Convolution not working with 'uint16'

2 Ansichten (letzte 30 Tage)
William Greenway
William Greenway am 31 Jan. 2022
Beantwortet: Star Strider am 31 Jan. 2022
I'm importing some .dat data into MATLAB to perform a discrete time convolution on a 'uint16' number.
I'm using the fread and conv functions to calculate a convolution. I find that when adding the 'uint16' variable to fread the convolution gives really small values.
test1 = [0.25 , 0.25 , 0.25 , 0.25];
% Answer gives very small convolution
x = fopen('in.dat');
A = fread(x, 'uint16');
conv(A,test1)
% Gives correct answer
z = fopen('in.dat');
C = fread(z);
v = nonzeros(C);
conv(v,test1)
Why does adding 'uint16' to the input produce the incorrect result? When looking at the difference between A and v, I cannot see any difference?

Akzeptierte Antwort

Star Strider
Star Strider am 31 Jan. 2022
That likely has to do with the precision of the calculations. I would convert them to double, do the calculations, and then convert back to unit16.

Weitere Antworten (0)

Community Treasure Hunt

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

Start Hunting!

Translated by