[Support] [Product Info] [Downloads] [Ordering] [Home <www.channld.com/software.html>]

Loudspeaker Frequency Response Measurement

The goals include: Finding the best room position for the speaker system (or subwoofer: adjusting subwoofer parameters - crossover frequency, level) balance Equalization of frequency response; especially high efficiency or other loudspeakers which may not exhibit balanced frequency response In the latter case, equalization for overall response is a reasonable option in moderate size rooms (up to about 15 feet in largest dimension, or somewhat larger rooms with low reverberation). Here, sound reflected from room boundaries will be dominated by the Haas effect. The Haas effect is a psychoacoustic phenomenon whereby sounds from multiple sources are perceived by the listener as coming from a single source, provided that the sound arrivals coincide within 30 milliseconds. Furthermore, the apparent sound source is the source with the first arrival time. Thus, equalization for overall sonic balance succeeds, because the listener cannot distinguish direct and reflected sounds - only their sum, which is perceived as originating at the loudspeaker. Moderate corrections can be made with an equalizer. Properly applied, equalization significantly improves the overall sonic balance of a sound system, minimizing the effect of engineering trade-offs made in the design of the sound system components. Particularly with background music (at any sound level) or non-critical listening situations, equalization artifacts (added noise, distortion or phase nonlinearity, depending on equalizer design) are offset by the general improvement in sonic characteristics. Different equalization settings for each channel (see Case 2) may cause an audio imaging shift, if applied in a critical listening environment. Here, it's best to compromise with equalization settings. For example, use a "common denominator" approach - settings that correspond to the channel with the smallest extremes. Large differences between channels may require other remedies for correction: keep loudspeakers away from reflective surfaces; experiment with loudspeaker placement or sound dampening materials in the listening room. Case 1. Measurement of high fidelity loudspeaker system Case 2. Measurement and equalization of minispeaker sound system The instrumentation used: Macintosh PowerBook 5300ce (117 MHz) / 32 MB RAM / System Software 7.6.1 Mac the Scope 2.6 Audix TR-40 Test and Recording Microphone Event Electronics EMP-1 Microphone Preamplifier

Figure 1. Diagram of the measurement setup. The equalizer was used for Case 2.

Measurement procedure: The electronics were connected as shown in Figure 1. The stereo preamplifier was set to minimum gain. Off-the-shelf cables and adaptors from an electronics shop may be used for making electrical connections. However, if handy with a soldering iron, it's best to build a custom cable, which will be easier to set up, be more reliable and neater in appearance. The parts needed: Several feet of 2 - conductor shielded microphone cable (2 - conductors plus shield) Two 1/8" stereo mini-plugs One 1/4" mono phone plug (this is one of the output connectors of the Event preamp) One solder-type phono (RCA or cinch) plug. RCA male to female shielded cable, of appropriate length (to connect the custom cable to the Stereo Preamplifier). Balanced type shielded 2-conductor cable was used, for convenience. Since the cable run was short (less than 1 foot) and impedances were relatively low, crosstalk was negligible. For longer cable runs, when using both conductors of the cable, different cable should be used. The cable connects Channel 1 Sound Out with the Stereo Preamplifier, Channel 1 Sound Out with Channel 1 Sound In, and Channel 2 Sound In with the microphone preamp. Channel 1 corresponds to the Tip connector on the stereo mini-plug. Wiring: 1. Cut 6" of cable. Solder a mini-plug to one end of cable. 2. Cut 3" of cable. Connect shield and one conductor of the cable to the shield and Channel 1 conductor of the cable from Step 1. 3. Solder an RCA plug to the Channel 1 conductor and shield of the cable from Step 2. 4. Cut 6" of cable. Solder a mini-plug to one end of the cable. 5. Connect shield and one conductor of the cable from Step 4 to the shield and Channel 1 conductor of the cable from Step 1. 6. Connect a length of cable (corresponding to the expected distance between the computer and the microphone preamplifier) to the shield and Channel 2 conductor of the cable from Step 4. 7. Connect a 1/4" phone plug to the shield and conductor at other end of the cable from Step 6. After wiring the cable, check all signal paths with an ohmmeter, looking especially for short circuts between channels, before connecting the cable. Mark the mini-plugs to indicate signal in or signal out - use colored electrical tape or heatshrink tubing. Place corresponding pieces of colored tape (or dabs of colored paint) near the computer sound jacks.

Mac the Scope Setup Mac the Scope was launched. (File Sharing was previously disabled, and the display was set to 256 colors. Consult Performance Tips and Mac the Scope FAQs for more information.) The Waavebox Console was activated. The Disclosure Triangle was used to reveal the bottom panel of the Waavebox Console. DIGITAL LOCK was activated. (With DIGITAL LOCK activated, the demands on the processor are minimized). The LATENCY control panel was activated and LATENCY set to maximum. The Waavebox loudspeaker icon was clicked, and the output level set to 50% of maximum in the OUTPUT LEVEL control panel. The Waavebox Noise icon was clicked, and PINK selected. Waavebox output was activated (ON button illuminated). The Mac the Scope VU Meter was activated and Channel 1 checked to verify the presence of an input signal. The Waavebox OUTPUT LEVEL and Monitors and Sound Control Panel were adjusted to obtain an input level of between -20 and -10 dB. The VU Meter window was closed, and the Mac the Scope Console bargraph meters deactivated by clicking on the color bar icon in the corner of the bezel area. Mac the Scope was set for continuous signal averaging (Console CONT button clicked). Horizontal grids were activated using the Console grid button (this also activates the grids in the RTA display). The Mac the Scope Console display was collapsed (reduced to just the title bar) by clicking in the collapse box. RTA Setup RTA Setup... (RTA menu) was selected and the setup dialog configured as shown: (Note that the Microphone Sensitivity corresponds to the microphone manufacturer's reported value, adjusted by 20 dB preamplifier gain. However, this parameter only is important when measuring sound pressure levels; further, the preamplifier gain is changed to an arbritary setting below.) The Mac the Scope RTA Channel 2 to Channel 1 Ratio display was activated by dismissing the RTA Setup... dialog. The stereo preamplifier level was increased until a moderately loud output level was achieved (ca. loud conversation, about 80 - 85 dB SPL). The signal level balance of Channel 1 and Channel 2 was adjusted by changing the gain of the microphone preamplifier. After each setting, the RTA Channel 2 to Channel 1 Ratio display Reset button was briefly clicked to reset the averaging counter and display the new results. The level setting procedure was repeated until the RTA curve was approximately centered at the 0 dB level. As a further measurement refinement, Mac the Scope was set to Don't Acquire Data in Background (Input menu) and the ADFB application launched, to determine the interchannel signal delay. After waiting several seconds for the delay value to be calculated, the delay value was noted and ADFB exited (Quit). The delay (absolute) value was entered in the Mac the Scope Input Signal Delay control panel (applied to Channel 1). The RTA Channel 2 to Channel 1 Ratio display Reset button was briefly clicked again. Signal averaging was allowed to proceed until the RTA level bars stopped moving appreciably, and the measurement result was saved.

Case 1. Measurement of high fidelity loudspeaker system

Mac the Scope was used to measure the frequency response of a loudspeaker system (Vandersteen Model 2 ce). The purpose of the measurement was to determine the loudspeaker performance in the listening room. The loudspeaker midrange and tweeter balance controls were set to minimum. The microphone was placed on axis, respective to the normal listening position, approximately 4 feet from the loudspeaker. Figure 2 shows a reduced (76%) screen grab of the PowerBook display, after the measurement averaging was complete.

Figure 2. Frequency response measurement of reference loudspeaker (screen grab from PowerBook, reduced to 76%).

The computer input and output were connected to determine the signal chain flatness. This ensured that the measured response was due to loudspeaker / room / microphone rather than the measurement system. The PowerBook straight-wire measurement is shown in Figure 3. The manufacturer supplied microphone frequency response curve is shown in Figure 4.

Figure 3. "Straight-Wire" combined frequency response of signal source (PowerBook Sound Out) and measurement (PowerBook Sound In).

Figure 4. TR-40 microphone, factory response curve (from supplied identification plate).

Case 2. Measurement and equalization of sound system

A pair of JBL Pro III miniature loudspeakers (2-way, ported box design) were chosen for their small size and relatively high power handling, for a background music and non-critical listening application. The initial measurement results from the left channel are shown in Figure 5. The sonic balance showed coloration in the midbass and midrange, which confirmed previous listening tests. Part of this coloration likely was due to wall mounting, which was required by the installation. An inexpensive, 2/3 octave (15 bands per channel) stereo equalizer (which was already on hand) was used to compensate the response curve, by cutting the midbass and midrange, and applying a small boost to the treble and lower midrange. Frequencies falling below louspeaker cutoff were not equalized. A parametric equalizer will provide the best results, adjusting the Q of the equalizer bands to match broad features in the loudspeaker frequency response curve, and applying small amounts of boost or cut, where appropriate. This is better than using a narrow band (ca. 1/3 octave) equalizer to try to correct narrow dips and peaks in response. The Reset button was clicked after each equalizer adjustment, to update the display. The adjustment was repeated until a satisfactory result (given the filter bandwidth limitations of the equalizer) was achieved. This procedure was repeated for the right channel of the sound system. The results are shown in Figure 6.

Figure 5. Baseline loudspeaker response.	Figure 6. Loudspeaker response after equalization.

Finally, a subwoofer was connected to extend the low frequency response. Minor adjustments were made to equalization, subwoofer crossover frequency and subwoofer level (results shown in Figure 7).

Figure 7. Overall system response, with subwoofer, after equalization.

Summary Purpose: Improving the sonic balance of a medium efficiency, miniature loudspeaker / subwoofer system, with electronic equalization. Result: The measurement system setup and equalization were performed in less than an hour, and the audio system's performance was dramatically improved. Unequalized: +/- 4.5 dB, 80 Hz - 20 kHz Equalized: +/- 2.5 dB, 80 Hz - 20 kHz Equalized, with subwoofer: +/- 2.5 dB, 40 Hz - 20 kHz

Copyright 1999 Channel D Corp. All Rights Reserved

[Support] [Product Info] [Downloads] [Ordering] [Home <www.channld.com/software.html>]