Reply by Vladimir Vassilevsky●December 31, 20092009-12-31
fsm12 wrote:
> HI,
>
> I want to know about error correcting capability of a convolution code
> i.e. how many errors can be corrected in a continuous input data pattern
> (e.g. a 1/2 rate code with constraint length of 7 has free distance of 10
> and is said to correct 4 errors, whether these errors cab be burst type
> also and what will be the relation of this error correction to input
> data)?
D = 2t + 1
VLV
Reply by Tim Wescott●December 31, 20092009-12-31
On Thu, 31 Dec 2009 08:58:07 -0600, fsm12 wrote:
> HI,
>
> I want to know about error correcting capability of a convolution
> code
> i.e. how many errors can be corrected in a continuous input data
> pattern (e.g. a 1/2 rate code with constraint length of 7 has free
> distance of 10 and is said to correct 4 errors, whether these errors cab
> be burst type also and what will be the relation of this error
> correction to input data)?
>
> Thanks
Your question isn't clear. Viterbi decoders (or any other decoders, for
convolutional or block codes) have a maximum-length burst that they can
correct for (I think that's your "4 errors"), and in general an error
rate that they can correct for.
Given a long enough burst, any decoder will fall down on the job. Most
practical channels tend to suffer from burst errors, so most practical
FEC schemes -- convolutional or block -- reorder the encoded data. Then
when the data is put back into the correct order at the receiver a burst
error is turned into a bunch of individual bit errors which are easier
for the decoder to deal with.
--
www.wescottdesign.com
Reply by fsm12●December 31, 20092009-12-31
HI,
I want to know about error correcting capability of a convolution code
i.e. how many errors can be corrected in a continuous input data pattern
(e.g. a 1/2 rate code with constraint length of 7 has free distance of 10
and is said to correct 4 errors, whether these errors cab be burst type
also and what will be the relation of this error correction to input
data)?
Thanks