Manufacturer's Specifications:
Frequency Response: 31 Hz to 16 kHz ±0.5 dB at line input, ±3 dB at microphone input.
Filter Center-Frequency Accuracy: 31 Hz to 1 kHz, ± 10%; 2 to 16 kHz, ±5%.
Amplitude Accuracy: ± 1 dB at 12 dB range, ±2 dB at 24-dB range, and ±3 dB at 36-dB range.
Input Sensitivity: Line, 150 mV; mike, 0.5 mV.
Input Impedance: Line, 100 kilohms.
Mike, 33 kilohms.
Pink-Noise Output Level: 150 mV.
Microphone Type: Electret condenser.
Pickup Pattern: Omnidirectional.
Dimensions: 17 in. (435 mm) W x 3 3/8 in. (86 mm) H x 9 1/2 in. (240 mm) D.
Weight: 6.5 lbs. (2.9 kg).
Price: $230.00.
In the Sound Shaper SA-1, ADC provides an octave-band real time analyzer (RTA) with a built-in pink-noise source and a matched test microphone at a most-attractive price.
The front panel is black with white designations, which are easy to read in any type of lighting. The display consists of columns of horizontal-bar LEDs for each of the 10 octave bands (31 Hz to 16 kHz), plus one that shows the average of the 10 bands. Each column is 12 LEDs high for a total range of 12, 24 or 36 dB as selected with a rotary switch, corresponding to 1, 2 or 3 dB/step, respectively. There are scales at the ends of the display to aid in checking relative levels. A nice feature is the white, side-lighted graticule with horizontal lines every 2 dB which make reading the display especially easy, much better than most such units.
There is a "Calibration" pot, which is really just a level-set control for shifting the display up or down. Each bar graph of turned-on LEDs shows the level in its respective band.
There are push-button switches for "Pink Noise" on/off, "Mic/Line," "Peak Hold," "Left" and "Right Mode," "Slow/ (Normal) and "Power" on/off. The peak-hold switch acts to hold whatever is in the display, and subsequent higher levels do not change what is held. "Peak Hold" is sort of a misnomer in ways, but it's quite useful just the same. The "Mode" selector allows feeding in just left or right channels, or both together with a push in of both buttons. The microphone phone jack is just below the mike/line switch.
On the back panel are phono jacks for left and right line in and a pair for the pink-noise generator output. A slide switch allows inserting 26 dB of attenuation to prevent overloading the line inputs if connection is made to power amplifier outputs. This is another useful feature, quite unexpected in an inexpensive unit. There is also an unswitched accessory a.c. outlet. Removal of the steel top and side cover revealed a large, almost chassis-size p.c. board with a neat, orderly layout and quality components. All parts and adjust pots were clearly labeled. The pots are trimmers for 12-, 24- and 36-dB ranges and for 1-, 2and 3-dB initial steps. These are potentially useful for the owner or serviceman for maintaining long-term accuracy. Soldering was very good, with a little flux residue in spots. Interconnections were made with multi-conductor cabling (soldered) or wirewrap. The hardboard back and bottom panels of the SA-1 contribute to its light weight, albeit giving it a possible vulnerability to interference from strong fields.
The provided test microphone is powered by a type-N battery, and close to shelf life should be possible if the switch is at "Off" whenever the mike is not in use. The cable is about 16 feet long, and extension is possible.
Performance
The first series of tests were on the filters of the analyzer.
With a CW test tone, the peak responses of the 10 filters were within ±0.4 dB, very good performance. The center frequencies were within 7% for the lower bands and within 4% for the upper bands, better than most such units. The crossovers from one band to the other were down 4.7 dB which is quite acceptable. The adjacent-channel response to a tone at center frequency was about -9.8 dB on the average, adequate for the intended purpose of the unit. By the centers of the second bands, the responses were down 17.5 dB (acceptable) and the final slopes of 6 dB/octave were established.
With an vie IE-20B as the pink-noise source, the band responses were all within ± 1 dB, with the minor exception that there were some "4-2" indications occasionally in the bands below 500 Hz, particularly when in fast mode. In switching to "Slow," the lower bands shifted down slightly (less than a dB), and the upper bands shifted up a dB. All bands were still within ± 1 dB. The "Average" indication appeared to be a rough average, and it was most accurate when band levels were fairly close to each other. It did indicate relative total power correctly as-3 dB when a bandpass filter was used to drive just half (five) of the filters to reference level. (Reference level energy in half of the filters equals one-half the power for the 10 filters at reference level, or 3 dB less.) "Peak Hold" maintained whatever was in the display at the time of switching for a long period of time, whether the signal was removed or not.
Tests of the vertical display thresholds were run, using "8" on the 12-dB range as the reference. All of the thresholds were very close for all three ranges, with most within 0.3 dB, which is excellent performance and quite superior to a number of other units. Total-range errors were less than 0.9, 0.2 and 1.3 dB for the 12-, 24- and 36-dB ranges, respectively--also excellent results. The line input sensitivity (with "Calibration" at maximum) was about 23 mV for pink noise to turn on the bottom row of LEDs. For a mid display 12-dB-range "6," 45 mV was needed. An input of 90 mV got a "12" on the 24-dB range, and 180 mV produced an "18" with the 36-dB range. With a 1-kHz test signal, "6" (12 dB range) turned on at 20 mV input. Input levels up to 25 V were controllable with the level-set pot. The "-26 dB" switch on the back panel shifted the 1-kHz sensitivity (pot at maximum) to 0.5V, and that would facilitate matching the voltage levels from even high-power amplifiers.
Attention was then turned to acoustic tests with the supplied microphone. By comparison with an Ivie IE-30A, it was determined that first indications on the display occurred at about 58 dB SPL. This was not a reliable level for measurements, however, both from the standpoint of system noise and external effects. The level from the reference loudspeaker was increased to 85 dB SPL, which is a good level for such tests. It is higher than what many would prefer for classical music, and it is lower than what might be desired for rock music. The Ivie IE-30A was operated in octave band mode to allow direct comparison with the ADC analyzer results. The tests were conducted in the direct field of a single loudspeaker to eliminate two-source and room effects. The hold functions of both analyzers were used to aid in making the response comparisons. In the main, the SA-1 showed close agreement with the known, excellent performance of the Ivie. The 31-Hz band (of the SA-1) was up perhaps 3 dB relative to the other bands, and the 8- and 16 kHz bands were slightly high. The supplied microphone did show some directivity, and the flattest response in the direct field, was obtained with the microphone at close to grazing incidence (not pointed at the loudspeaker). In a reverberant field, such as might exist at a listening position, the pointing of the microphone would be less significant, but the user should avoid pointings that cause a sharp 16-kHz rise.
The line input impedance was 86 kilohms over most of the band, down to about 48 kilohms at 10 kHz. These figures are somewhat lower than the specification, but they are more than high enough for any possible application. A 1-kHz tone burst was used to check the analyzer's dynamic responses. In fast mode, a burst of 200 mS duration was needed for a 20-dB change in indication, with a decay time of about 7 seconds. In "Slow," the on-time required was 355 mS, and the decay time was about 18 seconds. The decay times are particularly long, but they are quite acceptable for the purposes of equalizing a sound system. They are limiting, however, for any sort of music monitoring, except to observe general spectral patterns on an averaging basis.
The pink-noise generator output level was 200 mV, dropping to 140 mV with a 10-kilohm load, indicative of a 4.2kilohm source impedance. The spectrum, as indicated on the analyzer display, was within ± 1 dB, with the exception that the two lowest bands were at +2 dB. Or it could be said that the 16-kHz band was the sole exception at -2 dB. Either way, the results were really quite good.
In-Use Tests
The instruction manual makes the point directly that the analyzer is designed as a complement to a graphic analyzer. The text is not detailed in most places, and some of the English would have to be classified as rather clumsy. The statements on white and pink noise are confused and confusing, but there are good points made about avoiding excessive bass boost and the desirability of some high-end roll-off. A full schematic is included, which is good; the addition of a block diagram would be helpful for the neophyte. A short time was spent doing some music monitoring, and the unit did give good average-spectrum indications. The unit's sensitivity was not quite high enough to show the lower levels with classical music from a particular FM station, but there was no such problem with pop-rock.
The advantages of the SA-1 were demonstrated much more completely with two simulated tests. In the first, a speaker-like response was generated with an MXR 1/3-octave equalizer. An MXR octave-band EQ was added to the output of the 1/3-octave EQ, feeding the SA-1 in turn. To show what was involved, the Ivie IE-30A was operated in octave-band mode and its output fed to a storage scope.
Using the ADC unit's display, the MXR octave-band EQ was adjusted to flatten the response. No reference was made to the IE-30A response until the task was considered complete. Figure 1 shows the original response with a boost at 125 Hz, a crossover dip at around 1 kHz and a peak around 4 kHz. Also shown is the result after using the SA-1 to make adjustments. The improved response is obvious.
Fig. 1--"Loudspeaker" response before (top) and after (bottom) equalization.
(Vertical scale: 10 dB/div.)
Fig. 2--"Tape recorder" response before (top) and after (middle) equalization, with actual equalization (bottom). (Vertical scale: 10 dB/ div.)
In the second test, the same approach was used, but the response of a tape recorder was simulated:-a rise at the low end, some droop around 2 to 4 kHz and a definite rise around 8 to 16 kHz. Figure 2 shows the set-in conditions, the response after adjustments per SA-1 display and the EQ that was introduced to effect the change. It usually seemed best to start adjustments in fast mode to speed up removal of the major deviations. Then, "Slow" was used with a more expanded scale for less display shifting and better resolution.
"Peak Hold" was very useful a number of times, and it had very little drift over long periods of time, except at the lowest-microphone levels. It was definitely much better than indicated by a statement in the manual to the effect that it would drift after 10 seconds. The bar graph form of the display made it particularly easy to read, but somewhat dim lighting seemed best for quick assimilation of the data presented by the red LEDs.
It would probably be helpful to most users if ADC had given some general guidelines in the manual as to what settings of the so-called calibration control would give center-display indications at various SPLs. Our tests showed that SPLs would be high enough with the pot near its center of rotation or above. A quiet environment might permit somewhat lower settings.
All in all, the SA-1 octave-band RTA is a well-performing instrument, and it will give good results with octave-band or parametric equalizers in smoothing sound system responses. With its low price, the ADC SA-1 analyzer is undoubtedly an excellent value.
-Howard A. Roberson
[adapted from AUDIO magazine/Sept. 1982]
Also see: Audiosource RTA-One Real-Time Analyzer (Dec. 1986)