AUDIBLE FREQUENCY RANGES 



625 



3 first violins, 1 second violin, 1 viola, 1 'cello, 1 string bass, 1 flute, 

 1 oboe, 2 clarinets, 1 bassoon, 2 French horns, 2 trumpets, 1 trombone, 

 1 drummer. The players were seated in concert arrangement with the 

 violins about 8' from the microphone. Ten engineers experienced in 

 quality judgments acted as observers. In these tests the filter condi- 

 tions were always presented as "B" and the observers were asked to 

 rate the quality of the "B" condition numerically, considering the "A" 

 condition to possess a quality of 1.0. The ratings could be either less 

 than 1.0, indicating a degradation, or greater than 1.0, indicating an im- 

 provement. Conditions were switched A-B-A-B^, continuing until 

 all observers had obtained a judgment, but the filters were presented in 



100 500 1,000 



CUT-OFF FREQUENCY IN CYCLES PER SECOND 



Fig. 6 — Quality of orchestral music as function of cut-off frequency. 



perfectly random order. Therefore the observers were never informed 

 as to what filter was being tested ; they only knew that "A" represented 

 a quality of 1.0 and "B" a condition to judge. 



The orchestra played the Strauss waltz, "The Beautiful Blue Dan- 

 ube," for the first test. It was orchestrated so that most of the instru- 

 ments played most of the time. All filters except 30 and 40 high pass 

 were presented to the observers. From the results a list of filters which 

 all observers had rated as better than 0.5 was compiled for presentation 

 during a second test. This time the music was "In The Village," a 

 composition of Godard. It was a selection in which many instruments 

 had solo parts and in which therefore the character of the music 

 changed rapidly. 



The average ratings for both runs are plotted in Fig. 6. The two 

 sets of data agree reasonably well except at the extreme ends where 



