Scribd
Upload
33 views

TD1 Sampling Correction

1. The document discusses digital signal processing and sampling theory. It includes exercises on sampling frequency, the discrete Fourier transform, and signal quantization. 2. One exercise calculates that a minimum sampling frequency of 44 kHz is needed to digitize an audio signal with maximum frequency of 22 kHz. At this rate, 70 minutes of two microphone audio would be 739.2 megabytes in size. 3. Another exercise discusses the mean, variance, and signal-to-quantization noise ratio of a signal after analog to digital conversion, showing how the noise ratio improves by 6.02 dB for each additional bit of quantization.

Uploaded by

Gabriel Blanchet
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
33 views

TD1 Sampling Correction

1. The document discusses digital signal processing and sampling theory. It includes exercises on sampling frequency, the discrete Fourier transform, and signal quantization. 2. One exercise calculates that a minimum sampling frequency of 44 kHz is needed to digitize an audio signal with maximum frequency of 22 kHz. At this rate, 70 minutes of two microphone audio would be 739.2 megabytes in size. 3. Another exercise discusses the mean, variance, and signal-to-quantization noise ratio of a signal after analog to digital conversion, showing how the noise ratio improves by 6.02 dB for each additional bit of quantization.

Uploaded by

Gabriel Blanchet
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

DIGITAL SIGNAL PROCESSING - CORRECTION TD N°1 -

EXERCISE N°1 EXERCISE N°3


+ the maximal frequency value is 22 kHz. Thus, the
X e ( f ) = Fe Xa(f − kFe )X ( f )
e
minimal sampling frequency is 44 kHz. If we consider
k =− two microphones at 44 kHz and 16 bits, we obtain 1.408
Mbits/s. Consequently for 70 minutes, we obtain : 739.2
Mbytes.

f Exercise n°4
-10F -5F -F +F +5F +10F 1)
e(x)

/2
The sampled signal is written : x

x(nTe ) = A cos(2f 0 nTe ) + B -/2 i (i+1) 


The Fourier Transform is written :
X a ( f ) = ( f − f 0 ) + ( f + f 0 ) + B( f )
A
2
2) The mean :
The Discrete Fourier Transform is expressed :

A
 
X e ( f ) = WFe ( f − f 0 ) + WFe ( f + f 0 ) + BW Fe ( f ) +
+
2 1
2 me =  pe (u)udu =  
udu = 0
− 
−
2
The variance
B
 
+ +

 
A/2 2 2 1 2 2
 e2 = E e 2 =  p e (u )u 2 du =  
u du =
12
-Fe Fe/5 Fe  
− −
2 2

1) The signal-to-quantization noise ratio:


EXERCISE N°2
1. The Shannon theorem is respected because Fe  2 B  12 2x 2 2b 
 = 10 log10  , d'où
( 90 MHz  50 MHz )  2 
 A 
 
2. |Xe(f)|  = 6.02b + 20 log10  x  + 10.8
25 MHz  A 
6.02 dB per additional bit.
f 5) For a sinusoidal signal, the amplitude (Peak value) la
is 2 x and the condition of non-saturation can be
45 67.5 112.5 MHz A
22.5 MHz MHz written : 2 x  . Thus,   97.8 dB .
MHz
90 2
MHz For a Gaussian signal with an estimation of its
3.
The solution is not valid for 100 MHz and Fe=50 MHz. A
amplitude of 4 x , we obtain: 4  x  , and
A sufficient condition by down-sampling at Fe  2 B 2
F consequently   88.7 dB .
and f = kFe + e
4
Fe/2

kFe (k+1)Fe

f=kFe+Fe/4

Signal

You might also like