|
|
View previous topic :: View next topic |
Author |
Message |
juanodiy
Joined: 07 Sep 2012 Posts: 17
|
How to separate an 8 bit int |
Posted: Sun Nov 04, 2012 9:11 pm |
|
|
Hi, again asking for help, I have a problem separating an 8 bit int. I am doing a QAM modulator with PIC 18f452.
The 16 QAM modulation use 4 bits, if i want to modulate a PCM word, with 8 bits, i need to separate them in groups of 4, or individually and copy to a char, for be compared.
Example:
int x
x=rand(x)
x is now ( 1 0 1 0 0 0 1 0 ) for example
i need to separate
1 0 1 0
and
0 0 1 0
and copy :
c[i]=x(i)
I appreciate any help, of any master in this forum thanks to all |
|
|
juanodiy
Joined: 07 Sep 2012 Posts: 17
|
|
Posted: Sun Nov 04, 2012 10:32 pm |
|
|
I think I could use these operators "<< >>", but if I want every bit in every position of a vector I don't know how, please any idea will be great. |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19480
|
|
Posted: Mon Nov 05, 2012 2:48 am |
|
|
Code: |
int x, nibble;
x=rand(x);
nibble=x & 0xF;
//nibble is now the low nibble of x
swap(x);
nibble=x & 0xF;
//nibble is now the high nibble of x
|
Key is that the PIC has a single instruction (swap) which swaps the high and low nibbles of a byte. Much more efficient than a four bit rotation. CCS allows direct access to this instruction.
Best Wishes |
|
|
Mike Walne
Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK
|
|
Posted: Mon Nov 05, 2012 9:48 am |
|
|
The CCS manual does explain how you could do it in 'C'.
As Ttelmah has shown, you might as well use the built in Microchip instruction.
Mike |
|
|
juanodiy
Joined: 07 Sep 2012 Posts: 17
|
|
Posted: Mon Nov 05, 2012 11:20 pm |
|
|
First of all, thanks for your time. I already have used the function swap, indeed it works, but I want to copy to a char vector. Here is my example:
Code: |
#include <18f452.h>
#device adc=8
#use delay(clock=20000000)
#fuses HS,NOWDT,NOPROTECT,NOLVP
#include <math.h>
#include <lcd.c>
int8 data=0,val,x,aux,i=0,m=0,y=0,p=0;
char c[8],high[4],low[4];
short res=0;
float ai,aq;
char s1[16][4]={
{'0','0','0','0'},
{'0','0','0','1'},
{'0','0','1','0'},
{'0','0','1','1'},
{'0','1','0','0'},
{'0','1','0','1'},
{'0','1','1','0'},
{'0','1','1','1'},
{'1','0','0','0'},
{'1','0','0','1'},
{'1','0','1','0'},
{'1','0','1','1'},
{'1','1','0','0'},
{'1','1','0','1'},
{'1','1','1','0'},
{'1','1','1','1'},
};
float compare()
{
while(low!=s1[p][])
{
p++;
}
printf(lcd_putc,"encontro");
}
void main()
{
lcd_init();
setup_adc(ADC_CLOCK_DIV_32);//
setup_adc_ports(RA0_RA1_RA3_ANALOG);
set_adc_channel(0);
lcd_putc("dato:");
do {
read_adc(ADC_START_ONLY); //start the ADC reading
//Now wait for the ADC to complete
while (!adc_done()) ;
val=read_adc(ADC_READ_ONLY); //read the value
delay_us(5); //ensure acquisition has completed
lcd_putc("\f");
lcd_gotoxy(1,1);
printf(lcd_putc,"%u",val);
for(x=0;x<=7;x++) //bucle for conversion dec to bin
{
res=val%2;
aux=val/2;
val=aux;
if(res==0){
c[x]='0';
}
else
{
c[x]='1';
}
}
for(i=0;i<=3;i++) //copy LSB part from c
{
low[i]=c[i];
}
for(m=0;m<=3;m++) //copy MSB part from c
{
high[m]=c[m+4];
}
lcd_gotoxy(1,2);
printf(lcd_putc,"%s",low);
lcd_gotoxy(6,2);
for(y=0;y<=3;y++)
{
printf(lcd_putc,"%c",high[y]);
}
compare(); //call function compare
delay_ms(5000);
} while(TRUE);
}
|
Now the idea is to compare to values of s1 with high and low, the combinations possible for a nibble. For this I create a function called compare, but it doesn't work. I don't know how to compare a matrix with a vector. Again thanks for your answers, any idea will be grateful. |
|
|
Mike Walne
Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK
|
|
Posted: Tue Nov 06, 2012 2:47 am |
|
|
I don't follow exactly what you're trying to do.
Please explain what you are attempting to compare.
Mike |
|
|
juanodiy
Joined: 07 Sep 2012 Posts: 17
|
|
Posted: Tue Nov 06, 2012 11:23 pm |
|
|
Ok, thanks mike, I'm trying to do a comparison, between 3 arrays.
For example: I need to modulate a 16-QAM signal; this modulation use 4 bits, thats what I need to separate them, then I need to compare the 2 nibbles:
1001 0101
high low
and I have a predetermined matrix with all possible combinations, ex:
0000
0001
0010
0011
...
I want an efficient method for compare this rows with the variables high and low, thats what i think in a matrix, but i want to compare complete array, it means:
{ 0 ,1 ,0 ,0} == {0 ,1 ,0 ,0 }
not position by position, like:
for(....)
{
for(....)
{
My question is, is there a method for compare a matrix with a vector rows by rows? not position by position. |
|
|
Mike Walne
Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK
|
|
Posted: Wed Nov 07, 2012 3:03 am |
|
|
Maybe you're going about the task the wrong way.
I don't want to appear thick, but I still don't get it.
Can you outline in pseudo code the complete process.
Then either:-
I/We can see how to do what you're asking for as you see it.
OR
Suggest an alternative method which achieves the efficiency you're after.
Mike
EDIT
I keep getting stuck on the idea that {'0','0','1','0'} translates to 2 (or 4 depending on which way you read the bits).
So your high and low nibbles are then much easier to compare. |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19480
|
|
Posted: Wed Nov 07, 2012 3:05 am |
|
|
Seriously, for is the way to go. There is nothing 'more efficient' about trying to compare the complete array, languages that offer features like this (C doesn't), do it 'behind the scenes', with simple 'for' loop tests....
However I'm not sure what you actually want to achieve. As a comment (for instance), your rows in the array you show are just binary numbers from 0 to 15. No comparison needed.....
Best Wishes |
|
|
Mike Walne
Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK
|
|
Posted: Wed Nov 07, 2012 4:37 am |
|
|
Thanks Ttelmah, I was beginning to believe I was the only one in the dark.
Mike |
|
|
juanodiy
Joined: 07 Sep 2012 Posts: 17
|
|
Posted: Wed Nov 07, 2012 10:22 am |
|
|
Ok, thanks to both, look I'm gonna show you my advance,
Code: |
#include <18f452.h>
#device adc=8
#use delay(clock=20000000)
#fuses HS,NOWDT,NOPROTECT,NOLVP
#include <math.h>
#include <lcd.c>
#include <stdlib.h>
#define RAND_MAX 256
int8 data=0,val,x,aux,i=0,m=0,y=0,p=0,flag=0,k=0,j=0;
char c[8],high[4],low[4];
short res=0,encontro=0;
float ai=0,aq=0.1;
char s1[16][4]={
{'0','0','0','0'},
{'0','0','0','1'},
{'0','0','1','0'},
{'0','0','1','1'},
{'0','1','0','0'},
{'0','1','0','1'},
{'0','1','1','0'},
{'0','1','1','1'},
{'1','0','0','0'},
{'1','0','0','1'},
{'1','0','1','0'},
{'1','0','1','1'},
{'1','1','0','0'},
{'1','1','0','1'},
{'1','1','1','0'},
{'1','1','1','1'},
};
void compare()
{
while(encontro!=1)
{
for(k=0;k<=15;k++)
{
if(high[0]==s1[k][j] && high[1]==s1[k][j+1] && high[2]==s1[k][j+2] && high[3]==s1[k][j+3])
{
flag=k++;
encontro=1;
}
}
}
if(encontro==1)
{
lcd_putc("\f");
lcd_gotoxy(2,1);
printf(lcd_putc,"%u %u",flag,j);
switch(flag) {
case 1:
ai=-0.821;
aq=0.821;
break;
case 2:
ai=-0.821;
aq=0.22;
break;
case 3:
ai=-0.821;
aq=-0.22;
break;
case 4:
ai=-0.821;
aq=-0.821;
break;
case 5:
ai=-0.22;
aq=0.821;
break;
case 6:
ai=-0.22;
aq=0.22;
break;
case 7:
ai=-0.22;
aq=-0.22;
break;
case 8:
ai=-0.22;
aq=-0.821;
break;
case 9:
ai=0.22;
aq=0.821;
break;
case 10:
ai=0.22;
aq=0.22;
break;
case 11:
ai=0.22;
aq=-0.22;
break;
case 12:
ai=0.22;
aq=-0.821;
break;
case 13:
ai=0.821;
aq=0.821;
break;
case 14:
ai=0.821;
aq=0.22;
break;
case 15:
ai=0.821;
aq=-0.22;
break;
case 16:
ai=0.821;
aq=-0.821;
break;
default:
lcd_gotoxy(2,2);
printf(lcd_putc,"error");
delay_ms(2000);
}
}
flag=0;
encontro=0;
lcd_gotoxy(1,2);
printf(lcd_putc,"listo");
lcd_gotoxy(7,2);
printf(lcd_putc,"%f %f",ai,aq);
}
void leer()
{
read_adc(ADC_START_ONLY); //start the ADC reading
//Now wait for the ADC to complete
while (!adc_done()) ;
val=read_adc(ADC_READ_ONLY); //read the value
}
void main()
{
lcd_init();
setup_adc(ADC_CLOCK_DIV_32);// a 32 div , 73 useg en el multiplexor a 8div, 60 us
setup_adc_ports(RA0_RA1_RA3_ANALOG);
set_adc_channel(0);
lcd_putc("dato:");
do {
leer();
//val=rand();
delay_us(5); //ensure acquisition has completed
lcd_putc("\f");
lcd_gotoxy(1,1);
printf(lcd_putc,"%u",val);
for(x=0;x<=7;x++)
{
res=val%2;
aux=val/2;
val=aux;
if(res==0){
c[x]='0';
}
else
{
c[x]='1';
}
}
for(i=0;i<=3;i++)
{
low[i]=c[i];
}
for(m=0;m<=3;m++)
{
high[m]=c[m+4];
}
lcd_gotoxy(1,2);
printf(lcd_putc,"%s",low);
lcd_gotoxy(6,2);
for(y=0;y<=3;y++)
{
printf(lcd_putc,"%c",high[y]);
}
compare();
} while(TRUE);
}
|
and it works, but not so efficiently , not so good, I think there is something that I'm doing wrong |
|
|
Mike Walne
Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK
|
|
Posted: Wed Nov 07, 2012 3:39 pm |
|
|
I, for one, do not intend deciphering your code.
Please supply a SIMPLE explanation of the complete process.
What is your criterion for efficiency?
Is it speed, code size, simplicity, elegance, or what?
Mike |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19480
|
|
Posted: Wed Nov 07, 2012 3:56 pm |
|
|
Seriously, it is insane. flag at the end of the first part of the 'compare' routine, is just the numeric value you started with. The code looks to dismantle the binary value into individual bits, then look for patterns that are just the bit numbers, and give a result that was basically the original number. I'd say that everything in the main and compare, from the point the adc value is read, up to the switch, needs a maximum of one line. It shows a complete lack of understanding of what numbers 'are' and how they work... |
|
|
oxo
Joined: 13 Nov 2012 Posts: 219 Location: France
|
|
Posted: Tue Nov 13, 2012 8:32 am |
|
|
Ttelmah wrote: | Seriously, it is insane. flag at the end of the first part of the 'compare' routine, is just the numeric value you started with. The code looks to dismantle the binary value into individual bits, then look for patterns that are just the bit numbers, and give a result that was basically the original number. I'd say that everything in the main and compare, from the point the adc value is read, up to the switch, needs a maximum of one line. It shows a complete lack of understanding of what numbers 'are' and how they work... |
I agree. This sort of approach is found in those who don't understand the requirements before writing code. |
|
|
Mike Walne
Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK
|
|
Posted: Wed Nov 14, 2012 2:34 am |
|
|
There are now several of us telling you that you're going about it the wrong way.
I asked a while ago, please explain you're entire process from start to finish, in simple terms.
Then someone here will point you in a much better direction than the one you seem to have a fixation for.
BTW. Is this work, hobby, or school project?
Mike |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|