Convolutional or trellis codes work by forcing specific sequences of state transitions and detecting errors when those transitions don't happen. Once the decoder makes an error it could take several symbols get get back into synchronization with the transmitter. If you mean the systematic (feedback free) form of a convolutional encoder it has the same limitations as the form using feedback as they generate the same encoded output. If you mean a block code like LDPC, Golay or Walsh codes, errors don't propogate beyond the block.
>From what I've read, turbo codes get very close to channel capacity but they >require very long block lengths (greater than 10,000 symbols) and an iterative >decoder to be effective. 73, John KD6OZH ----- Original Message ----- From: Rud Merriam To: digitalradio@yahoogroups.com Sent: Wednesday, November 07, 2007 02:24 UTC Subject: RE: [digitalradio] Testing Digital Codes at Bit Level In addition, if you have a systematic code versus convolutional or trellis encoding is the bit flipping not applicable? One of my other research activities is on Low Density Parity Codes since they approach channel capacity better than other codes. Rud Merriam K5RUD ARES AEC Montgomery County, TX http://TheHamNetwork.net -----Original Message----- From: digitalradio@yahoogroups.com [mailto:[EMAIL PROTECTED] On Behalf Of John B. Stephensen Sent: Tuesday, November 06, 2007 2:30 PM To: digitalradio@yahoogroups.com Subject: Re: [digitalradio] Testing Digital Codes at Bit Level If you have an inner and outer code that would be the situation, but I'm not sure that flipping one bit would always be accurate. A Viterbi decoder might generate small bursts of errors. HDTV uses TCM with an outer Reed-Solomon code. Even though there are 12 interleaved convolutional encoders, they still use an RS code that is capable of correcting bursts of 8 errors. 73, John KD6OZH ----- Original Message ----- From: Rud Merriam To: digitalradio@yahoogroups.com Sent: Tuesday, November 06, 2007 20:10 UTC Subject: RE: [digitalradio] Testing Digital Codes at Bit Level I understand about the use of soft decoders. If the protocol uses a soft decoder and another hard decoder the latter works at the bit level. A standard example is using Reed-Solomon for the hard decoder. Would the bit flipping be representative of the atmospheric effects for the outer hard decoder, e.g. RS decoder? Rud Merriam K5RUD ARES AEC Montgomery County, TX http://TheHamNetwork.net
