Istituto
Elettrotecnico Nazionale
Galileo
Ferraris
Sede di Corso Massimo d'Azeglio, 42
- 10125 TORINO - Telefax (+39) 011.650.76.11
Sede di Strada delle Cacce, 91 - 10135
TORINO - Telefax (+39) 011.34.63.84
Telefono (+39) 011.3919.1 (selezione
passante) – sito Internet: http://www.ien.it
________________
RELAZIONE
N.
34380 constituted of n. 4 pages released in date 20/09/2001
to Mr.
Michele Greci. via XX Settembre 98/E - 00187 Roma [RM] - Italy
conforming
to the application prot. IEN 6747/01 in date 18/09/2001
Object:
comparative measurements of sound pressure levels on two guitars.
1)Preliminary statements
The customer
has asked for a comparative measurement of sound pressure levels on two guitars
which he furnished himself:
Purpose of
the measurements is to establish the superiority of the sound level produced by
the Greci guitar.
2) EXECUTION
OF MEASUREMENTS
With the purpose to eliminate from the results the
human performance procedures on the instrument, a mechanical device was
conceived and created at the university La Sapienza in Rome, Faculty of
Engineering, under the direction of professor Enzo Marchetti of the Department of
Mechanics and Aeronautics. The guitar is placed on a horizontal surface; a
cylinder, which carries a projecting part similar to a plectrum, is set in slow
rotation by an electric scaling engine,
which at every turn pinches the string of the guitar with a procedure analogous
to that of a human performance but highly repeatable. It is also
possible to shift the guitar horizontally so that it offers its six strings
consecutively to the point of stimulation of the plectrum.
The device, set on a bearing surface, has been
situated in a semi-anecoic chamber. Above it, at a distance of 1 metre from the
soundboard of each guitar and on the perpendicular of the point of stimulation,
has been situated a omnidirectional mesure microphone for free field, turned toward
the instrument. By means of the microphonic preamplifier, the microphone has
been connected to a real time two channel digital frequency analyser of 1/3,
1/12 and 1/24 of octave.
Measuring instrumentation is therefore the following:
With the described sequence,
measurements of the level of sound pressure have been effected through
acquisition of the spectrum of 1/24 of octave and at the same time measurements
have been effected of global, linear and ponderable A level; measurements have
been repeated three times on each open string of each guitar, so as to verify
repeatability. The analyser was started using the trigger on the global signal:
when a particular sound level was reached, varying according to the string and
the instrument, correspondent to the first beginning of the sound, the
acquisition began, with linear integration time equal to 4 seconds, such as to
include the part of the sound decadence distinguished by greater energy.
Transferring on a calculator the levels of each band of 1/24 of octave and elaborating them on a
spreadsheet, the energetic average value and the standard deviation of sound
pressure level of the three measures have been then calculated, for each band of 1/24 of octave. The datum of standard
deviation regarding average value calculated on global levels (linear and
weighted A.) has been, in the worst case, lower than 1,5 dB, with values on the
average lower than 1,0 dB, thus confirming repeatability. Also the level of
residual noise has been measured (due to rotation of the electric engine of the
device for mechanical stimulation), so as to check that it didn't interfere
with the sound produced by the instrument. The result was a difference between
the level of the background noise and that of each string of each instrument
constantly higher than 20 dB for every band of the spectrum, for each string
and each instrument, so as to confirm that in no case the measure effected on
the guitars has been influenced by background noise (error lower than 0,1 dB).
3) PRESENTATION OF
MEASUREMENTS’ RESULTS
As previously expressed, the
data of the spectrums of 1/24 of octave (three for each string and for each
guitar) have been transferred on a spreadsheet so as to calculate the energetic
average value and the standard deviation. We have so obtained 12 average spectrums (six for each guitar and for each instrument,
of which the data are expressed in dB (A) and dB corresponding to the global
levels, weighted Á. and linear, in tab.1).
Table 1): gives the comparison
of the global and weighed sonorous levels A of the two guitars, in dB.
|
|
Greci Guitar |
Classic Guitar |
||
|
NOTA (Note, string) |
A |
Linear |
A |
Linear |
|
MI (E – sixth) |
69,7 |
73,7 |
60,7 |
65,7 |
|
LA ( A – fifth) |
66,2 |
72,1 |
61,7 |
67,8 |
|
RE (D – fourth) |
68,9 |
73,6 |
60,4 |
64,5 |
|
SOL (G –
third) |
63,0 |
68,3 |
58,9 |
63,1 |
|
SI (B – second) |
65,8 |
69,6 |
60,7 |
63,5 |
|
MI Cant. (E – first) |
65,3 |
66,5 |
58,4 |
59,7 |
The data of the table are also
expressed in the diagrams of the two accompanying tables (fig. 1 and fig. 2),
which set a comparison, respectively, between the global weighted A and linear
levels of all the strings of the two instruments on histograms and diagrams.
The data regarding the spectrums of 1/24 of octave have been here
expressed exclusively in graphic form
due to the great number of data (194 values for each string and each
instrument). In figures 3- 8 in the two accompanying tables the spectrums of each instrument are
compared, each string separately, on a diagram with lines of different colours,
according to the instrument. The numerical data are, in any case, available on
the floppy disk enclosed, which is integral part of this paper, as a file
(chitarre.xls) in Microsoft Excel 2000, which contains the whole spreadsheet
used for the elaboration.
4) INTERPRETATION OF THE RESULTS