Lab9: Spectral Transformations

The sms.m program can be extended by adding transformations. In the sms.m file locate the line "Transformations". This is the place where the sinusoidal plus residual data can be transformed before a sound is synthesized.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%===== Transformations =====
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

...

The variable iloc holds an array of peak positions, whereas the variable ival holds the corresponding amplitude values. We can use these two arrays in order to fill syniloc and synival, the corresponding arrays that get synthesized.

1. Filtering with arbitrary resolution

We can implement a band-pass filter defined by (x, y) points where x is the frequency value in Hertz and y is the amplitude factor to apply. In the example code given below, we define a band pass filter which suppresses frequencies not included in the range [2100 3000].

Download functions interp1.m and lookup.m.

Implement a function

function [syniloc synival] = transform(iloc,ival,mode)

%===== Filtering with arbitrary resolution =====

Filter=[ 0  2099 2100 3000 3001 22050 ;
         0  0    1    1    0    0 ];

[syniloc,ind] = sort(iloc);
FilterEnvelope = interp1(Filter(1,:)',Filter(2,:)',syniloc/N*SR);
synival = ival(ind)+(20*log10(max(FilterEnvelope,10^-9)));
synival(ind)=synival;
syniloc(ind)=syniloc;

2. Filter out specific partials

%=== voice to clarinet ===

syniloc = iloc;
synival = ival;
if (isHarm == 1)
  pitchvalue
  for i=1:nSines
    harmNum = round(((iloc(i)-1)/w1Length*SR/2)/pitchvalue);
    if (mod(harmNum,2)==0)  % case of an even harmonic number
    synival(i) = -100;
    end
  end
end

3. Frequency Scaling

In a similar way, we can apply a frequency scaling to the sinusoidal components of our modeled sound. In that way, we can transpose all the partials in the spectrum or reproduce pseudo-inharmonicities like the higher partials frequency stretching characteristical in a piano sound. In this first example we introduce a frequency shift factor to all the partials of our sound. Note, though, that if a constant is added to every partial of a harmonic spectrum, the resulting sound will be inharmonic. Implement this as mode 3 of the transform function.

%==== Frequency Shift =====
fstretch = 100;  % frequency shift in Hz
syniloc = iloc + round(fstretch/SR*N);
syniloc = syniloc.*(syniloc<=N/2);
synival = ival; 

%===== Frequency Stretch  =====
fstretch = 1.1;
[syniloc,ind] = sort(iloc);
syniloc = syniloc.*((fstretch).^[0:nSines-1]');
syniloc = syniloc.*(syniloc<=N/2);
synival = ival;

Finally mode 5 is frequency scaling, where the pitch of the sound is changed, while the duration stays the same.

%==== Frequency Scale =====
fscale = 1.2;  % frequency scaling factor
syniloc = iloc * fscale;
syniloc = syniloc.*(syniloc<=N/2);
synival = ival;

4. Transformation plots

Add the function transform to the sms.m file and make a plot of the magnitude values of a significant frame of DAFx_outresynth, the resynthesized sound, and the same frame of DAFx_in, the input sound sax.wav. (In one plot, you can use the "hold on" function).

5. Pitch Transposition with Timbre Preservation

Pitch transposition is the scaling of all the partials of a sound by the same multiplying factor. Here we introduce the concept of timbre preservation by leaving the spectral shape unmodified. For that reason we scale the frequency of each partial applying the original spectral shape.

where PitchTransposition is the following function:

1. Change the sms.m program into a pitch transposition with timbre preservation function. The input to the function should be the sound file to be changed and the pitch transposition factor.

a.       Use the function with the sound file elvis.wav and apply different transposition factors (for example: 0.2, 0.5, 2, 3, …). Make sure that the output sound equals the original sound when no transposition is applied (change analysis parameters if necessary).

b.      Compare the effect of applying pitch transposition with and without timbre preservation. Describe what is the difference.

c.       Plot a magnitude spectrum of the input file (at a meaningful place) and one magnitude spectrum for a few pitch transpositions. Explain the results.

6. Vibrato and Tremolo

Vibrato and tremolo are common effects used with different kinds of acoustical instruments, including the human voice. Both are low frequency modulations: vibrato is applied to the frequency and tremolo to the amplitude of the partials. Note, though, that in this particular implementation, both effects share the same modulation frequency.

%===== vibrato and tremolo =====
if (isHarm == 1)
vtf = 5;       % vibrato-tremolo frequency in Hz
va  = 10;     % vibrato depth in percentil
td  = 3;       % tremolo depth in dB
synival = ival + td*sin(2*pi*vtf*pin/SR);   % tremolo
pt = 1 + va/200*sin(2*pi*vtf*pin/SR);    % pitch transposition factor  
[spectralShape, shapePos] = CalculateSpectralShape(iloc, ival, MinMag, N);
[syniloc, synival] = PitchTransposition(iloc, ival, spectralShape, shapePos, pt, N);
%--- comb filtering the residual
CombCoef = 1;
resfft = combFilter(resfft, N, SR/(pitchvalue*pt), CombCoef);
end

1. Change the sms.m program into a function to add vibrato and tremolo into a sound. The input to the function should be (1) the sound file to be changed, (2) the vibrato-tremolo frequency in Hz, (3) vibrato depth in percentil, and (4) tremolo depth in dB.

a.       Use the function with the sound file Trumpet-01-mf-A3.wav and apply different vibrato and tremolo values. Make sure that the output sound equals the original sound when no vibrato is applied (change analysis parameters if necessary).

b.      Increase the vibrato frequency to values higher than 50 Hz and the depth to values higher than 200. Describe and explain what happens.

c.       Increase the tremolo frequency to values higher than 50 Hz and the depth to values higher than 50 dB. Describe and explain what happens. What is the difference with the effect of the question (b).