Protecting 3x16 bits with 16 bits ?
|
Prev: Call for Papers: The 2010 International Conference on Computer Design (CDES'10), USA, July 2010 (updated)
Next: Protecting 3x16 bits with 16 bits ? (Raptor codes)
From: Ian Bell on 13 Feb 2010 10:12 mpm wrote: > On Feb 12, 9:17 pm, "m...(a)sushi.com" <m...(a)sushi.com> wrote: >> On Feb 12, 8:13 am, "Skybuck Flying" <IntoTheFut...(a)hotmail.com> >> wrote: >> >> >> >> >> >>> "Dave -Turner" <ad...(a)127.0.0.1> wrote in message >>> news:suOdnXyGC-Hf3-jWnZ2dnUVZ7v-dnZ2d(a)westnet.com.au... >>>> so use crc16 >>> I will try... >>> However I still need an error correcting code for mode 3 (See new >>> posting)... >>> Any idea's ? >>> I have 0% experience with error correcting codes I am afraid ! ;) :) >>> Bye, >>> Skybuck. >> Have you considered paying a consultant? >> >> I took a class in grad school by some NASA/Ames Phd on error detection >> and correct. It had to be the class with the most people dropping it >> that I ever took. When the dust settled, he gave out one A, one B >> (me), and a lot of Cs. He figured he didn't have to use a curve since >> so many people dropped it, so he spared D and Fs. >> >> Anyway, my point is this a specialty where it would pay to have some >> expert do the coding for you.- Hide quoted text - >> >> - Show quoted text - > > Trust me (and everyone else here). > If Skybuck were to ever pay for a specialist, that person should have > nothing to do with electronics!! > The first thing you need to do is decide: 1. What type of errors are likely to occur - single or multiple bit errors in a single packet of 4x16 bits or bursts in a stream of packets. 2. Whether to reformat groups of packets so as to spread bursts of errors amongst a number of packets. 3. How many bits in error do you want to be able to detect and how many of these do you want to be able to correct Only then can you begin to determine a suitable coding scheme. Cheers Ian
From: Skybuck Flying on 13 Feb 2010 11:29 "Ian Bell" <ruffrecords(a)yahoo.com> wrote in message news:hl6fh3$k5i$1(a)localhost.localdomain... > mpm wrote: >> On Feb 12, 9:17 pm, "m...(a)sushi.com" <m...(a)sushi.com> wrote: >>> On Feb 12, 8:13 am, "Skybuck Flying" <IntoTheFut...(a)hotmail.com> >>> wrote: >>> >>> >>> >>> >>> >>>> "Dave -Turner" <ad...(a)127.0.0.1> wrote in message >>>> news:suOdnXyGC-Hf3-jWnZ2dnUVZ7v-dnZ2d(a)westnet.com.au... >>>>> so use crc16 >>>> I will try... >>>> However I still need an error correcting code for mode 3 (See new >>>> posting)... >>>> Any idea's ? >>>> I have 0% experience with error correcting codes I am afraid ! ;) :) >>>> Bye, >>>> Skybuck. >>> Have you considered paying a consultant? >>> >>> I took a class in grad school by some NASA/Ames Phd on error detection >>> and correct. It had to be the class with the most people dropping it >>> that I ever took. When the dust settled, he gave out one A, one B >>> (me), and a lot of Cs. He figured he didn't have to use a curve since >>> so many people dropped it, so he spared D and Fs. >>> >>> Anyway, my point is this a specialty where it would pay to have some >>> expert do the coding for you.- Hide quoted text - >>> >>> - Show quoted text - >> >> Trust me (and everyone else here). >> If Skybuck were to ever pay for a specialist, that person should have >> nothing to do with electronics!! >> > > > The first thing you need to do is decide: > > 1. What type of errors are likely to occur - single or multiple bit errors > in a single packet of 4x16 bits or bursts in a stream of packets. > > 2. Whether to reformat groups of packets so as to spread bursts of errors > amongst a number of packets. > > 3. How many bits in error do you want to be able to detect and how many of > these do you want to be able to correct > > Only then can you begin to determine a suitable coding scheme. Not really lol ! ;) :) One first has to detect the errors ! LOL. Bye, Skybuck ;) :) P.S.: Sorry to burst your bubble ! ;) :)
From: Ian Bell on 13 Feb 2010 18:39 Skybuck Flying wrote: > "Ian Bell" <ruffrecords(a)yahoo.com> wrote in message > news:hl6fh3$k5i$1(a)localhost.localdomain... >> mpm wrote: >>> On Feb 12, 9:17 pm, "m...(a)sushi.com" <m...(a)sushi.com> wrote: >>>> On Feb 12, 8:13 am, "Skybuck Flying" <IntoTheFut...(a)hotmail.com> >>>> wrote: >>>> >>>> >>>> >>>> >>>> >>>>> "Dave -Turner" <ad...(a)127.0.0.1> wrote in message >>>>> news:suOdnXyGC-Hf3-jWnZ2dnUVZ7v-dnZ2d(a)westnet.com.au... >>>>>> so use crc16 >>>>> I will try... >>>>> However I still need an error correcting code for mode 3 (See new >>>>> posting)... >>>>> Any idea's ? >>>>> I have 0% experience with error correcting codes I am afraid ! ;) :) >>>>> Bye, >>>>> Skybuck. >>>> Have you considered paying a consultant? >>>> >>>> I took a class in grad school by some NASA/Ames Phd on error detection >>>> and correct. It had to be the class with the most people dropping it >>>> that I ever took. When the dust settled, he gave out one A, one B >>>> (me), and a lot of Cs. He figured he didn't have to use a curve since >>>> so many people dropped it, so he spared D and Fs. >>>> >>>> Anyway, my point is this a specialty where it would pay to have some >>>> expert do the coding for you.- Hide quoted text - >>>> >>>> - Show quoted text - >>> Trust me (and everyone else here). >>> If Skybuck were to ever pay for a specialist, that person should have >>> nothing to do with electronics!! >>> >> >> The first thing you need to do is decide: >> >> 1. What type of errors are likely to occur - single or multiple bit errors >> in a single packet of 4x16 bits or bursts in a stream of packets. >> >> 2. Whether to reformat groups of packets so as to spread bursts of errors >> amongst a number of packets. >> >> 3. How many bits in error do you want to be able to detect and how many of >> these do you want to be able to correct >> >> Only then can you begin to determine a suitable coding scheme. > > Not really lol ! ;) :) > > One first has to detect the errors ! LOL. > > Bye, > Skybuck ;) :) > > P.S.: Sorry to burst your bubble ! ;) :) > > YAWN
From: Michael A. Terrell on 13 Feb 2010 18:51 "miso(a)sushi.com" wrote: > > On Feb 12, 8:13 am, "Skybuck Flying" <IntoTheFut...(a)hotmail.com> > wrote: > > "Dave -Turner" <ad...(a)127.0.0.1> wrote in message > > > > news:suOdnXyGC-Hf3-jWnZ2dnUVZ7v-dnZ2d(a)westnet.com.au... > > > > > so use crc16 > > > > I will try... > > > > However I still need an error correcting code for mode 3 (See new > > posting)... > > > > Any idea's ? > > > > I have 0% experience with error correcting codes I am afraid ! ;) :) > > > > Bye, > > Skybuck. > > Have you considered paying a consultant? > > I took a class in grad school by some NASA/Ames Phd on error detection > and correct. It had to be the class with the most people dropping it > that I ever took. When the dust settled, he gave out one A, one B > (me), and a lot of Cs. He figured he didn't have to use a curve since > so many people dropped it, so he spared D and Fs. > > Anyway, my point is this a specialty where it would pay to have some > expert do the coding for you. If his parents used error correcting code, Skyduck would have never been conceived. -- Greed is the root of all eBay.
From: miso on 14 Feb 2010 00:10
On Feb 12, 8:47 pm, "Skybuck Flying" <IntoTheFut...(a)hotmail.com> wrote: > Coding the hamming code thing... I think I can manage that ;) It might take > some time but that's ok :) > > LDPC codes is a different matter though ;) > > For now it seems LDPC has large overhead ?! so that's kinda shitty... but do > correct me if I am wrong ;) :) > > If the overhead ain't that bad I might be willing to help some "scientist" > with coding up something "good" in delphi ;) :) You do the thinking, I do > the programming =D > > When it comes to saving time maybe some people could code for me but I doubt > there are any good "Delphi" codes out there to work for absolutely "budget" > hours ;) :) > > And I am very demanding when it comes to Delphi code quality ! ;) :) It most > be top notch ! ;) :) =D > > And that's exactly what's gonna happen: > > 1. First generic Delphi code. > 2. Then maybe optimized Delphi code for this specific case. > 3. Then gpu code. > 4. Then maybe further optimized gpu code if possible ;) :) > > However shader model 3.0 /cg doesn't seem to good at bit fiddling... so that > might get a bit tricky ;) :) > > Also error correction is something I wanted to do for some time now ;) :) > > Thanks to better documentation explaining the idea's/concepts I should be > able to program it now and ofcourse thanks to my increased programming skill > level ;) :) > > Bye, > Skybuck ;) :) All the coding I studied was not shortened, so you could only pad the extra bits. This is complicated enough that to do the job justice, you would need to take a few courses. You would probably have to take one class in discrete systems so you understand the Galois math. Then maybe 2 classes error detection and correction theory. It's far easier to pay a professional. This is a specialty. Even consultants use consultants. |