MODELLING OF AUDIO EFFECTS FOR

VOCAL AND MUSIC SYNTHESIS IN REAL TIME

Mr. Dony Armstrong D’Souza1 Dr. Veena Devi Shastrimath.V 2
Sr. Asst. Professor, Dept. of ECE. MITE, Mijar Professor, Electronics and communication Department
(Research scholar N.M.A.M.I.T Nitte.) N.M.A.M.I.T Nitte.
1 2
[email protected] [email protected]

Abstract- Sound effects play an important role in today’s music day musicians. The proposed system is to be developed by
industry. The various effects, modulations to the voice is done. an DSP processor or by utilizing the computer’s line–in or
here we present a musical sound effects processing system Mic-in ports. The High speed processing capabilities of the
based on virtual analog modelling and Digital Signal computer is nowadays used for these kind of applications.
Processing techniques. The modelling is using in matlab and
where in the live streaming of audio has to be converted to
the order of effects are sequenced depending on the musicians
choice. The various techniques of Digital signal processing are digital data, then given to computer where it is processed
used .The comparison of results obtained are done with the and given out through line-out port of the computer. The
available system. commercial available music processors do not have the
flexibility to interchange the chain of effects within the loop
Keywords—Fuzz, Flanger, Chorus, Delay, State variable filter in a required order. This proposed system overcomes this
limitation by independently running the effects and good
I. INTRODUCTION memory or register banks.
The music industry uses various sound effects which are
used in various audio tracks or movies to augment the
experience of movie watching, audio listening, public II. SYSTEM OVERVIEW
address systems and live music concerts. Musical effects
processing and synthesis in real time takes an important
place in most of the sounds of contemporary music Musical Impedance
whenever they are encountered in the theater or surrounding instrument ADC DSP DAC matching
and Proce and and
environment. Almost all the music produced in last few with buffer ssor buffer Amplifier
decades by recording or electrically amplifying uses lots of Transducer speakers
effects processing which may include dynamic compression
or artificial reverberation, and synthesizers or synthetic
instrument sounds play and occupy an increasingly greater
part in the total musical spectrum. In addition, most of the Computer
or a vast majority of these effects are currently implemented
using digital signal processing(DSP)[1][12] techniques,
majorly because of the flexibility and lower cost of modern
digital devices. However during live music or performance,
operation of these effects[7] and synthesis algorithms to be Power Supply
able to work in real time is obviously of supreme
Fig 1. Block diagram of the proposed system
importance. Recorded music however, also typically
requires real-time operation [9] of these devices and The system put forward consists of Musical instrument
algorithms, due to the fact that performers normally wish to with the transducer which may be a pick up or piezo electric
hear the final, processed sound of their instrument while or the output of any musical instrument with electrical
playing live. In recent years while reviews of virtual analog output like aux or mic output. The ADC and Buffer unit for
modeling and synthesis of digital sound’s by synthesizers the signal to be digitized and the buffer is used for the
are to be found and discussed [3] and [4] respectively. storage of data in order to have an easy flow of the sequence
virtual analog oscillator algorithms, tutorials are available of data from ADC to the next stage[2]. The DSP processor
and which is not dealt in this paper as it not in scope and to do the processing of the signal and then buffer and DAC
which has been Documented discussed[5]. For Real-time and then to impedance matching and Amplifier section and
simulation of an interesting analog effects device, the to speakers .The proposed system can also utilize the
voltage-controlled filter or state variable filter is being used. Computer with very high speed processing to run the
In this proposed project the importance is given for the program in matlab[15] and do the processing in real time.
versatile, hassle free, portable and easy to utilize musical The power supply unit supplies the power to the required
synthesizer which is the prime requirement of the modern sections. The computer approach consists of running Matlab

Authorized licensed use limited to: KLE Technological University. Downloaded on October 08,2024 at 14:28:18 UTC from IEEE Xplore. Restrictions apply.
 or octave with streaming capability got by using tool like signal.yh(n) is the high pass signal,yb(n)is the band pass
playrec. signal and x(n) is the input signal.

III. METHODOLOGY where F1=2sin(fc/fs) and Q1=2d----------(3)

The proposed System incorporates the various musical
sound effects in real time into a digitally[10] organized IV. RESULTS AND ANALYSIS
system in which the shortcomings of analog version can be The results of three effects generated are obtained from a
eradicated. The basic construction of the project is proposed 18sec clip played by an guitar which is captured by
by keeping the following design objectives in mind. windows sound recorder and processed. The same clip is
1-To be a versatile and reliable system. played through the commercial synthesizer.
2-To have a system which will be of low cost very compact
,organized and the option of many in one kind of gadget.
3-To be able to process data in real time with the latency 0.6

being negligible to human ear

4-Low power consumption 0.4

5-Smaller in size and should be portable.

The Methodology involves basically 5 steps; 0.2

A. Data Acquisition 0
The data to be acquired is the signal from the musical
instrument and then this may converted to digital by using a -0.2
flash ADC with high resolution. The input can be directly
fed to computer mic or line-in port. -0.4

B. Data storage
-0.6
Data storage is required in order to have proper feeding
of Data to the DSP processor and from processor to DAC .
-0.8
0 1 2 3 4 5 6
C. Selection of Desired effects 5
x 10
The effects required have to be selected[6] and the order
Fig. 2 The input audio clip
must be specified an term as the effects in the sequence as in
Fig 2. This is done by having a switch and choice statements The input clip shown Fig. 2 is of duration 18sec and is
in the program and it switches the sequence of the functions played from an electric guitar and captured through
to be executed. windows sound recorder utility and converted to wavefile.
D. Processing 0.6
The processing used for the effects are for effects like
fuzz[1],wahwah[1],delay[1],chorus[1],Flanger[1],Modulator[ 0.4
1]Can be modelled by the algorithms in[1].The function of
the fuzz is given in [2] by the equation as 0.2

( ) = {2 0≤ ≤ 1/3 0

f(x)= [3-(2-3x) 2] /3 for 1/3 x 2/3----(1)

f(x)=1 for 2/3 x 1 -0.2

Here the equation is given and the matlab program -0.4

written for it gives the values for the function from the array
of the audio signal to be processed. The wah wah effect is -0.6

obtained[11] by the shifting of the band of a band pass filter

over the spectrum by the state variable filters using the -0.8
0 1 2 3 4 5 6
equation, 5
x 10

Fig. 3 Simulated result for the fuzz effect using the proposed system
Yl(n)=F1 yb(n)+yl (n-1)
Yb(n)=F1yh(n)+yb(n-1) --------------(2) The waveform of the Fig. 3 shown is the obtained after
Yh(n)=x(n)-yl(n-1)-Q1yb(n-1) processing from the proposed system with input signal clip
depicted in shown in Fig 2
Here F1 and Q1 are tuning coefficients related to cut-off
frequency fc and damping d. y1(n) is the low pass

Authorized licensed use limited to: KLE Technological University. Downloaded on October 08,2024 at 14:28:18 UTC from IEEE Xplore. Restrictions apply.
 The waveform Fig. 6 is the wah wah effect generated by the
Digitech music processing unit and the input signal is the same as
1
Fig 1.which is captured with windows soundrecorder utility.
0.8 .
0.6
0.4
0.4
0.3
0.2
0.2
0

-0.2 0.1

-0.4 0

-0.6
-0.1
-0.8
-0.2
-1
0 1 2 3 4 5 6 7 8 9
5 -0.3
x 10
Fig. 4 The fuzz effect out of Digitech RP
-0.4
0 1 2 3 4 5 6 7
The waveform of Fig. 4 is the waveform obtained from 5
x 10
Digitech guitar sound processor for the same note played as
the input. Fig. 7 The flanger effect out of proposed project.

The waveform of Fig 7 is obtained from the Flanger

1
effect by executing the file [13] in matlab for the input
waveform of the same sound clip of above examples
0.8

0.6 Expander Signal

0.6
0.4

0.4
0.2

0 0.2

-0.2
0
-0.4
-0.2
-0.6

-0.8 -0.4

-1
0 1 2 3 4 5 6 -0.6
5
x 10
-0.8
Fig. 5 The wah wah effect from proposed system 0 1 2 3 4 5 6 7 8
5
x 10

The Waveform of Fig. 5 is the wah wah effect Fig. 8 The expander effect of proposed project.
generated using [8] a state variable filter with
sweeping frequency varying from 500 to 5kHz. The Waveform of Fig 8 shows the effect of Expander
signal by expandng the wave[14] over a values by the
matlab program in the proposed project.
1
Compressed and Boosted Signal
0.6
0.8

0.6 0.4

0.4
0.2
0.2

0
0

-0.2 -0.2

-0.4
-0.4
-0.6

-0.8 -0.6

-1
0 1 2 3 4 5 6 7 8 9 10 -0.8
5
0 1 2 3 4 5 6 7 8
x 10 5
x 10

Fig. 6 The wah wah effect from proposed Digitech RP Fig. 9 The compressor effect of proposed project.

Authorized licensed use limited to: KLE Technological University. Downloaded on October 08,2024 at 14:28:18 UTC from IEEE Xplore. Restrictions apply.
 The Waveform of Fig 9 shows the effect of Compressor [7]. David Moffat, David Ronan, Joshua D. Reiss. “An Evaluation Of
signal by compressing the values above threshold by the Audio Feature Extraction Toolboxes”. Proc. of the 18th Int.
matlab program in the proposed project. Conference on Digital Audio Effects (DAFx-15), Trondheim,
Norway, Nov 30 – Dec 3, 2015
Overdriven Signal [8]. V. V¨alim¨aki, J. Pakarinen, C. Erkut, and M. Karjalainen, “Discrete-
1
time modelling of musical instruments,” Reports on Progress in
0.8
Physics, vol. 69, no. 1, pp. 1–78, 2006
0.6
[9]. C. M. Cooper and J. S. Abel, “Digital simulation of brassiness and
0.4
amplitude-dependent propagation speed in wind Instruments,” in
0.2
Proceedings of the International Conference on Digital Audio Effects,
0
Graz, Austria, September 2010.
-0.2
[10]. J. Pakarinen, H. Penttinen, V. V¨alim¨aki et al., “Review of sound
-0.4
synthesis and effects processing for interactive mobile applications,”
-0.6
Report 8, Department of Signal Processing and Acoustics, Helsinki
-0.8
University of Technology, 2009
-1
0 1 2 3 4 5 6 7
5
8 [11]. V. Lazzarini, J. Timoney, and T. Lysaght, “The generation of natural-
x 10
synthetic spectra by means of adaptive frequencymodulation,”
Computer Music Journal, vol. 32, no. 2, pp. 9–22, 2008.
Fig. 10 The Overdrive Distortion effect obtained by the [12]. C. Poepel and R. B. Dannenberg, “Audio signal driven sound
proposed project. synthesis,” in Proceedings of the International ComputerMusic
Conference, pp. 391–394, Barcelona, Spain, September 2005.
Fig 10 shows the waveform obtained by symmetrically [13]. J. Pekonen, “Coefficient modulated first-order allpass filter as
clipping with gain to moderate gain. distortion effect,” in Proceedings of the International Conference on
Digital Audio Effects, pp. 83–87, Espoo, Finland, September 2008.
V. CONCLUSION AND DISCUSSION [14]. J. Kleimola, J. Pekonen, H. Penttinen, V. V¨alim¨aki, and J. S. Abel,
As seen by the waveforms the results obtained aare “Sound synthesis using an allpass filter chain with audio-rate
satisfactory.By comparing Fig. 3 and Fig 4 and listening to coefficient modulation,” in Proceedings of the International
the audiowave file obtained it is seen that the clarity is Conference on Digital Audio Effects, Como, Italy, September 2009.
better in proposed system and the other effect wah wah [15]. Ren Gang,Gregory Bocko,Justin Lundberg and Stephen Rossener “A
from Fig 5 and Fig 6 it is seen that there is clarity in Real-Time signal processing Franmework of Musical Expressive
Feature Extraction using MATLAB” 12th International Society for
proposed system waveform and the audio output generated.
music information retrieval Conference (ISMIR 2011) .
The further effects flanger, delay, chorus, reverberation are
to be tested. Further the virtual analog models using analog
equations have to be tested and implemented.The subjective
analysis was found to be good.The power for the proposed
system is the regular power supply used with less power
consumption in mind. The processing time required depends
on the samples.This has to be optimized as we use buffers of
3 seconds for recording and playing the signal.However
with a devolpment board the response time required is less
with optimized number of effects.

REFERENCES
[1]. U. Z¨olzer, Ed, DAFX—Digital Audio Effects, John Wiley &Sons,
New York, NY, USA, 2002.
[2]. Dony D’souza, Sunitha Lasrado, Ganesh V.N.-“Virtual analog and
Digital effects model for vocal and musical sound
Synthesis”.ICECE- 2012 conference..Vol-1,pp-247-251. IRnet
[3]. J. Pakarinen, Vesa valimaki,Federico Fontana,Victor Lazzarani,and
Jonathan S.Abel “Recent Advances in Real-Time Musical
effects,Synthesis,and Virtual Analog Models”.Journal volume
2011,article ID940784.Sound and Music technology Research
Group,National university of Ireland Maynooth,Ireland.
[4]. J. O. Smith, “Physical Audio Signal Processing,”
2010,https://ccrma.stanford.edu/ jos/pasp/.
[5]. V. V¨alim¨aki, F. Fontana, J. O. Smith, and U. Z¨olzer, “Introduction
to the special issue on virtual analog audio effects and musical
instruments,” IEEE Transactions on Audio,Speech and Language
Processing, vol. 18, no. 4, pp. 713–714, 2010.
[6]. V. V¨alim¨aki and A. Huovilainen, “Antialiasing oscillators in
subtractive synthesis,” IEEE Signal Processing Magazine, vol. 24,
no. 2, pp. 116–125, 2007.

Authorized licensed use limited to: KLE Technological University. Downloaded on October 08,2024 at 14:28:18 UTC from IEEE Xplore. Restrictions apply.