Forums

Why there is an extra error when I try to correct more than the capacity?

Started by kittuis4u February 9, 2007
Hi I am Krishna Chaitanya. I am trying to implement reed-solomon codes and
check what is the performance of the coding scheme. 

I am using n=31 and k=29. This means that it has to correct 2 symbols(5
bits each). I would like to know few things about the error correction
codes. 

When ever I induce more than 1 symbol error, that is lets say I introduce
and error of 7 bits, now the error correction code has to correct the 1
symbol that it has to and then leave the remaining errors right? 

But when I try to check this it not only gives all the 7 bits as error but
some times it also increases the number of errors.

Is there any specific reason for that. Please let me know about this. 

Thanks and Regards,
Krishna Chaitanya.



kittuis4u wrote:

> Hi I am Krishna Chaitanya. I am trying to implement reed-solomon codes and > check what is the performance of the coding scheme. > I am using n=31 and k=29. This means that it has to correct 2 symbols(5 > bits each).
???? (31,29) is a Hamming code in GF(2^5). It can correct for one symbol only. I would like to know few things about the error correction
> codes.
R. Blahut "The theory and practice of the error control codes"
> When ever I induce more than 1 symbol error, that is lets say I introduce > and error of 7 bits, now the error correction code has to correct the 1 > symbol that it has to and then leave the remaining errors right? > But when I try to check this it not only gives all the 7 bits as error but > some times it also increases the number of errors. > > Is there any specific reason for that. Please let me know about this.
No wonder. If there are more errors then the particular code can correct, then the attempt to correct the errors can only make the things worse. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
> > >kittuis4u wrote: > >> Hi I am Krishna Chaitanya. I am trying to implement reed-solomon codes
and
>> check what is the performance of the coding scheme. >> I am using n=31 and k=29. This means that it has to correct 2
symbols(5
>> bits each). > >???? >(31,29) is a Hamming code in GF(2^5). It can correct for one symbol
only.
> > > I would like to know few things about the error correction >> codes. > >R. Blahut "The theory and practice of the error control codes" > >> When ever I induce more than 1 symbol error, that is lets say I
introduce
>> and error of 7 bits, now the error correction code has to correct the
1
>> symbol that it has to and then leave the remaining errors right? >> But when I try to check this it not only gives all the 7 bits as error
but
>> some times it also increases the number of errors. >> >> Is there any specific reason for that. Please let me know about this. > >No wonder. If there are more errors then the particular code can >correct, then the attempt to correct the errors can only make the things
>worse.
I also tried to implement a cyclic code with very high values of (n,k). When I tried that and check if the code is able to correct all the 2 bit errors, it was not able to do that. The minimum distance for the code is 6 but it was not able to correct all the 2 bit errors. Could you let me know why this is happening. the (n,k)=(117,102) in my code.

kittuis4u wrote:


> I also tried to implement a cyclic code with very high values of (n,k). > When I tried that and check if the code is able to correct all the 2 bit > errors, it was not able to do that. The minimum distance for the code is 6 > but it was not able to correct all the 2 bit errors. Could you let me know > why this is happening. the (n,k)=(117,102) in my code.
You should do your homework yourself. VLV
> >
On 2007-02-09, kittuis4u <kittuis4u@gmail.com> wrote:
> > I am using n=31 and k=29. This means that it has to correct 2 symbols(5 > bits each). I would like to know few things about the error correction > codes.
Are you confusing errors and erasures? -- Ben Jackson AD7GD <ben@ben.com> http://www.ben.com/