|
|
|
|
|
THE NATURAL TRUMPET |
|
AND ITS VIRTUAL SOUND |
|
|
|
L.Tronchin (*), A.Farina (**), A.Cocchi |
|
|
|
(*) DIENCA – CIARM, Viale Risorgimento, 2 |
|
40136 Bologna (Italy) - http://ciarm.ing.unibo.it |
|
|
|
(**) Industrial Engineering Department., |
|
University of Parma, Via delle Scienze 181/A, |
|
43100 Parma (Italy) - http://pcfarina.eng.unipr.it |
|
|
|
|
|
|
|
|
A) Virtual reconstruction of the sound
quality of musical
instruments. |
|
Violins, trumpets, flutes |
|
|
|
B) Virtual reconstruction of spaces
for music and speech. |
|
Theatres, churches , auditoria |
|
|
|
|
Restoration cultural heritage |
|
Storage
museums |
|
Sound and music: “visualisation” and “auralisation”. |
|
Multimediality and acoustics: could they live together? |
|
|
|
|
Evaluation of sound quality of sound chests |
|
in different violins |
|
Measurements of
IR (force pressure) |
|
on the bridge |
|
Calculation of inverse filter |
|
Recordings in anechoic chamber |
|
Deconvolution:
rec * F.I.R. (getting “dry” music) |
|
Convolution virtual sound |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Direct (up) and reciprocal (down) method of
measurements |
|
|
|
|
|
|
Modern trumpets |
|
Vincent Bach |
|
Yamaha |
|
Yamaha custom |
|
|
|
Baroque age trumpets |
|
Natural |
|
Hunting |
|
|
|
|
The impulse response (mouthpiece -> radiated
field) characterizing different instruments are measured |
|
A music piece is played on one instrument, and
recorded in anechoic environment |
|
The impulse response of this instrument is
inverted |
|
The anechoic recording is deconvolved by
convolution with the inverse filter |
|
The deconvolved recording is used as the
starting point for subsequent reconvolutions with the IRs of the different
instruments |
|
|
|
|
Shift register of order N |
|
|
|
|
CoolEditPro is employed for simultaneous
playback and recording – thereafter, a special plug-in is invoked for
deconvolving the impulse response |
|
|
|
|
The test signal is a sine sweep: |
|
the frequency exponentially increases with time |
|
|
|
|
The impulse response is recovered by convolving
the system’s response with a proper inverse filter, obtained by the
time-reversal mirror of the excitation signal, with a 6dB/octave
equalization |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Conforzi-Callegari natural trumpet |
|
(a tonal copy of the "Wilhelm Magnus Ehe
I" model, XVII century) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Hunting-trumpet Meinl&Lauber |
|
|
|
|
|
|
The original response h(t) is first FFT
transformed: |
|
|
|
(1) |
|
|
|
|
|
Then the complex spectrum C is inverted: |
|
|
|
(2) |
|
|
|
|
|
And the result is back-transformed to time
domain |
|
|
|
(3) |
|
|
|
|
|
|
|
|
IR measurements: |
|
- Maximum length sequence |
|
- Exponential sine sweep |
|
|
|
|
|
|
|
Inverse Filter calculation: |
|
- Toeplitz technique
(Morjoupolous 1985) |
|
- Kirkeby technique
(Kirkeby, Farina 1998) |
|
|
|
|
|
|
|
|
|
|
|
|
A B |
|
|
|
|
|
|
|
|
|
C D |
|
|
|
|
|
A-original (Conforzi-Callegari) B-dry (deconvolved) |
|
C-reconvolved (as A) D-with IR of
hunting-trumpet |
|
|
|
|
|
|
The measurement of IR is feasible in trumpets –
the sweep method is better than MLS |
|
The computation of inverse filter is better with
Kirkeby-Farina method |
|
The measurements could be improved with smaller
microphones (1/4”) |
|
|
|
|
As the new sweep measurement method also
characterizes the not-linear response of the instrument, a multiple order
convolution will be employed for attempting the virtual recreation of the
harmonic distortion. |
|
The reciprocity method will be attempted (as
already done on violins) |
|
A subjective listening experiment will be
started, for trying to understand the connection between measurable
objective parameters and the perceived musical quality |
|
|
|
|
CIARM site on musical acoustics: |
|
HTTP://www.ciarm.ing.unibo.it |
|
|
|
Aurora site (software plug-ins for measurements
and convolution): |
|
HTTP://www.ramsete.com/aurora (info) |
|
FTP://pcangelo.ramsete.com/aurora (download) |
|
|
|
Notes