Page 97 - The Master Handbook Of Acoustics
P. 97

72   CHAPTER THREE



                                   adjusted until the tone is just hidden or masked by the noise. Only
                                   the noise in the critical band centered on 1,000 Hz is effective in
                                   masking the tone. If the noise is expressed in sound-pressure level
                                   of a band 1 Hz wide, the voltage of the tone then corresponds to the
                                   1-Hz sound-pressure level of the noise at the masked point. Adjust-
                                   ing the voltage until the tone is just masked should yield one of the
                                   points on our noise spectrum graph. For convenience, let us assume
                                   that this voltmeter is calibrated in dB referred to some arbitrary base
                                   such as 1 volt (dBv). Referring to Fig. 3-11, note that the critical
                                   band centered on 1,000 Hz is 160 Hz wide. This can also be
                                   expressed in decibels by taking log 10 160 = 22 dB; this 22 dB, repre-
                                   senting the width of the critical band as it does, must be subtracted
                                   from the voltmeter reading in dB. This gives one point on the noise
                                   spectrum graph. By repeating the process for other frequencies, a
                                   series of points is obtained that reveal the shape of the noise spec-
                                   trum. If the recording and the entire measuring system (including
                                   the observer’s ears) were calibrated, the absolute levels for the noise
                                   spectrum could be obtained.
                                      The important point here is that there is such a set of filters in our
                                   head that could be put to such a task, not that this method will ever
                                   replace a good sound level meter equipped with octave or one-third
                                   octave filters. Surely human variables would far exceed sound-level
                                   meter fluctuations from day to day, and what the observer eats for
                                   breakfast has no effect on the sound-level meter, although it might
                                   affect the dependability of the readings made with physiological
                                   equipment.

                                   Meters vs. the Ear


                                   There still remains a great chasm between subjective judgments of
                                   sound quality, room acoustics, etc., and objective measurements. Con-
                                   siderable attention is being focused on the problem. Consider the fol-
                                   lowing descriptive words, which are often applied to concert-hall
                                   acoustics 19,20 :
                                                       warmth              clarity
                                                       bassiness           brilliance
                                                       definition          resonance
                                                       reverberance        balance
   92   93   94   95   96   97   98   99   100   101   102