Stereophile's Test CD 2 Tracks 15-19

Track Information, Tracks 15-19, Subjective room & speaker diagnostic tracks

[15] Pink Noise at -20dBFS (DDD) 1:24
Correlated between channels, then uncorrelated from 0:42

[16] Bass Decade 1/3-octave warble tones at -20dBFS (DDD) 2:47
Center frequencies: 200Hz (Index 1), 160Hz (Index 2), 125Hz (Index 3), 100Hz (Index 4), 80Hz (Index 5), 63Hz (Index 6), 50Hz (Index 7), 40Hz (Index 8), 31.5Hz (Index 9), 25Hz (Index 10), and 20Hz (Index 11)

[17] Midrange Decade 1/3-octave warble tones at -20dBFS (DDD) 2:32
Center frequencies: 250Hz (Index 1), 315Hz (Index 2), 400Hz (Index 3), 500Hz (Index 4), 630Hz (Index 5), 800Hz (Index 6), 1kHz (Index 7), 1.25kHz (Index 8), 1.6kHz (Index 9), 2kHz (Index 10)

[18] Treble Decade 1/3-octave warble tones at -20dBFS (DDD) 2:32
Center frequencies: 2.5kHz (Index 1), 3.15kHz (Index 2), 4kHz (Index 3), 5kHz (Index 4), 6.3kHz (Index 5), 8kHz (Index 6), 10kHz (Index 7), 12.5kHz (Index 8), 16kHz (Index 9), 20kHz (Index 10)

[19] Music Articulation Test Tone (DDD) 1:19

These tracks are intended to enable audiophiles to get a handle on how their system and loudspeakers interface with their listening room, even if they don't have any test equipment.

The sound on Track 15 is random noise with equal energy per musical octave, recorded in dual mono. Called "pink" noise by engineers, it easily enables the listener to hear loudspeaker problems. It should sound like absolutely smooth rushing water with no band of frequencies sticking out any more than any other. It should also not sound hollow or colored in any way, while the image of the noise should appear to come from a narrow point midway between the speakers. If the sound of this track fails to meet any of these criteria, then try sitting higher or lower, closer or farther away, or moving the speakers and/or nearby furniture. The second noise signal is the same as the first except that each channel of noise was recorded independently. The overall sound should therefore appear to be very spacious. Listen, however, to see if any frequency bands stick out or appear to give a central image, particularly in the treble. If this happens, then your speakers probably have some resonant problems in those regions.

The warble-tone tracks (which roughly illustrate the extent of the terms "bass," "midrange," and "treble") were recorded on a JVC XD-Z1010TN DAT recorder from the output of an Old Colony Sound Lab warble-tone generator, the frequency quoted being the approximate center frequency of each. The generator contains a sinewave oscillator that is frequency-modulated at a rate of 5Hz or so; this is fast enough that the effect of low-frequency room resonances on the perceived level will be minimized, the test tone changing sufficiently quickly that the resonance doesn't have time to fully develop.

The bass warble tones can therefore be used to give a good idea of a loudspeaker's subjective bass extension in the listening room, either by listening or by using a sound-level meter—Radio Shack sells quite a good one for around $30. Set a reference level with Track 17, Index 7, the 1kHz band, then note by how much the sound level drops with each successive warble tone (footnote 7). (If your cassette deck came with a microphone, put the mike at your listening chair, set the recorder to "Record," and monitor the sound level with its meters.) The 200Hz-100Hz bands can be considered the upper bass, 80Hz-40Hz the midbass, and the remaining bands the low bass. If these bass warbles sound or measure uneven, with some either sticking out more than others or missing in action, then try moving the speakers or your listening chair around the room. The object is to get the tones as even-sounding/measuring as possible.

Tracks 17 and 18 offer sets of warble tones covering the Midrange and Treble decades (footnote 8), so that you can measure the in-room response of your loudspeakers without having to pop for an expensive spectrum analyzer. The 1kHz warble tone can also be used to get a relative idea of a loudspeaker's sensitivity: measure the sound-pressure level with a loudspeaker whose sensitivity you know, then, without changing the playback level, measure the spl of an unknown loudspeaker substituted into the system.

The Music Articulation Test Tone was supplied by Acoustic Sciences Corporation, the progenitor of the Tube and Studio Traps (see Stereophile Vol.9 No.3 and Vol.15 No.2), and again allows you to get a handle on your loudspeaker positioning and your listening-room acoustics. Essentially a "musical range intelligibility" or "Modulation Transfer Function" test, according to Acoustic Sciences' Arthur Noxon, the test signal consists of a rapid sequence of tone bursts that starts at 28Hz, rises in pitch to a peak at 780Hz, then descends back down the scale to 28Hz (see fig.5). Each burst is 2Hz above the previous one, there are eight bursts every second, and each burst last 1/16s, followed by 1/16s of silence. The frequency at any point in the test is easily determined with a stopwatch, therefore, as the tone burst changes frequency by 16Hz every second. For ascending tones, the frequency = 28 + 16t Hz, where t is the time in seconds; for descending tones, the frequency = 1532 - 16t Hz.

The best way to audition the MATT test is to first listen to the quality of the tone sequence over headphones, then over your loudspeakers. (The sound level should be set to your normal listening level.) "It is always a surprise how unintelligible audio playback systems can be," notes Arthur Noxon. "Some sections of the test will sound out articulate 'TAT, TAT, TAT,' representing clean, fast attack transients, stable sustains, and rapid decays. But interspersed among these articulate passages will be heard totally garbled 'BUDDULA, BUDDULA, BUDDULA' sounds. If you listen close to the speakers, the garbled passages will disappear, but as you back away from them into the room, the amount of garbled signal will increase rapidly."

Fig.5 ASC Music Articulation Test Tone, plotting toneburst frequency (triangular curve) and amplitude (ragged curve) against track playing time

The Radio Shack SPL meter can be used with the MATT signal. Set the speed to FAST and the weighting network to C SCALE. Adjust the speaker volume and meter gain so that the display does not peg the meter. Simply watching the meter needle is instructive: first experience a fully articulate signal by holding the meter a couple of feet from the speaker, on-axis. The needle will vibrate rapidly as it follows the rapid loud-quiet-loud-quiet of the tone-burst sequence. Now move back to the listening position and the needle will not move nearly as wildly. It will barely quiver during the inarticulate passages and mildly fluctuate during the more articulate passages.

The object of the test is to move your speakers and/or listening position and/or add acoustic treatment to your listening room to obtain the most articulate sound/SPL meter reading across the entire frequency range of the test. A 1dB swing of the needle indicates your room needs a lot of help in that frequency region; 10dB or more is good to excellent performance.

If you have any comments on the MATT test, we are sure that Acoustic Sciences—Tel: (503) 343-9727, Fax: (503) 343-9245—would be delighted to hear from you. In addition, recordings of MATT tests made in your room from the SPL meter's line output—be sure to turn off your cassette deck's Automatic Gain Control, if present—can be sent to ASC for processing into graphs for a nominal charge (phone first).

Footnote 7: Whereas the warble tones on the succeeding tracks increase in frequency, those on Track 17 decrease, to make it easier to judge bass extension by ear.

Footnote 8: The 20kHz warble tone is included for completeness's sake, but because of its proximity to the edge of the CD passband, its actual center frequency measures closer to 18kHz.