speech production

I am trying to generate the /a/ vowel by using speech production model (digital filters)by following codes.

I am not able to hear the /a/.

function dfcoeff;

imp=[1 zeros(1,33075)]; %create impulse response
[b,a]=gp(0.88); % generate glottal pulse filter coefficients
gp88=filter(b,a,imp); % create a single glottal pulse period, 120 long

Ff=5512.5; % folding frequency is 5000 Hz
a1=0.98; %  radius for pole 1
p1=710; % angle for pole 1
a2=0.96; %  radius for pole 2
p2=1100.0; %  angle for pole 2
a3=0.85; %  radius for pole 3
p3=2450.0; %  angle for pole 3
a4=0.85;  %radius for pole 4
p4=3600.0; % angle for pole 4
sf=10; % scale factor for z-plane

[b,a]=allpoles(Ff,p1,a1,p2,a2,p3,a3,p4,a4,sf);%  all-pole filter
%freqz(a,b) %view frequency response

y=filter(b,a,gp88);%  filter response

x=conv(y,[1 -1]);% model of sound passing lips

   soundsc(x,11025);

function [b,a]= gp(ap)
% returns digital filter coefficients using parameter ap.
e=2.71828;
b=[0 -ap*e*log(ap)];

a=[1 -2*ap ap^2];

function [b,a]=allpoles(Ff,p1,a1,p2,a2,p3,a3,p4,a4,sf)
 % Ff = folding frequency
% p# = radius of pole pair # from origin
% a# = frequency location of pole pair #
% sf = pole-zero plot scale factor
% b = vector of input coefficients
% a = vector of output  coefficients
% [b,a] = allpoles(p1,a1,p2,a2,p3,a3,p4,a4)
p1a=p1*(cos(a1*pi/Ff)+sin(a1*pi/Ff)*j);
 p1b=conj(p1a);
p2a=p2*(cos(a2*pi/Ff)+sin(a2*pi/Ff)*j);
 p2b=conj(p2a);
p3a=p3*(cos(a3*pi/Ff)+sin(a3*pi/Ff)*j);
 p3b=conj(p3a);
p4a=p4*(cos(a4*pi/Ff)+sin(a4*pi/Ff)*j);
 p4b=conj(p4a);
[b,a]=zp2tf([],[p1a;p1b;p2a;p2b;p3a;p3b;p4a;p4b],sf);

Any suggestion?