I have an analog signal(ECG) ranging in amplitude from 0.1 to 2.5 and I want to quantize it, to convert the ECG signal into bit stream. The problem I am facing is I don't know what parameters in quantize function will lead to the required quantization range.
Can anyone help me in how to do this using quantize function and what should be the parameters that will lead me to the required range of quantization values.

1 Kommentar

Walter Roberson
Walter Roberson am 27 Mai 2019
Due to the laws of the USA, we are not permitted to discuss encryption.

Melden Sie sich an, um zu kommentieren.

Antworten (3)

Rick Rosson
Rick Rosson am 12 Apr. 2012

2 Stimmen

Quantize Function
function y = quantize(x)
xMin = 0.1;
xMax = 2.5;
N = 64;
y = floor( N * log(x/xMin) / log(xMax/xMin) ) ;
end
Test Script
x = 0.1 + 2.4*rand(100,1);
y = quantize(x);
scatter(x,y);
HTH.
Rick

5 Kommentare

Anas Imtiaz
Anas Imtiaz am 16 Mär. 2013
This does not work for a sinusoidal function. I don't think this is a valid generic solution.
Rick Rosson
Rick Rosson am 17 Mär. 2013
It's kind of difficult to provide a general solution when the questioner has not provided all of the necessary information to answer the question. That's why I parameterized my solution, so that it can act as an example of one possible method, and it can be modified depending on the specific requirements.
Neha Gupta
Neha Gupta am 26 Mär. 2019
i have an Speech signal whose amplitudes are always in the range of -1 and 1. Acoording to your solution my quantized output will always be 0. What changes would you suggest?
Walter Roberson
Walter Roberson am 26 Mär. 2019
The above code assumes non-negative numbers (notice the log() calls).
Neha Gupta, are you looking for uniform quantization, or non-uniform? If you want non-uniform, then how do you want to determine the widths of the bins ?
Image Analyst
Image Analyst am 27 Mär. 2019
Why not use imquantize() from the Image Processing Toolbox where you can specify the edges of the ranges to quantize?

Melden Sie sich an, um zu kommentieren.

Rick Rosson
Rick Rosson am 10 Apr. 2012

1 Stimme

  1. How many quantization levels do you want to have?
  2. Do you want uniformly spaced levels, or non?
  3. Is the signal signed or unsigned?

2 Kommentare

Shravan Kumar Pasupuleti
Shravan Kumar Pasupuleti am 12 Apr. 2012
I want to have 64 quatization levels.
I want them to be non- uniform as the signal I am dealing is an ECG signal.
I don't know what are signed and unsigned signals?
Rick Rosson
Rick Rosson am 12 Apr. 2012
Is the value of the signal always greater than zero (one-sided), or is it less than zero some of the time (two-sided)?

Melden Sie sich an, um zu kommentieren.

Image Analyst
Image Analyst am 17 Mär. 2013

0 Stimmen

If it's already in MATLAB then it's already been quantized. This just changes the quantization to some other quantization. I got the impression he had an analog signal that he needed to quantize (digitize). For that he'd probably need the Data Acquisition Toolbox and an appropriate A-to-D converter device, such as those from Measurement Computing Corporation (like I use) or any of the others listed here. That will take the actual input analog voltage signal and digitize it into a MATLAB variable.

Kategorien

Mehr zu Signal Processing Toolbox finden Sie in Hilfe-Center und File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by