Page 26 - Fall 2007

group of speakers, but not the spectrum contour. Zones have several functions, one of which is noted in the previous para- graph. They also permit adjustments for differences in the acoustical environment of the speakers, such as plenum depths, suspended or structural ceiling heights, when the additional expense of a separate channel is not warranted. Zone controls also give the user some personal control over the system and the controls should be stepped so changes do not detune the system. A rule of thumb, when higher degrees of privacy have been achieved, is that a change of 3 dB equates to a nine point change in the Privacy Index. (See the steep part of the curve in Fig. 2.) This change is enough to alter the degree of privacy; thus level steps should be on the order of 1 to 2 dB. Several manufacturers now have zone con- trols in the form of separate autotransformers, amplifier input controls, or centralized digital controls that meet this requirement.
Initial ramp function
A common adage says “Everyone is for progress; it is change they hate.” There is considerable change when employ- ees enter a new facility. To greatly reduce the impact of sound masking on these individuals, several modern masking sys- tems include an initial ramp function that will increase the lev- els automatically from background to final over many days. This function is in series with other gain controls.
Power ramp function
The above concept also applies when power to the mask- ing system is lost. To avoid the abrupt onset of sound mask- ing, a few systems now include a function that will raise the level slowly over several minutes.
Programmed level control
Most masking specifications ignore the time factor even though the need for privacy is a function of time. Employees need to communicate at times, to be private at other times, but always to retain a sense of community with their fellow employees. During night, or early morning hours, occupan-
Fig. 3. A sample of the two important percentile levels found in a busy open office with 44 dB(A) sound masking is shown. Significant potential distractions occur during work hours despite the presence of sound masking.
Fig. 4. The duration of exposure of an employee to acoustical distractions as a per- centage of the workday for several masking levels is strongly influenced by the masking level. It is based on the criterion that L10-L90 is greater than 5 dB.
cy is very low. During those hours, employees need to be aware of the presence of others, and security guards need wide aural horizons. In the early work hours, social aspects are most important and then transition to a need for privacy occurs and finally, near the end of the workday, the social aspects again prevail. One example of this effect is that com- plaints about excessive masking in fixed level systems often occur in early morning or late afternoon and this strongly suggests the need for temporal variations of masking level.
The earliest patent for this function was issued in 19788 and now a number of manufacturers have what may be called a programmed level control function that will automatically change overall masking levels with time. The installer can pro- gram the generator for a range of levels in hourly increments over any day of the week, no increment of change ever being greater than 1 dB. Typically, there is one level rise and one level decrease during the day. The rise starts about one hour before normal working hours and the decrease starts about one hour before quitting time. The length of the rise and fall time is gen- erally two hours. The total amount of change is generally 10 dB during the workday, i.e., the rise starts 10 dB below the design level, rises to the design level (0 dB attenuation) and then drops 10 dB toward evening levels. On weekends and holidays, the population density is much less and so the change is often only 5 dB, i.e., the gain increases from –10 dB to –5 dB. This feature is best implemented in clock-based digital signal pro- cessing systems where the required versatility is controlled by software and is quite simple to program. Programmed level control has been used successfully for over twenty-five years, but few specifications include it.
Adaptive level control
The weakness of programmed level control is obvious— the installer must predict how much activity sound there is going to be during any hour of any day. Adaptive level con- trol is a further refinement, although it requires more thought about how much privacy is needed.
The basic structure of such a system is as follows. A spa- tially diverse set of microphones is placed in a room and overall level data are mixed and sent to the generator. The data are converted to one second Leq values and a running level histogram is constructed. The histogram data are con- verted to a running percentile table. The L10 and L90 (or L99)
24 Acoustics Today, October 2007