dbx 165 Compressor / Limiter Instruction Manual

dbx 165 Compressor / Limiter

Dear dbx customer:
Your new model 165A is identical to the old model 165 except for a special feature called Peak top. Until a new manual is produced, this memo has all of the necessary PeakStop information and is intended to supplement the accompanying 165 manual.
The PeakStop feature allows you to control th maximum peak levels at the output of the 165A irrespective of any other control. PeakStop comes after the compression and other circuitry, including the output gain, so it lets an absolute limit be put on the peak-to-peak excursions of the output. PeakStop works instantaneously; you are able to apply moderate amounts of dbx’s Over Easy compression and will still be protected from large transients, other short-term overloads, and overmodulation. Technically, PeakStop consists of a sophisticated voltage-controlled clipper that produces a minimum of audible distortion. It rounds the corners of a peak rather than cutting it off sharply, as the word “clipping” usually implies. By making a signal’s leading and trailing edges curved instead of sharp corners, it reduces the amount of higher odd-order, offensive-sounding harmonics that conventional clipping causes. The level at which PeakStop is activated is adjustable from -2 to +24 dBm. Note that small signal excursions above the set value of PeakStop are possible, to allow the rounding to take place; therefore, for any applications where you must not exceed a given ceiling, set the PeakStop control 1-2 dB below it to be sure. The red PeakStop LED flashes whenever peaks attempt to exceed PeakStop level and get reduced in amplitude. To disable the PeakStop function altogether, simply set the control to +24 dBm (which is the maximum output level of the 165A anyway). In use, the PeakStop function can prevent an amplifier from being driven into hard clipping, where it can lose control over the speaker system. PeakStop is a smooth, well-controlled clipper whose behavior is sonically similar to the gentleness of Over Easy compression; its clipping is much preferable to a power amp’s. As mentioned, control of speaker over excursion, broadcast overmodulation, and harsh electronics clipping are all applications of PeakStop. With it and Over Easy, you have the best of both worlds: virtually inaudible rms compression and peak protection downstream at the end.

Sincerely,
dbx Inc. Professional Products

WARNING: TO PREVENT FIRE OR SHOCK HAZARDS, DO OT EXPOSE THIS APPLIANCE TO RAIN OR MOISTURE.

BRIEF OPERATING INSTRUCTIONS

•”Threshold” is defined as the point of 6 dB gain reduction when the unit is set for a maximum (infinite) compression ratio.
0 dB is referenced co 0.775 V RMS.

“dbx” and “Over Easy” are trademarks of dbx Inc.

INTRODUCTION

• The dbx Model 165 is a professional single-channel compressor/limiter that features the new dbx Over Easy compression curve. The dbx Over Easy compression curve permits extremely smooth, almost inaudible com­pression due to the gradual change of compression ratio around the threshold, instead of the customary sharp threshold. This curve, plus dbx’s true RMS level detector coupled to a wide-range voltage-controlled amplifier in a feed-forward circuit, makes it possible to achieve larger amounts of compression without adverse audible side effects.
• The RMS level detector is separately accessible, allowing various signal conditioners – equalizers, filters, delay lines, etc. – to be inserted in either the level detector or signal path independently. This feature aids in the compression of certain difficult types of material, as well as in the creation of special effects. In addition to dbx’s automatically variable attack and release time circuitry (as on Models 160, 161, 162, 163, and 164), a front panel switch and controls provide for manually adjustable attack and release rates over a very wide range.
• The Model 165 also features a stereo coup! ing for two 165s (at the touch of a front panel button), LED indicators showing whether the signal is at. above or below the threshold of compression, a 30 dB dynamic range meter (indicating input, output, and gain change levels), and a rear panel zero VU calibration adjustment. A hard-wired bypass switch, also located on the front Jjanel, is convenient for checking the effect of the compression and assures the failsafe flow of audio through the unit. The maximum input level is +24 dB (12.3 V RMS) and the maximum output level is +23 dBm. The output amplifiers have a source impedance of 47 ohms and will drive input loads of 600 ohms or greater impedance. Output gain is also adjustable over a very wide range(± 20 dB). so the Model 165 is compatible with virtually all professional sound and creative audio equipment.
• Perhaps the most flexible and useful compressor/ limiter ever offered, the dbx Model 165 is well suited to a wide range of applications including tape recording, disc mastering, radio and TV production and broadcast, live concert sound reinforcement, and theatrical production.

More About Separate Level Detector Access
Because auxiliary sound equipment can be used to process the level detector signal but not the main audio input signal (or vice-versa). the 165 offers the user an opportunity to create many unusual effects. By con­necting it to additional signal processors, such as a parametric equalizer, the Model 165 can be converted to a de-esser, a vocal stressor or a level-sensitive filter. Certain musical or vocal elements in a program can be suppressed without affecting others. In addition to these signal conditioning functions, many creative special effects are possible.

More About Over Easy Compression
Conventional compressors/limiters have a sharp knee at the threshold point (see Figure 3A). That is, when the input signal is below the threshold, the gain remains fixed ( 1: 1 ratio or no compression), but when the input signal goes above the threshold, the gain abruptly decreases according to the compression ratio for which the unit is adjusted (several fixed compression ratios are shown in Figure 3A). This abrupt change in gain in a conventional compressor/limiter is often audible and therefore undesirable to the user.
The dbx Over Easy approach utilizes a soft knee at the threshold of compression (see Figure 38). The Over Easy compressor/limiter gradually increases its compression ratio from 1: 1 towards the set compression ratio as the input signal rises through the threshold region. Thus an age-old dream can be realized … dynamic range restriction without audible, abrupt gain changes. When properly operated, the Model 165 is a highly effective compressor/limiter that you don’t hear working.

Since dbx Over Easy compressor/limiters have no distinct point at which the gain changes, the threshold on such units is defined differently from conventional units. We define the threshold to occur approximately mid-way between the fixed gain portion of the curve and the point where the curve “levels off” at the selected compression ratio. At an infinite compression setting, the threshold is defined as the level at which 6 dB gain reduction is realized. At this setting, the maximum permissible output level is 5 dB above the threshold. At lower compression settings, the threshold represents somewhat less gain reduction. ( In contrast, a conventional compressor/limiter yields O dB of gain reduction at the threshold; gain reduction begins just above the threshold.) To see how the 165’s Threshold indicator LEDs correlate with the compression curves, refer to Figure 4.

Make input and output connections to the barrier strip on the rear panel ( Figures 5 and 6).

Input Connection
For balanced or floating lines, connect the signal leads to the (+) and(-) terminals, and the shield to the chassis ground () terminal. For unbalanced lines, connect the signal high lead to the(+) terminal, and jumper the(-) and chassis ground ( ) together for connection of the shield. When using an unbalanced connection, reversing the(+) and(-) input terminals will cause the output signal to be 180° out of phase (reverse polarity) relative to the input signal.

Level Detector Input Connection
For normal compressor operation, leave the factory­installed straps connected between the Detector(+) and the Input Signal (+) terminals and the Detector(-) and Input Signal (-) terminals. ( Refer to Figure 5.) If you wish to gain access to the 165’s Detector Input for insertion of an auxiliary device, you may do so by removing the strapping, wiring the auxiliary device’s output to the 165’s Detector Input terminals, and feeding the auxiliary device’s input with the same signal fed to the 165’s Signal Input. In certain situations, the auxiliary device may need to be inserted in the signal path, not in the detector path. In such cases, the signal is fed to the input of that device, and also to the detector input, and the auxiliary device’s output is fed to the 165 signal input. Unless the installation is permanent, this connection can be awkward. Therefore, we recom­mend wiring a few Tip/Ring/Sleeve phone jacks to the 165 instead and using the jacks for Signal and Detector Inputs as well as for the Signal Output. (Refer to Figure 6.)

Fig. 6 – Signal Input & Output Jacks With A Normalling Jack For The Detector Input Dotted lines indicate the location of optional mult jack (parallel wired) for feeding input of auxiliary device with same program as 165 Signal Input.

Output Connection
The output of the dbx 165 is designed to feed balanced or unbalanced 600 ohm or greater loads. The output stage is single-ended so that in normal operation, the (-) signal output terminal is internally connected to the ( ) terminal. When the system bypass switch is engaged, the signal inputs are directly connected to the signal outputs (so a balanced input would produce a balanced output).

Grounding
For maximum hum rejection, avoid common ground­ing at the input and output (i.e., double grounding). One method that usually works is to ground the shield at the 165’s output Ground() terminal and also ground it at the input of the following device. Do not connect the shield at the 165’s input Ground terminal; leave the input shield connected only to the output of the device feeding the 165.

Stereo Coupler Cable
When you wish to link two Model 165’s for processing a stereo program, a cable must be constructed to join the Stereo Coupler connectors on the two units. Use the mating connectors supplied with each Model 165, and wire the cable according to Figure 7. Use connectors supplied with units, or equivalent (Cinch-Jones P-312-CCT). The cable should be 6-pair, twisted 24 ga. wire with shield (Belden 9506). Once the cable is connected to the two 165’s, it may be left in place at all times. When both units’ Stereo Coupler switches are placed in “master” mode ( button OUT). the 165’s operate completely independently; stereo operation is achieved simply by switching one of the two units to “slave” mode (button IN, Slave LED ON).

Input Impedance & Terminations
There is sometimes a misunderstanding regarding the nature of matching and bridging inputs, the use of terminating resistors. and the relationship between actual input impedance and nominal source impedance,. Most electronic outputs work well when “terminated” by an input (connected to an input) having the same or a higher actual impedance. Outputs are usually over· loaded when terminated by an impedance that is lower than the source impedance. When the input impedance is nearly the same impedance as the source, it is known as a “matching” input. When an input is 10 times the source impedance, or more, the input is considered to be a “bridging” input. The dbx Model 165 signal input has an actual impedance of 22,000 ohms in balanced configuration or 11,000 ohms in unbalanced (it has a high-Z” input). This makes the 165 signal input suitable for use with virtually any nominal source impedance, low or high. The dbx signal input will bridge 150-ohm or 600-ohm (low-Z) lines.

The dbx 165 detector input has an actual impedance of 600,000 ohms in balanced configuration or 300,000 ohms unbalanced. This very high impedance means the level detector will bridge virtually any nominal source impedance. It also means that when the jumper connectors link the detector and signal inputs, the detector has a negligible effect on the 165’s input impedance as “seen” by the source device. The very high impedance of the detector does suggest, however, that cables from an external processor to the detector input be kept as short as practical to reduce susceptibility to hum and RFI. Terminating resistors are not needed for the dbx 165 signal or detector inputs to operate correctly, but may be required at these inputs when they are fed from devices (such as passive equalizers) designed for a specified load impedance. The dbx 165 output is capable of driving loads of 600 ohms or greater. While it may be terminated by a low impedance, such termination is not required. *”Z” is an abbreviation for “impedance.”

OPERATION & APPLICATIONS

Additional Information on the Setting and Function of Certain Controls.

Compression Control
This control is continuously adjustable from a com­pression ratio of 1: 1 to infinite compression (that is, no change in output level regardless of changes in input level above the set threshold). High compression (settings greater than 6) significantly level out program materials. Infinite compression virtually stops music levels from exceeding the threshold setting. Lower compression settings (below 4) still permit dynamic range to exist. They are used to tighten up the sound of a bass guitar, lead guitar, snare drum, kick drum and vocals. Moderate over al I compression is typically used during stereo mixdowns. Here a variety of settings could be used. While any compression ratio can be set with this control, remember that the Over Easy curve causes that ratio to be approached gradually as the input signal level rises through the threshold. Only when the input signal is considerably above the threshold does the 165 closely approach the set ratio of compression. (See Figure 3.)

Threshold Adjustment & LED Indicators
The Threshold control sets the level at which the 165 begins to compress the signal (i.e., it sets the point where the Over Easy compression curve intersects the input signal). The control has a wide range so that the desired results can be obtained with any line level input signal. When the Threshold control is set too low, the 165 will compress most if not all of the input signal (red LED ON most of the time). At low compression ratios, the very low threshold setting can be used to gently reduce the overall dynamic range of the program. Low threshold settings (-20 to -40) are used when program materials are to be compressed. Low settings, with high input levels, cause the 165 to “work” all or more of the time than with lower input levels. The major portion of program material is processed with a low threshold setting. Normal compression and “leveling” of vocals and instruments typically use lower settings. Compression of the whole program, however, may not sound natural, especially at compression ratios of 10: 1 or greater. High threshold settings (-10 to +10) are used for limiting program levels or where only peaks are to be compressed. The major portion of the program material is not processed with a high threshold setting. Speaker protection and peak overload prevention are just two applications.

NOTE: Whenever threshold level and compression ratio settings are made, it is important to watch the LED’s and meter for reference and confirmation. Remember that a compressor/limiter is a tool that can provide desirable effects when used properly. When used to excess, the results can sound unusual, and may be of value only for special effects. With the Compression control set for the desired maximum compression ratio (or an approximation thereof), rotate the Threshold control counterclockwise until the desired sound, special effect, or amount of gain reduction is achieved.

Auto Switch & Auto Attack/Release LED Indicator
When the Auto switch is IN (Auto Mode). the LED indicator will be illuminated and the Model 165 will automatically adjust its attack rate and release time to suit the program envelope. (This Auto Model sets the Model 165 for the same attack and release characteristics as dbx Models 160, 161, 162, 163 and 164 compressor/limiters.) When the Auto switch is OUT (manual mode). the LED indicator above it turns OFF, and the front panel Attack and Release rate controls determine the maximum rate of gain change and the behavior of the level detector circuitry (see below).

Attack Rate & Release Rate Controls
The Model 165 is the first dbx compressor/limiter to offer a choice of automatic or user adjustable attack and release characteristics. In Auto Mode, the 165 utilizes the patented dbx RMS level detector with its program dependent attack/release characteristics to obtain natural-sounding compression or limiting. For special effects and certain signal situations, however, it is often desirable to set fixed attack and release characteristics. Manual mode affords this capability. The Auto Mode is recommended for vocals as well as instruments. When determining separate attack and release rate control settings, it provides a good starting point. Because the Auto Mode has a variable attack rate, the 165 may compress or limit some program materials smoother than in the manual mode which has a fixed rate of attack. This is especially true on vocals.

Where To Set Attack & Release Controls
There is no “right” way to set the Attack and Release controls. General! y, one would want a slow enough Release to avoid “pumping” or “breathing” sounds caused when background sounds are audibly modulated by the dominant signal energy, yet the release must be fast enough to avoid suppression of the desired signal after a sudden transient or a loud note has decayed. Depending on the desired effect, one might want a very slow Attack so that percussive or transient sounds are not restricted, but average volume levels are held within the desired range. A very fast attack setting (control maximum counter­clockwise) will cause the 165 to act like a peak limiter even though RMS detection circuitry is used. Slower attack settings cause the 165 to act like an RMS or averaging detecting compressor/limiter. Don’t forget, attack and release controls operate together and with the compression ratio control. Changing any one control may necessitate changing the other settings.

Meter Calibration & Use
The meter in the 165 is factory-calibrated to indicate “O” when the signal level is +4 dB ( 1. 23 V) at either the input or output of the 165, depending on the meter function switch position. (When the meter is in Gain Change mode, “O” indicates no gain change; the meter calibration control does not affect this mode.) To recalibrate the meter, engage the Input meter function switch and feed a 1 kHz signal at the selected nominal operating level (the level desired for a “O VU” meter indication) to the 165’s signal input. Then adjust the rear panel METER CALIBRATION control until the meter indicates “O dB”.

Use of an Equalizer in the Level Detector Circuit for Frequency-Weighted Compression, Dressing, or Increasing Sustain

• It is possible to “separate” certain instruments from a mix by frequency-weighted compression. This function is created by inserting an equalizer ahead of the Model 165’s level detector input, but not in the main signal input path. Peaking the equalizer in a certain frequency range will tend to suppress any frequencies (notes) in that register. A relatively high threshold setting can allow normal sounds to be unaffected while solo or very loud sounds are held back. Of course, when compression does occur, the level of the entire program is affected. For this reason, it may be more useful to reserve this combined EO/compression technique for isolated sound sources, such as a single channel of vocal or a single instrument in a multi-track program. Unlike overall program equalization, EO of the level detector will take effect only when signals are above tthe hreshold (or when those frequencies affected fall above the set threshold). Depending on the threshold setting, lower-level funda­mentals or harmonics will not cause compression. and the program is not subject to the phase shift normally caused by program equalization.
• For example, consider a single channel carrying the preamplified signal from a microphone placed near a cymbal and a torn tom. Set up the 165 with an equalizer in the level detector path, as depicted in Figure 8. The equalizer can be adjusted for boost with a peak at about 5 kHz, causing the cymbal to be compressed on a very loud crash, preventing tape saturation at high frequencies where there is less headroom. However, gentle tapping of a drumstick or brushing of the cymbal will not be held back. Assuming the torn-tom is a lower frequency instrument and can be better tolerated by the tape, there is less need for compression on it. The equalization in the detector circuit means that the compressor will not be triggered as readily by a loud torn beat as by an equally loud cymbal crash.
• Another application for this type of EQ boost in the level detector is for increasing the sustain of a guitar, bass, etc; this requires EO boost in the dominant frequency range of the instrument, along with a fairly low threshold and a moderate compression ratio.
• Still another related application involves de-essing of vocals (reduction of sibilance). Use a parametric equalizer in the level detector circuit and set it for a high-frequency boost in the specific frequency range where the vocal “hiss” or lisp occurs. This pre-emphasizes the already “hissy” vocal input to the detector. Used in conjunction with a moderate to high threshold and compression ratio, th is arrangement greatly attenuates the “missing” without affecting the basic sound quality or balance of the voice. While it is true that all frequencies are lowered in level when the compressor is triggered, generally the “sss” sound occurs alone, before or after the dominant tone in the voice. ( If this seems unlikely, just try to hum and hiss at the same time.)
• The converse of the above EQ techniques may be used; dipping the equalizer will cause any sound in the affected register to pull the level up because it will seem to require less compression than the other frequencies.
  

Use of a Filter in the Level Detector Circuit
The results of inserting a filter in the level detector circuit are basically the same as obtained with an equali· zer, as previously described. Those frequencies passed by the filter are subject to compression (or at least they are subject to considerably more compression than those frequencies outside the passband). Because a passive filter can have insertion loss, it may be necessary to lower the 165’s Threshold setting to maintain a given amount of gain reduction within the tilter passband; this can be determined, as usual, by monitoring the 165’s threshold indicator LEDs.

Use of a Time Delay Line in the Signal Path but not in the Level Detector Circuit for Zero or Negative Attack & Release Times
While the Model 165 can be set for incredibly fast attack times, there will always be some small transient that “gets past” the level detector. In some cases – such as maximum modulation broadcasting – it may be desirable to preserve the sonic quality obtained with a slower attack time. yet it may not be permissible for even the slightest overshoot to get past the compressor/ limiter. A delay line (digital or analog) can be used in this instance. By feeding the program directly to the 165’s detector input, but delaying the feed to the 165’s signal input, the unit can “anticipate” the need for a gain change. (Refer to Figure 9.) With some experimen­tation, the effect can be that of “zero” attack time.

Additional signal delays beyond the “zero” time established above would then cause the compressor to finish changing gain before the leading edge of the loud passage enters the signal input, suppressing the program which is not above the threshold. Also, the 165 would begin to recover from compression (release) before the input signal has dropped back to the set threshold, causing the output to surge higher in level as the note or passage is decaying. This special effect obtained with the time delay might sound akin to reverse playback of a tape recording.

Access to the 165’s level detector makes possible a whole range of effects not normally available. The more you think about it and experiment, the more use­ful this capability can become.

The 165 As A Line Amplifier
To use the 165 as a line amplifier, adjust the COMPRESSION RATIO control to fully counterclock­wise (1:1 position). THRESHOLD to full clockwise position (+10) and OUTPUT GAIN to whatever setting is required for the application. Remember that, as wtth any amplifier, excessive gain may lead to output clipping of high-level signals. To add compression, adjust the COMPRESSION RATIO and the THRESHOLD controls to the desired settings.

SPECIFICATIONS

dbx PRODUCT WARRANTY & FACTORY SERVICE

All dbx products are covered by a limited warranty. Consult your warranty card or your local dealer for full details.
The dbx Customer Service Department is prepared to give additional assistance in the use of this product. All questions regarding interfacing dbx equipment with your system, service information or information on special applications will be answered. You may call during normal business hours – Telephone: 617-964-3210. Telex: 92-2522, or write to:

dbx, Inc .
71 Chapel Street
Newton, MA 02195
Attn: Customer Service Department

Should it become necessary to have your equipment factory serviced:

1. Please repack the unit, including a note describing the problem along with the day, month and year of purchase.
2. Send the unit freight prepaid to:
   dbx, Inc.
   224 Calvary Street Waltham, MA 02154 Attn: Repair Department
3. We recommend that you insure the package and send it via United Parcel Service wherever possible.
4. Please direct all inquiries to dbx Customer Service Department.
   Outside the U.S.A. – contact your nearest dbx dealer for the name and address of the nearest authorized repair center.

BLOCK DIAGRAM

SCHEMATIC

GLOSSARY

• Asperity Noise
  This is a swishing type of background noise that occurs with tape recordings in the presence of strong low frequency signals, especially when there are no high frequency signals to mask the hiss. Asperity noise is caused by minute imperfections in the surface of the tape, including variations in the magnetic particle size in the tape’s oxide coating. The imperfections increase or decrease the strength of the magnetic field passing the play head in a random manner, resulting in audible noise. Asperity noise may be present even when no program is recorded. When a program is recorded, asperity noise becomes superimposed on the signal, creating modulated asperity noise, or “modulation noise.” Using high-quality tape with a calendered surface helps reduce asperity and modulation noise (calendered tape is pressed smooth by high-pressure rollers).
• Attack Time
  Attack time may mean different things, depending on the context. In music, the time it takes for a note to reach its full volume is the attack time of the note. Percussive instruments have short attack times (reach maximum volume quickly) and wind instruments have lon’g attack times (reach maximum volume more gradually). When a compressor (or expander) changes the level of an incom­ing signal, the circuitry actually requires a finite amount of time to complete that change. This time is known as the attack time. More precisely, the attack time is the interval (usually measured in milli­seconds or microseconds) during which the compressing or expand­ing amplifier changes its gain from the initial value to 90% of the final value.
• Aux Input (Aux Level)
  Aux inputs, an abbreviation for auxiliary inputs, are low-sensitivity jacks provided on most hi-fi and semi-professional equipment. Aux inputs (also known as “aux level” or “line level” inputs) have “flat” frequency respanse and are intended to be used with preamplified signals. Aux-level (line-level) signals are medium­level, higher than microphone levels, but not enough power to drive a speaker. The advantage to these levels is that they are less susceptible to hum and noise than are microphone levels. Typical items that might be connected to aux inputs are tape machine “play” outputs, tuner outputs, and dbx “play” outputs. Mic-level or phono-level signals are considerably lower 1n level than aux inputs (approx. -60 to -40dBV), so they will not produce adequate volume when connected to an aux input. Moreover, phono cartridge outputs require RIAA equalization which 1s not provided by aux inputs.
• Bandwidth
  Bandwidth refers to the “space” between two spec1f1c frequencies which are upper and lower limits; alternately, band­width refers to the absolute value of the range of frequencies between those llmlts. Thus, a filter that passes frequencies from 1 ,OOOHz to 1 O,OOOHz may be said to have a bandwidth of 1 kHz- 10kHz, or it may be said to have a 9kHz bandwidth (10kHz minus 1 kHz equals 9kHz).
  Bandwidth is not necessarily the same as frequency response. Bandwidth may be measured at low levels, and frequency response at higher levels. Moreover, bandwidth may refer only to certain portions of the circuitry within a piece of equipment, whereas frequency response may refer to the overall performance of the equipment. Thus, while the overall input-to-output frequency response of dbx type 11 equipment is 20Hz to 20kHz, the band­width of the RMS detection circuitry within that equipment 1s 30Hz to 10kHz.
• Bass
  The low audio frequency range below approximately 500Hz. For the purpose of discussion or analysis, the bass range may be further divided into upper bass (250 to 500Hz), mid bass ( 100-200Hz), low bass (50-1 OOHz). and ultra-low bass (20-50Hz).
• Bass Boost
  An accentuation of the lower audio frequencies (bass frequencies), whereby they are made louder than other frequencies.
• Biamplified
  Descriptive of a sound system that utilizes a low-level cross­over network to divide the full-spectrum audio signal into low and high-frequency ranges. These ranges are then fed to separate power amplifiers, which in turn feed low-frequency speakers (woofers) and high-frequency speakers (tweeters).
• Bias
  Bias, as the term is used in tape recording, is a very high-frequency signal (usually over 1 OOkHz) that is mixed with the program being recorded in order to achieve linear magnetization of the tape. If only the audio program were applied to the recording head, a very distorted recording would result because lower-energy portions of the program would not be able to overcome the initial magnetization threshold of the tape (known as hysteresis). The frequency of the bias signal is not critical, so long as the record and erase bias are synchronized. However, the bias energy level has a direct effect on the recorded level, background noise, and distortion. It is sometimes necessary to reset the bias level for optimum performance with different types of recording tape, and professional tape machines are equipped with continuously variable bias controls; many consumer tape machines are now equipped with bias selector switches.
• Clipping
  Clipping is a very distorted sound. It occurs when the output capabilities of an amplifier are exceeded, and the amp can no longer produce any more voltage, regardless of how much additional gain or how much more input signal is present. Clipping is relatively easy to see on an oscilliscope, and it is sometimes audible as an increase in harmonic distortion. In severe cases of clipping (hard clipping), sine- waves begin to resemble square waves, and the sound quality is very poor. Often, the maximum output level of an amplifier is defined as 3s the level where clipping begins to occur. There is a phenomenon known as input clipping, and this may occur where the input signal is so high in level that it exceeds the level-handling ability of the transformer and/or of the input amplifier. Clipping also occurs when the ape is saturated by excessive record levels. So-called “soft clipping” is usually the result of transformer saturation, and it may be somewhat less objectionable than the “hard clipping” that occurs when output voltage limits are reached. Aside from degrading the sound quality, clipping can damage loudspeakers. Output clipping may be avoided by reducing the level of the input signal , reducing the gain of the amplifier, or using a larger amplifier, Input clipping may be avoided by reducing the level of the incoming signal, and then increasing the gain of the amplifier.
• Clipping Level
  This is the signal level at which clipping just begins to occur. The clipping level is not always easy to define. It may be a matter of visually judging the waveform on an oscilliscope as the level is increased; alternately, a clipping level may be defined as the level at which harmonic distortion reaches a given value. Tape clipping, or saturation, is defined as the 3% harmonic distortion level.
• Compression
  Compression is a process whereby the dynamic range of program material is reduced. In other words, the difference between the lowest and highest audio levels is “squeezed” into a smaller dynamic range. A compressed signal has a higher average level, and therefore may have more apparent loudness than an uncompressed signal, even though the peaks are no higher in level. Compression is achieved with a compressor, a special type of amplifier that decreases its gain as the level of the input signal increases. The amount of compression is expressed as a ratio of the input dynamic range to the output dynamic range, thus, a compressor that takes a program input with 100dB of dynamic range and yields an output program of 50dB dynamic range may be said to have a 2:1 compression ratio.
• Compressor
  A compressor is an amplifier that decreases its gain as the level of the input signal increases to reduce the dynamic range of the program (see “compression”). A compressor may operate over the entire range of input levels, or it may operate only on signals above and/or below a given level (the threshold level).
• Crossover Frequency
  In loudspeaker systems and multi-amplifier audio systems, the transition frequency (actually a frequency range) between bass and midrange or midrange and treble speakers or amplifiers.
• Crossover Network
  A circuit that divides the audio spectrum into two or more frequency bands for distribution to different speakers (high-level crossover) or different amplifiers which then feed different speakers (low-level crossover). High-level crossovers are usually built into the speaker cabinet and are passive (they require no power supply). Low-level cross-overs are used in amplified or tri-amplified sound systems. They are usually self-contained, and come before the power amplifiers Low-level crossovers may be passive or active; active low-level crossovers are known as “electronic crossovers. logarithm of a power ratio or 20 times the logarithm of a voltage or sound pressure ratio.) If the number of “dB’s” are referenced to a given level, then the value of the dB number becomes specific.
  • dbv expresses a voltage ratio. OdBv is usually referenced to 0.775V rms. Thus, OdBv=0.775V, +6dBv=1.55V (twice OdBv), +20dBv=7.75V (ten times OdBv), etc.
  • dBV expresses a voltage ratio and is similar to DVB, but OdBV is usually referenced to 1 V rms. Thus, OdBV is 2.2dB higher than OdBv.
  • dB SPL expresses a Sound Pressure Level ratio. dB SPL is a measure of acoustic pressure (loudness), not acoustic power, which would be measured in acoustic watts. OdB SPL is equal to 0.0002 dynes/square centimeter (the threshold of human hearing at 1 kHz). As with dBV, an increase of 6dB SPL is twice the sound pressure, and an increase of 20dB SPL is an increase of 10 times the sound pressure.
  • dBm expresses a power ratio. OdBm is 1 milliwatt (.001 watts), or 0.775V rms delivered to a 600-ohm load. +3 dBm=2 milliwatts, or 1.55V into 600 ohms (twice OdBm), +10dBm=10 milliwatts, or 7.75V into 600 ohms (ten times OdBm), etc. dBV and dBm are numerically equal when dealing with 600-ohm circuits. However, when the impedance is other than 600 ohms, the value of dBV remains the same if the voltage is the same, whereas the value of dBm decreases with increasing impedance.
  • dB alone, without any suffix, doesn’t mean anything unless it is associated with a reference. It may express the difference between two levels. Thus, the difference between 10dBV and 15dBV, the difference between DdBm and 5dBm, and the difference between 90dB SPL and 95dB SPL are all differences of 5dB.
• Decay Time
  Decay time may mean different things, depending on the con­text. A compressor’s decay time is also known as its release time or recovery time. After a compressor (or expander) changes its ga,n to accommodate an ,ncoming signal, and the signal ,s then removed, the decay time is the amount of time required for the circuitry 10 return 10 “normal.” More precisely, the decay 1, me is the interval (usually measured 1n microseconds) during which the compressing or expanding amplifier returns to 90% of the normal gain. Very fast decay times can cause “pumping” or “breathing” effects, whereas very slow decay times may cause a moderate-level program that follows a high-level program or program peaks to be too low in level.
• Decoder
  When a circuit restores an original program from a specially treated version of that program, the circuit may be said to decode the program. The equipment or circuit wh,ch performs this function is known as a decoder. Decoders must be used only with programs that have been encoded by complementary encoding circuitry. Typical decoders include: FM tuners that use multiplex decoders to extract left and right stereo signals from left-plus-right and left-minus-right signals, matrix quadraphonic decoders that extract four channels of program from the stereo program on encoded recordings, and dbx decoders that retrieve wide-dynamic range programs from the compressed programs on dbx-encoded recordings.
• D&-emphasis & Pre-emphasis
  De-emphasis and pre-emphasis are related processes that are usually done to avoid audio noise in some storage or transmission medium. Pre-emphasis is a boost at specific higher frequencies, the encoding part of an encoding/decoding system. De-emphasis ,s an attenuation at the same frequencies, a reciprocal decod1n11 that counteracts the pre-emphasis. In dbx noise reduction, de-emphasis is performed by the decoder (the play circuitry). The de-emphasis attenuates high frequencies, thereby reducing tape modulation noise and restoring the original frequency response of the program before it was dbx encoded. There are other types of pre-emphasis and de-emphasis. For example, in FM tuners, de-emphasis is used to compensate for special equalization (known as 75-microsecond pre-emphasis) applied at the s1a11on’s transmitter.
• Dynamic Range
  The dynamic range of a program is the range of signal levels from the lowest to the highest level. In equipment, the dynamic range is the “space,” in dB, between thr residual noise level and the maximum undistorted signal level. A program with wide dynamic range has a large variation from the softest to the loudest passages, and will tend 10 be more lifelike than programs with narrow dynamic range.
• Encoder
  When a circuit processes an original program to create a specially treated version of that program, the circuit may be said to encode the program. The equipment or circuit which performs this function is known as an encoder. Encoded programs must decoded only with complementary decoding circuitry. Typical encoded programs include: FM multiplex broadcasts, matrix quadraphonic recordings, and dbx encoded recordings.
• Envelope
  In music, the envelope of a note describes the change in average signal level from initial attack, to peak level, to decay time, to sustain, to release time. In other words, the envelope describes the level of the note as a function of time. Envelope does not refer to frequency.
  
  ln fact, any audio signal may be said to have an envelope. While all audio frequencies rise and fall in instantaneous level from 40 to 40,000 times per second, an envelope may take many milliseconds, seconds or even minutes to rise and fall. In dbx processing, the envelope is what “cues” the rms level detection circuitry to com· press and expand the signal; the peak or average level of individual cycles of a note would be useless for level detection because the gain would change much too rapidly for audibly pleasing sound reproduction.
• EQ(Equalization)
  EQ or equalization, is an intentional change in the frequency response of a circuit. EQ may be used for boosting (increasing) or cutting (decreasing) the relative level of a portion of the audible spectrum. Some EQ is used for achieving sound to suit personal listening tastes, while other types of EQ are specifically designed to correct for non-linearities in the system; these corrective EQ “curves” include tape (NAB or CCIA) equalization and phono­graph (RIAA) equalization. In a sense, the pre-emphasis and de­emphasis used in dbx processing are special forms of equalization. There are two common types of EOualizat1on curves (characteristics): PEAKING and SHELVING. Shelving EQ is used in most Hi-Fi bass and treble tone controls. Peaking EQ is used in Hi-Fi midrange tone controls, in graphic equalizers. and many types of professional sound mixing equipment. EQ is performed by an equalizer, which may be a specially built piece of equipment, or it may be no more than the tone control section of an amplifier. Graphic equalizers have many controls, each affecting one octave, one-half octave, or one-third octave of the audio spectrum. (An octave is the interval between a given tone and its repetition eight tones above or below on the musical scale: a note that is an octave higher than another note is twice the frequency of the first note.
• Expander
  An expander is an amplifier that increases Its gain as the level of the input signal increases, a characteristic that “stretches” the dynamic range of the program (see “expansion”). An expander may operate over the entire range of input levels, or it may operate only on signals above and/or below a given level (the threshold level).
• Expansion
  Expansion is a process whereby the dynamic range of program material is increased. In other words, the difference between the lowest and highest audio levels is “stretched” into a wider dynamic range. Expansion is sometimes used to restore dynamic range that has been lost through compression or limiting done in the original recording or broadcast; expansion is an integral part of compander-type noise reduction systems, including dbx. Expansion Is achieved with an expander. a special type of amplifier that increases its gain as the level of the input signal increases. The amount of expansion is expressed as a ratio of the input dynamic range to the output dynamic range; thus, an expander that takes a program input with 50dB of dynamic range and yields an output program of 1 0OdB dynamic range may be said to have a 1:2 compression ratio.
• Fundamental
  A musical note is usually comprised of a basic frequency, plus one or more whole-number multiples of that frequency. The basic frequency is known as the fundamental, and the multiples are known as harmonics or overtones. A pure tone would consist of only the fundamentals.
• Ground Compensated Output
  This is a sophisticated output circuit that senses the! the potential difference between the ground of the dbx unit and the shield ground of unbalanced inputs to which the dbx unit is connected. Ideally, the dbx unit and the input of the following device should be at the same level (potential). However, where grounding is not “right” (where so-called “ground loops” exist), this circuit calculates the ground error and adds a correction signal to the high side of the output, thereby canceling much of the hum, buzz and noise that might otherwise have been introduced by ground loops.
• Harmonic Distortion
  Harmonic distortion consists of signal components appearing at the output of an amplifier or other circuit that were not present in the input signal. and that are whole-number multiples (harmonics) of the input signal. For example, an amplifier given a pure sine wave input at 1 00Hz may produce 200Hz. 300Hz. 400Hz, 500Hz, 600Hz and even 700Hz energy, plus 100Hz, at its output (these being the 2nd, 3rd. 4th, 5th, 6th and 7th order harmonics). Usually, only the first few harmonics are significant. and even-order harmonics (i.e. 2nd and 4th) are less objectionable than odd-order harmonics (i.e. 3rd and 5th); higher harmonics may be negligible in comparison to the fundamental (100Hzl output. Therefore, rather than specifying the level ot each harmonic com• ponent, this distortion is usually expressed as T.H.D. or Total Harmonic Distortion. While T.H .D. ,s the total power of all harmonics generated by the c,rcu1try, expressed as a percentage
  of the total output power, the “mixture” of different harmonics may vary 1n different equipment with the same T.H. D. rating.
• Harmonics
  Overtones which are integral multiples of the fundamental.
• Headroom
  Headroom refers to the “space,” usually expressed in dB. between the nominal operating signal level and the maximum signal level. The input headroom of a circuit that is meant to accept nominal – 10dB levels, but can accept up to +18dB without overdrive or excessive distortion, is 28dB (from -10 to +18 equals 28dB). Similarly, the output headroom of a circuit that is meant to supply nominal +4dBm drive levels, but that can produce +24dBm before clipping is 20dB. A circuit that lacks adequate headroom is more likely to distort by clipping transient peaks since these peaks can be 10 to 20dB above nominal operating signal levies.
• I.M. (Intermodulation Distortion)
  Intermodulation distortion consists of signal components appearing at the output of an amplifier or other circuit that were not present in the input signal, that are not harmonically related to the input, and that are the result of interaction between two or more input frequencies. I.M. distortion, like harmonic distortion, is usually rated as a percentage of the total output power of the device. While some types of harmonic distortion are musical, and not particularly objectionable, most I.M. distortion is unpleasant to the ear.
• Impulse Response
  Related to the rise time of a circuit, the impulse response is a measurement of the ability of a circuit to respond to sharp sounds. such as percussion instruments or plucked strings. A circuit with good impluse response would tend to have good transient response.
• Level Match
  The dox noise reduction system is unlike competitive systems in that there is no one threshold at which compression or expansion begins. Instead, compression occurs linearly, with respect to decibels, over the full dynamic range of the program. By necessity, there is an arbitrary signal level that passes through the encoder and decoder without being changed in level. This level is known as the level match point (transition point). Some dbx equipment provides for user adjustment of the level match point, for monitor. ing purposes only. Although this is not necessary for proper encode/ decode performance, by setting the level match point to be approximately equal to the nominal (average) signal level, there will be no increase or decrease in level as you switch from monitoring “live” program to monitoring dbx-processed program.
• Limiter
  A limiter is a type of compressor, one with a 10:1 or greater compression ratio. A limiter with a high compression ratio (120:1) can be set so that no amount of increase in the input signal will be able to raise the output level beyond a preset value. The difference between limiting and compression is that compression gently “shrinks” dynamic range, whereas limiting is a way to place a fixed “ceiling” on the maximum level, without changing the dynamic range of the program below that “ceiling,” or threshold.
• Line Level (Line Input)
  Line level refers to a preamplified audio signal, in contrast to mic level, which describes a lower-level audio signal. The actual signal levels vary. Generally, mic level is nominal -50dBm (with ta ypical dynamic range of -64dBm to +10dBm). Line level signals vary, depending on the audio system. Hi-Fi line levels are nominally Line inputs are simply inputs that have sensitivities intended for line level (preamplified) signals. Often, the nominal impedance of a line-level input will be different than the nominal impedance of a mic-level input.
• Modulation Noise
  Modulation noise is a swishing type of background hiss that occurs with tape recordings in the presence of strong low-frequency signals. The noise depends on the level of the recorded signal; the higher the recorded signal level, the higher the modulation noise. Modulation noise has typically been “masked,” hidden by the dominant signal and/or by the background hiss of the tape. However, when the background hiss is removed, as with dbx processing, modulation noise could become audible. This would happen primarily with strong, low-frequency signals, but in fact it is minimized by dbx’s pre-emphasis and de-emphasis.
• Octave
  In music or audio, an interval between two frequencies having a ratio of 2:1.
• Overshoot
  When a compressor or expander changes its gain in response to a fast increase or decrease in level, the maximum gain change should be directly proportional to the actual signal level. However, in some compressors the level detection and gain changing circuitry develop a kind of “inertia,” over-reacting to changes in level, increasing or decreasing the gain more than the fixed ratio asked for. This over-reaction is known as overshoot, and it can cause audibly non-linear compression (distortion). dbx circuits have minimal overshoot, so they provide highly linear compression and expansion.
• Peak Level
  An audio signal continuously varies in level (strength, or maximum voltage) over any period of time, but at any instant, the level may be higher or lower than the average. The maximum instantaneous value reached by a signal is its peak level (see RMS level).
• Phase Shift
  “Timeshift” is another way to describe phase shift. Some circuitry, such as record electronics and heads, will delay some frequencies of an audio program with respect to other portions of the same program. In other words, phase shift increases or decreases the delay time as the frequency increases. On an absolute basis, phase shift cannot be heard, but when two signals are compared to one another, one having a phase shift relative to the other, the effects can be very noticeable, and not very desirable. Excessive phase shift can give a tunnel-like quality to the sound. Phase shift also can degrade the performance of compander-type noise reduction systems which depend on peak or average-level detection circuitry.
• Power Amplifier
  A unit that takes a medium-level signal (e.g., from a pre-amplifier) and amplifies it so it can drive a loudspeaker. Power amplifiers can operate into very low impedance loads (4-16 ohms), whereas preamplifiers operate only into low impedance (600 ohms) or high impedance (5,000 ohms or higher) loads. Also known as a main amplifier, the power amplifier may be built into an integrated amplifier or a receiver.
• Preamplifier
  A device that takes a small signal (e.g., from a microphone, record player), or a medium-level signal (e.g., from a tuner or tape recorder), and amplifies it or routes it so it can drive a power amplifier. Most preamplifiers incorporate tone and volume con-trols. A preamp may be a separate component, or part of an integrated amplifier or of a receiver.
  Pre-Emphasis (See “de-emphasis”)
• Receiver
  A single unit that combines tuner, preamp and power amplifier sections.
• Release Time or Release Rate
  (See “decay time” and “attack time”)
• Rise Time (Attack Time)
  This is the ability of a circuit to follow for “track”) a sudden increase in signal level. The shorter the rise time, the better the frequency response. Rise time is usually specified as the interval (in microseconds) required to respond to the leading edge of a square-wave input.
• RMS Level
  RMS level (Root Mean Square) is a measurement obtained by mathematically squaring all the instantaneous voltages along the waveform, adding the squared values together, and taking the square root of that number. For simple sine waves, the RMS value is approximately 0.707 times the peak value, but fer complex audio signals, the RMS value is more difficult to calculate. The RMS level is similar to an average level, although not identical (The average level is a slower measurement).
• Sub Harmonic
  A sub-multiple of the fundamental frequency. For example, a wave the frequency of which is half the fundamental frequency of another wave is called the second subharmonic of that wave.
• Sub Woofer
  A loudspeaker made specifically to reproduce the lowest of audio frequencies, usually between 20Hz and 1 OOHz.

  Synthesizer
  An ELECTRONIC MUSIC SYNTHESIZER is an audio processor that has a built-in sound generator (oscillator). and that alters the envelope of the sound with voltage-controlled circuitry. Synthesizers can produce familiar sounds and serve as musical instruments, or they can create many unique sounds and effects of their own. A SUB HARMONIC SYNTHESIZER is a device that is not used to create music but to enhance an existing audio program. In the case of the dbx Model 100, the unit creates a new signal that corresponds to the volume of the input signal but is at 1 /2 the frequency of the input signal.

• Tape Saturation
  There is a ‘maximum amount of energy that can be recorded on any given type of magnetic tape. When a recorder “tries” to record more energy. the signals become distorted but are not recorded at an,y higher levels. This phenomenon is called tape saturation because the magnetic oxide particles of the tape are literally saturated with energy and cannot accept any more magnetization. T.H.D. (Total Harmonic Distortion) (See “Harmonic Distortion”)
• Threshold
  Threshold is the level at which a compressor N limiter ceases to have linear gain, and begins to perform its gain-changing function (i.e., where the output level no longer rises and falls in direct proportion to the input level). In most systems, the threshold is a point above which the level changes, although there are compressors that raise signal levels below a threshold point. Some compander­type noise reduction systems, such as Dolby have upper and lower thresholds between which the gain changes; these systems require careful level calibration for proper encode/decode perfor­mance. dbx noise reduction systems have no threshold at which compression or expansion factors change, so level calibration is not critical.
  
• Tracking Accuracy
  Tracking refers to the ability of one circuit to “follow” the changes of another circuit. When two volume controls are adjusted in exactly the same way, the corresponding “sameness” of the output levels can be expressed as the tracking accuracy of the controls.
  The level detection circuitry in a dbx encoder senses the signal level, changes the gain, and creates an encoded signal. The corre­sponding “sameness” of the original signal and the encoded/ decoded signal can be expressed as the tracking accuracy of the noise reduction system. (dbx systems are non-critical for the operator and are built to close tolerances, so that tracking accuracy is excellent, even if the encoder and decoder are in different pieces of dbx equipment.
• Transition Level (See Level Match)
  When a circuit has uniform compression or expansion through­out its full dynamic range, there must be some level that passes through the unit without being raised or lowered (where gain is unity). This unity gain level is the transition level or transition point. The transition point is a “window” 1dB wide, in a dbx encoder (compressor!, all signals above the transition point are decreased in level, and all signals below the point are increased in level. Con­versely, in a dbx decoder (expander), all signals above the transition point are increased in level, and all signals below the point are decreased in level. The transition level is similar to a “threshold,” except it does not refer to a point at which compression or expansion factors change.
• Trlamplified
  Similar to blamplified. A sound system where a passive cross­over network creates three frequency ranges, and feeds three power amplifiers: one for bass, one for mid, end one for high frequencies. The amplifiers are connected directly to the woofers, midrange drivers and tweeters without a passive, high-level crossover network.
• Tuner
  A unit that receives radio broadcasts and converts them into audio frequency signals. May be part of a receiver.
• VCA (Voltage Controlled Amplifier)
  Traditionally, amplifiers have been designed to increase signal levels (to provide gain). If an amplifier were required to decrease the level (to attenuate), it could become unstable, and might even oscillate. The gain (amount of amplification) in these traditional amplifiers would be adjusted by one of three methods (1 I attenuating the audio signal fed to the Input of the amplifier, (21 attenuating the audio output of the amplifier, or (31 changing the negative feed­back (feeding more or less signal from the output back to the input, but in reversed polarity). The VCA is a special type of amplifier that can be used to increase or decrease levels over a wide dynamic range. Instead of using signal attenuation or negative feedback, the gain (or loss) is adjusted by means of an external de-control voltage, dbx has a unique, patented VCA design that has extremely low noise and very wide dynamic range; the dbx VCA is the heart of dbx noise reduction equipment.
• Woofer
  A loudspeaker that reproduces only low frequencies.

Manufactured under one or more of the following U.S. patents: 3,681,618; 3,714,462; 3,789,143; 4,101,849; 4,097,767.
Other patents pending.