Re: error correcting codes
Hi Resistance
>the reason i think dmin =1 is difficult to correct is wat u told plus the syndrome >for one bit error is the same as the property of LBC goes right?
I dont know what LBC is.
But when dmin = 1 there is no detection or correction. If just one bit changes the new word will be valid and the error goes undetected.
dmin is selected by first defining how many bit errors you want to be able to detect and correct. Obviuosly the larger is better. But to increase dmin you should increase the space of possible words or decrease the number of valid words of the code.
In other words you have to add redundancy to the data. All systems of error correction add redundancy to the data by adding more bits than those needed to just send the desired message.
In a convolutional code (ratio 1/2) you are sending 2 bits for each bit of the message. That is you are doubling the req. bandwidth.
In a memory ECC 7 bits can protect 32 bits or 8 bits can protect 64 bits
So there is always a tradeoff between the error correcting capability and the redundancy you add.
The choice of an ECC depends on the application.
There are Block and convolutional (sequential) ECCs. And in each category you have many choices. Sometimes more than one is used at the same time. It is said that they are "concatenated". As an example in a modem you may find a block ECC as Reed Solomon, followed by a convolutional encoder. Which are widely used.
Check this page:
http://www.eccpage.com/
There are some books that you can upload from edaboard.
Including "Introduction to error correcting codes " by Michael Purser