On Thu, 1 Oct 2009 01:23:07 +0000 (UTC), spope33@speedymail.org (Steve Pope) wrote:>Muzaffer Kal <kal@dspia.com> wrote: > >>On Wed, 30 Sep 2009 23:14:34 +0000 (UTC), spope33@speedymail.org > >>>Here's an example. Suppose all errors occur in three-bit bursts. > >>The question is how can one suppose such a situation. If you know the >>channel to that degree then you design a encoder/decoder which takes >>advantage of that information so you remove that certainty and in the >>end you're left with an iid source and channel with awgn which is >>where you have started and which is the main problem to solve. > >Except that the three-bit-burst channel has more capacity >than a random-error channel with the same bit-error rate. >You can do better than just randomizing it and treating it >as if it were random. >Which is exactly what I said "If you know the channel to that degree then you design a encoder/decoder which takes advantage of that information so you remove that certainty">A similar thread came up here not long ago, in which it was >discussed that an AWGN channel is the worst possible channel.This is certainly true and I already stated it " in the end you're left with an iid source and channel with awgn which is where you have started and which is the main problem to solve". The bottom line is when ever you know that your channel is anything other than iid + awgn you can design an encoder/decoder to your advantage but you're in the end left with the basic problem with some higher SNR at the decision point. -- Muzaffer Kal DSPIA INC. ASIC/FPGA Design Services http://www.dspia.com
Reed Solomon vs Convolutional coding
Started by ●September 30, 2009
Reply by ●October 1, 20092009-10-01
Reply by ●October 1, 20092009-10-01
On 9/30/2009 6:46 PM, Jerry Avins wrote:> JAlbertoDJ wrote: >>> JAlbertoDJ wrote: >>>>> cpshah99 <cpshah99@rediffmail.com> wrote: >>>>> >>>>>> Sorry to say this but can you please explain why with interleaving >> the >>>>>> information is lost.? >>>>> Here's an example. Suppose all errors occur in three-bit bursts. >>>>> Suppose you apply a pseudo-random interleaver. In the interleaved >>>>> stream, the ability to predict bit errors based on the previous >>>>> bit is lost. >>>>> >>>>> You could assert of course that if you know the interleaver >>>>> pattern, no information is lost but this may not help you out >>>>> practically. >>>>> >>>>> Steve >>>>> >>>> But, you cannot apply a pseudo-random interleaver, you should apply a >>>> interleaver type fodney, for example. So you separate the symbols in >> time >>>> transforming a channel with memory to a memoryless one, and thereby >> enables >>>> the random-error-correcting codes to be useful in a burns-noise >> channel. >>> You keep writing burns noise. Do you mean burst noise? >> >> Means High levels of noise during a time interval. > > Can you cite a reference that uses it that way? > > JerryStill sounds like he means burst noise to me. I've never heard of burns-noise, either. -- Eric Jacobsen Minister of Algorithms Abineau Communications http://www.abineau.com
Reply by ●October 1, 20092009-10-01
Eric Jacobsen <eric.jacobsen@ieee.org> wrote:>Still sounds like he means burst noise to me. I've never heard of >burns-noise, either.Sounds like long-winded Scottish poetry. (GD&R) Steve
Reply by ●October 1, 20092009-10-01
>On 9/30/2009 6:46 PM, Jerry Avins wrote: >> JAlbertoDJ wrote: >>>> JAlbertoDJ wrote: >>>>>> cpshah99 <cpshah99@rediffmail.com> wrote: >>>>>> >>>>>>> Sorry to say this but can you please explain why withinterleaving>>> the >>>>>>> information is lost.? >>>>>> Here's an example. Suppose all errors occur in three-bit bursts. >>>>>> Suppose you apply a pseudo-random interleaver. In the interleaved >>>>>> stream, the ability to predict bit errors based on the previous >>>>>> bit is lost. >>>>>> >>>>>> You could assert of course that if you know the interleaver >>>>>> pattern, no information is lost but this may not help you out >>>>>> practically. >>>>>> >>>>>> Steve >>>>>> >>>>> But, you cannot apply a pseudo-random interleaver, you should applya>>>>> interleaver type fodney, for example. So you separate the symbolsin>>> time >>>>> transforming a channel with memory to a memoryless one, and thereby >>> enables >>>>> the random-error-correcting codes to be useful in a burns-noise >>> channel. >>>> You keep writing burns noise. Do you mean burst noise? >>> >>> Means High levels of noise during a time interval. >> >> Can you cite a reference that uses it that way? >> >> Jerry > >Still sounds like he means burst noise to me. I've never heard of >burns-noise, either. > >-- >Eric Jacobsen >Minister of Algorithms >Abineau Communications >http://www.abineau.com >Excuse me, i want say burst-noise. I have some problems with english.
Reply by ●October 1, 20092009-10-01
Eric Jacobsen wrote:> On 9/30/2009 6:46 PM, Jerry Avins wrote: >> JAlbertoDJ wrote: >>>> JAlbertoDJ wrote: >>>>>> cpshah99 <cpshah99@rediffmail.com> wrote: >>>>>> >>>>>>> Sorry to say this but can you please explain why with interleaving >>> the >>>>>>> information is lost.? >>>>>> Here's an example. Suppose all errors occur in three-bit bursts. >>>>>> Suppose you apply a pseudo-random interleaver. In the interleaved >>>>>> stream, the ability to predict bit errors based on the previous >>>>>> bit is lost. >>>>>> >>>>>> You could assert of course that if you know the interleaver >>>>>> pattern, no information is lost but this may not help you out >>>>>> practically. >>>>>> >>>>>> Steve >>>>>> >>>>> But, you cannot apply a pseudo-random interleaver, you should apply a >>>>> interleaver type fodney, for example. So you separate the symbols in >>> time >>>>> transforming a channel with memory to a memoryless one, and thereby >>> enables >>>>> the random-error-correcting codes to be useful in a burns-noise >>> channel. >>>> You keep writing burns noise. Do you mean burst noise? >>> >>> Means High levels of noise during a time interval. >> >> Can you cite a reference that uses it that way? >> >> Jerry > > Still sounds like he means burst noise to me. I've never heard of > burns-noise, either.In a masters-level course I took at Rutgers, the professor -- yes, a full professor -- persisted in talking about casual circuits. He insisted that causal wasn't a word. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 1, 20092009-10-01
JAlbertoDJ wrote: ...> Excuse me, i want say burst-noise. I have some problems with english.No problem. I just wanted to be clear. Many people say "nucular" when they mean "nuclear". It's funny, though. Those same people don't sat "uncular" when they mean "unclear", even though the only difference is the order of the first two letters. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 1, 20092009-10-01
On 2009-10-01 13:41:45 -0300, Jerry Avins <jya@ieee.org> said:> Eric Jacobsen wrote: >> On 9/30/2009 6:46 PM, Jerry Avins wrote: >>> JAlbertoDJ wrote: >>>>> JAlbertoDJ wrote: >>>>>>> cpshah99 <cpshah99@rediffmail.com> wrote: >>>>>>> >>>>>>>> Sorry to say this but can you please explain why with interleaving >>>> the >>>>>>>> information is lost.? >>>>>>> Here's an example. Suppose all errors occur in three-bit bursts. >>>>>>> Suppose you apply a pseudo-random interleaver. In the interleaved >>>>>>> stream, the ability to predict bit errors based on the previous >>>>>>> bit is lost. >>>>>>> >>>>>>> You could assert of course that if you know the interleaver >>>>>>> pattern, no information is lost but this may not help you out >>>>>>> practically. >>>>>>> >>>>>>> Steve >>>>>>> >>>>>> But, you cannot apply a pseudo-random interleaver, you should apply a >>>>>> interleaver type fodney, for example. So you separate the symbols in >>>> time >>>>>> transforming a channel with memory to a memoryless one, and thereby >>>> enables >>>>>> the random-error-correcting codes to be useful in a burns-noise >>>> channel. >>>>> You keep writing burns noise. Do you mean burst noise? >>>> >>>> Means High levels of noise during a time interval. >>> >>> Can you cite a reference that uses it that way? >>> >>> Jerry >> >> Still sounds like he means burst noise to me. I've never heard of >> burns-noise, either. > > In a masters-level course I took at Rutgers, the professor -- yes, a > full professor -- persisted in talking about casual circuits. He > insisted that causal wasn't a word. > > JerryHe must have been a colleague of another one I heard of when I was a summer student. It seems the speaker/researcher has used "miserable" functions as part of his analysis. Most other folks used measureable functions so had an easier time in obtaining their results. ;-) The chap telling the story had other comments on the heavy accent of the speaker.
Reply by ●October 1, 20092009-10-01
On Oct 1, 11:46=A0am, Jerry Avins <j...@ieee.org> wrote:> Many people say "nucular" when they mean "nuclear". It's funny, though. > Those same people don't sat "uncular" when they mean "unclear", even > though the only difference is the order of the first two letters.In the mid-1930s, my grandfather, who was a government official in India, was told by the government, ".................... is making a huge donation to establish a new university, and we need a complete set of documents: University Constitution, Bylaws, Degree Requirements, Curricula, Course Syllabi, etc. in two weeks' time." Faced with an impossible task, he arranged for a bunch of typists to re-type the corresponding documents for Nagpur University, and merely substitute "Saugor University" wherever a document said "Nagpur University", and this got the job done in a hurry as was needed by the government. For about twenty years after its founding, Saugor University had on its books, and on the transcripts/mark-sheets issued, a graduate course titled Unclear Physics.... None of the faculty noticed, and all of the students thought the course title described the course perfectly. Dilip Sarwate
Reply by ●October 1, 20092009-10-01
dvsarwate@yahoo.com wrote: ...> For about twenty years after its founding, Saugor University had > on its books, and on the transcripts/mark-sheets issued, a > graduate course titled Unclear Physics. ...:-) Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
