I would recommend using the B filter setting when matching the loudness of different loudspeaker channels. Research has shown that the B-weighted response gives the most reliable results when matching loudspeaker loudness.

For most measurements you'll need a test signal. No problem. You can jump directly to the Signal Generator menu by clicking on the SLM's Gen field. Signals available are sine wave, square wave, white and pink noise, including band-limited pink noise. For most SLM measurements I would recommend pink noise. For one thing it's relaxing, approximating the sound of a water fall. It also simulates an acoustic musical spectrum, decreasing in power with increasing frequency. I'm not suggesting that you listen to pink noise for fun, but even after only a moderate exposure period, it becomes easy, for example, to pick out loudspeaker colorations in the mid and high frequency ranges because they significantly sharpen the timbre of pink noise.

Finally, be sure to select "Mic" in the microphone field if you are using the mic provided by TerraSonde. For reproducible measurements, it is best to stand-mount the mic. A ¼-inch mic stand clip adapter and a 6-foot cable are provided for this purpose. Disconnect the mic from the ATB and clamp it onto the mic stand. Be sure to point the business end of the mic toward the sound source. Turn off you amplifier and connect a sufficiently long unbalanced cable (RCA jack terminations) from the unbalanced output on the ATB to both the left and right line inputs on your preamp or integrated amplifier using a Y-adapter cable. Turn down the volume and turn on the amp and the Generator output on the ATB. Adjust the volume and/or generator output for about 80 dB - C weighted.

And now its time for measurements… Let's say that you're attempting to optimize your listening room setup for a particular speaker placement. Measure the SPL over the listening area in a grid of about 1-square foot. Record the reading for each grid element. This gives you a good picture of any major dips and peaks in the room and allows you to select the most uniform soundfield location for your listening seat.

Another common application would be to balance two speaker channels. Place the mic at the listening seat, pointing at the center spot between the speakers, and measure the SPL in dB - B weighted, first from the left channel, and then from the right channel. Be sure to turn off the power amp while you switch the signal generator input from one channel to the other. Adjust the balance control or speaker placement until the two channels agree within a fraction of a dB. This will ensure the most stable stereo sweet spot.

Real Time Analyzer (RTA)

Real-Time Analyzers have become the workhorse tool in the application of equalization (EQ) to listening rooms, studios, and public address systems. The basic idea is to play a test signal whose level is the same at each frequency (i.e., pink noise) and then measure the in-room response of the loudspeaker. SPL deviations from flat are displayed in real time.

The ATB's RTA analyzes the audio signal at the microphone in 1, 1/3, 1/6, or 1/12 octaves and displays the spectrum as a bar graph that shows the dB levels of each frequency band. This provides quick and extremely useful in-room spectral information for loudspeakers and drivers. Two resolution modes are provided: the 'Full" mode covers the frequency range from 20 Hz to 20 kHz, while the "Low" mode coverage is 10 Hz to 332 Hz. Various averaging speeds may be selected. The longer the time constant selected, the smoother (and more accurate) the display. Constants in the 1 the 3 seconds worked best for me. A resolution setting of 1/3 octave most closely approximates the critical bandwidth of the filters in our ears, and gives the most meaningful perceptual picture of speaker response.

The biggest problem in the application of RTA results, in my experience, is the user. As long as the mic is in the speaker's nearfield, the impact of the room on the measurements is minimal. However, should you wish to EQ your sound system at the listening seat, which may two or more meters away from the drivers, the effect of the room becomes substantial. It's conceptually useful to think of the soundfield at the listening seat as a room-loudspeaker system. Reflected energy and room resonances contribute over 50% of the sound energy in a typical domestic listening environment. Now, here is a surprising fact: the equal loudness contours are not the same for direct and reflected sound. Relative to direct sound, reflected sound is perceived to be brighter. In other words, due to diffraction and at the head and external ear, the auditory system is more sensitive to diffuse treble energy. Therefore, the worst mistake you can make is to EQ your sound system to measure flat at the listening seat.

Many years ago I watched an Audio Research engineer EQ J. Gordon Holt's system to measure flat at the ratty sofa that served as his listening seat. The result was an unbearably bright sound, that reinforced Gordon's long-held belief of "Down with Flat." What he meant, and I've heard others echo the same notion, is that to sound natural, a speaker's treble response should roll off at some predetermined rate per octave. Of course, this is all wrong and stems from a misunderstanding of the psychoacoustics of reflected sound. It's perfectly all right to design loudspeakers that measure flat to 20 kHz under anechoic conditions. However, in the far field, especially in a large room, where diffuse sound dominates the soundfield, the reflected sound must not be equalized to flat. Flat in this case means too bright. For example, the ear is about +4 dB more sensitive to diffuse sound in the bandwidth from 6 kHz to 10 kHz. This has been known for many years, and has led to the development of a "house curve" for EQ purposes.

Such curves typically feature a high-frequency rolloff because they represent a combination of the speaker's direct sound and diffuse sound. An good example is the X curve: the measured SPL frequency response presented to the ears of listeners in a dubbing stage or motion picture theater. The curve is to be measured under specific conditions, and is to be adjusted for room volume as specified in the standard. Tomlinson Holman of THX fame, writing in Surround Professional (January 2000, Volume 3, Number 1) does a nice job of describing the technical history of the X curve.

Basically, the curve is flat to 2 kHz, then down 1 dB per one-third octave, to -6 dB at 8 kHz, and falling more steeply to 16 kHz. In smaller rooms, a curve which is flat to about 6 kHz may yield a more balanced response. Holman points out that the biggest problem in the application of any house curve has to do with the accuracy of the measurements. Lab standard 1/2-inch microphones demonstrate very different high-frequency response when used anechoically on axis and with a diffuse field. Most such microphones are calibrated in a free field, and would therefore seriously underestimate the contribution of diffuse treble sound. Differences of 6 dB in the top octave are possible between the two. He suggests the use of small, low-diffraction microphones, such as 1/4-inch or smaller diaphragm mics, to minimize measurement differences between direct and diffuse sound. In practice, it is a good idea to collect data with the mic pointed toward the ceiling and then compare the results with those obtained with the mic pointed at the speaker. For EQ above 6 kHz, use the data set giving the higher SPL reading in the treble.

The ATB's RTA function is easy to use. Plug the output into the amp or preamp. Turn on the pink noise generator, adjust the output level for a decent dynamic range and you're in business. I used it to measure the 1-meter on-axis frequency response of the Lowther BassZilla loudspeaker, and I was surprised by how faithfully it captured the essential features of the frequency response I had previously obtained with the CLIO MLS measurement system.

Frequency & Impedance Response Sweeps

A traditional sine wave sweep is also facilitated by the ATB using 1/3 or 1/12 octave resolution. This test offers enhanced accuracy relative to the RTA, but with the tradeoff of a much longer time interval to complete. The impedance sweep will plot the impedance of a load versus frequency. A special test cable is required for this function, which connects the test probes to the tip and ring of a ¼" stereo plug.

Energy-Time Graph

Suppose you wanted to see how room reflections contribute to the soundfield at the listening seat and how successful acoustic treatment is in mitigating specific wall reflections. Then this function is just for you. It shows the decay pattern of sound at the microphone after a pulse has been generated and played through a speaker. Initial time delay is also shown. The plot gives SPL vertically, and units of time (or distance) horizontally. A-weighting of the response is suggested for the most reliable results. This test truly allows you to see how coherent (or should I say incoherent) your multi-way speaker system really is and which early reflections are most troublesome.

Distortion Meter

The distortion meter uses notch filters to compare the level of the incoming filter with its harmonics and noise (THD + noise). The meter generates a sine wave at a user selectable frequency of either 63, 125, 250, 500, 1000, or 2000 Hz. It then sums the energy in the harmonics to compute the total harmonic distortion. The meter is capable of spanning the range of 0.02% to 50% THD. Care should be taken not to overdrive the speaker or the measurement mic.

Now you can perform the sort of measurement audio magazines rarely pay any attention to. The emphasis is always on frequency response, with anal-retentive audiophiles agonizing over a -1 dB loss at 20 kHz. You may be surprised to learn that audiophile-grade speakers, darlings of the mainstream press, routinely dish out several percent of THD in the midband at normal listening levels. And it only gets worse in the bass range. It makes you wonder what all the fuss is over low-distortion solid-state power amps (e.g., 0.001%) when the associated speaker features about an order of magnitude higher THD. Gee, I now feel better about using a single-ended, directly heated, triode power amps with over 1% THD to drive my speakers.

Polarity Tester

Under System Tools I found the Polarity Tester to be extremely useful in determining the absolute polarity of speakers, drivers, and audio chains. The polarity of any driver in a multi-way system can be checked. This test also works for microphones, amps, or preamps in an audio chain.

For best accuracy ensure that only one speaker is playing or position the mic much closer to the speaker or driver under test. Plug the ATB output into the preamp or integrated amp and turn down the volume. Slowly turn up the volume until the meter shows a mid-scale reading. The display shows either a plus sign for positive polarity or a minus sign for negative polarity. It's a simple as that, and accurate.

Conclusion

The Audio Tool Box Plus is literally stuffed with a multitude of useful and easy to use test functions. It is well engineered, accurate, easy to use , very portable for field applications, and fairly priced. It practically obsoletes many of the test gadgets I've accumulated over the years. I highly recommend it for the hobbyist or audiophile looking to perform some serious measurements. 

Specifications

Acoustic Analysis Sound Level Meter 
+/-0.3dB steady state, resolution to 0.1 dB. 35 -125 dBA with supplied microphone. Type 1 ANSI A/ B/ C weighting filters. Class 2 octave band filters (if present). 

Real Time Analyzer +/-1.0 dB with supplied microphone. Filters +/-0.20 dBu at band center frequency. 

Reverb Decay Time +/-2% of full signal, 10ms resolution. RT60 time is extrapolated from available room dynamic range. 
Energy Time Curve +/-one graph pixel. Delay calculation +/-50usec. 125usec minimum graph resolution. 

Polarity 0 to 500 feet, speaker to Audio Toolbox. Shows polarity of closest driver's mid-frequency wavefront.

Test Functions Signal Generator 
Distortion <- 76db (0.015%) typical @6 dBu, sine wave output 11-22kHz, output level range -75dBu to +14dBu. +/-0.05dBu @1000Hz, 
+/-0.10dBu 16Hz-20kHz. Pink noise +/ 0.1db, 20Hz-20kHz, >6 hour repeat cycle. Impedance meter range 1 Ohm to 8000 Ohms, +/-2% 
(or 0.2ohms) from 1 Ohm to 50 Ohms, +/-10% overall. 

Level/ Frequency Meter
+/-0.05dBu @1kHz, +/-0.10dBu, 16Hz-16kHz, +0.10/-0.80dBu, 16-20Khz. Input range from -98 dBu to +40dBu. Frequency counter range 16Hz to 40kHz, +/-1Hz. 
Signal-to-Noise Ratio > 135dB range checking. -95dBu noise floor. 
Sweeps +/-0.2dB or pixel of graph, 20Hz to 20Khz. 
Sample Scope Runs at full sample rate, 48kHz. +/-one pixel of display. 20-22kHz. 156uV to 3.0V/ div vertical scale range. 0.1ms to 8.0ms/ div screen sweep range. Auto-triggers. 
Distortion Meter 0.020% to 50%, 63, 125, 250, 1k, and 2kHz frequencies. 

Session Helpers 
Guitar Tuner +/-2 cents. 
Tempo Computer All functions, +/-0.2 BPM. Time to <1ms. 
MIDI Functions One input, two outputs, standard 5-pin DIN MIDI sockets. Tests 
MIDI cables at 2x MIDI data rate. 
Time Code Computer +/-1 frame, all functions. Supports 24 fps (film), 25 fps (EBU), 30 fps (SMPTE) and 30 fps drop (SMPTE drop). 29.97 read as 30. Sub-frame lock indicator +/-1 sub-frame. Frame per second 
reader, +/-0.001 fps. 
Hum Cancellation 45 dB reduction in line-frequency related hum components.

Utility Monitor Amp 
Up to 60dB of gain. Minimum headphone load impedance 32 ohms. 
Cable Tester Tests XLR, ¼:" balanced and single-ended, all XLR to ¼" combinations. RCA cables also tested.

Power Tests Phantom checker shows voltage +/-0.2V on + and -XLR pins. +50VDC absolute maximum Battery tester accurate to +/-.02VDC. 
Transmit Data Transmit Toolbox memories using serial protocol, uses MIDI to serial adapter cable ATB-PSC for PC and ATB-PSC & ATB-MSC for Mac. 
Electrical Specifications: Input Impedances/AC Power 
XLR and ¼" balanced inputs 40k ohms 6v-7.5vDC @450ma, unregulated, 1V ripple max @load. RCA unbalanced 20k ohms

Output Impedances Batteries XLR and ¼" balanced outputs 300 ohms 6 AA cells provide 3-5 hours of service, depending on usage. 
Internal lead-acid battery system, 3 hours with LCD backlight off. RCA unbalanced 150 ohms

ATB-1 $999 
ATB-Plus $1699 
ATB-R1 $1785

Company Information

TerraSonde 
1751 Redwood Ave 
Boulder CO 80304

Voice: (303) 545-5848
Fax: (303) 545-6066 
Website: www.terrasonde.com