I finally got around to picking up the April 4th EET while eating supper tonight and who do I see on the inside of the front page? Rick's got an article on embedded.com on CIC filters. Cool stuff, man. The competition is tough around here! There is at least one thing in this article I don't get, Rick. You write: Although the gain of an Mth-order CIC decimation filter is (NR)M individual integrators can experience overflow. (Their gain is infinite at DC.) As such, the use of two's complement arithmetic resolves this overflow situation just so long as the integrator word width accommodates the maximum difference between any two successive samples (in other words, the difference causes no more than a single overflow). Using the two's complement binary format, with its modular wrap-around property, the follow-on comb filter will properly compute the correct difference between two successive integrator output samples. If you place the integrator before the comb, and the integrator overflows on one sample (without saturating), then how does that mean the comb filter will compute the right value? I know that an overflowed, two's complement accumulator will come back "in-range" as long as the final sum is in range of the accumulator, but if you're processing each output one-by-one, you're going to get a bogus value into the comb filter, no? What am I missing? -- % Randy Yates % "...the answer lies within your soul %% Fuquay-Varina, NC % 'cause no one knows which side %%% 919-577-9882 % the coin will fall." %%%% <yates@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO http://home.earthlink.net/~yatescr
Rick Lyons in EETimes!
Started by ●April 20, 2005
Reply by ●April 20, 20052005-04-20
On Wed, 20 Apr 2005 03:18:50 GMT, Randy Yates <yates@ieee.org> wrote: Hi Randy,>I finally got around to picking up the April 4th EET while eating >supper tonight and who do I see on the inside of the front page? >Rick's got an article on embedded.com on CIC filters. Cool stuff, >man. The competition is tough around here!EETimes! That's news to me! (I submitted that article to the Embedded Systems Programming magazine in May of 2004!) Is EETimes a hardcopy newspaper that you received in the mail? If so, I wonder how I can get my hands on a copy.>There is at least one thing in this article I don't get, Rick. You >write: > > Although the gain of an Mth-order CIC decimation filter is (NR)M > individual integrators can experience overflow. (Their gain is > infinite at DC.) As such, the use of two's complement arithmetic > resolves this overflow situation just so long as the integrator word > width accommodates the maximum difference between any two successive > samples (in other words, the difference causes no more than a single > overflow). Using the two's complement binary format, with its modular > wrap-around property, the follow-on comb filter will properly compute > the correct difference between two successive integrator output > samples. > >If you place the integrator before the comb, and the integrator overflows >on one sample (without saturating), then how does that mean the comb >filter will compute the right value? I know that an overflowed, two's >complement accumulator will come back "in-range" as long as the final >sum is in range of the accumulator, but if you're processing each output >one-by-one, you're going to get a bogus value into the comb filter, no? > >What am I missing?Ah, such a good question. Fred Harris' book has a neat example, with data sequence sample numerical values, of this whole process. I'll send ya' that example in the next day or two. See ya' Randy, [-Rick-]
Reply by ●April 20, 20052005-04-20
Rick Lyons wrote:> On Wed, 20 Apr 2005 03:18:50 GMT, Randy Yates <yates@ieee.org> wrote: > > Hi Randy, > > >>I finally got around to picking up the April 4th EET while eating >>supper tonight and who do I see on the inside of the front page? >>Rick's got an article on embedded.com on CIC filters. Cool stuff, >>man. The competition is tough around here! > > > EETimes! That's news to me! > (I submitted that article to the Embedded Systems Programming > magazine in May of 2004!) > Is EETimes a hardcopy newspaper that you received in the mail? > If so, I wonder how I can get my hands on a copy. >The article itself is on the embedded.com website, but EETimes is part of the CMP empire and they cross-reference one another.> > >>There is at least one thing in this article I don't get, Rick. You >>write: >> >> Although the gain of an Mth-order CIC decimation filter is (NR)M >> individual integrators can experience overflow. (Their gain is >> infinite at DC.) As such, the use of two's complement arithmetic >> resolves this overflow situation just so long as the integrator word >> width accommodates the maximum difference between any two successive >> samples (in other words, the difference causes no more than a single >> overflow). Using the two's complement binary format, with its modular >> wrap-around property, the follow-on comb filter will properly compute >> the correct difference between two successive integrator output >> samples. >> >>If you place the integrator before the comb, and the integrator overflows >>on one sample (without saturating), then how does that mean the comb >>filter will compute the right value? I know that an overflowed, two's >>complement accumulator will come back "in-range" as long as the final >>sum is in range of the accumulator, but if you're processing each output >>one-by-one, you're going to get a bogus value into the comb filter, no? >> >>What am I missing? > > > Ah, such a good question. > > Fred Harris' book has a neat example, with data > sequence sample numerical values, of this > whole process. I'll send ya' that example > in the next day or two. > > See ya' Randy, > [-Rick-] >-- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Reply by ●April 20, 20052005-04-20
"Rick Lyons" <R.Lyons@_BOGUS_ieee.org> wrote in message news:42669815.187395703@news.sf.sbcglobal.net...> On Wed, 20 Apr 2005 03:18:50 GMT, Randy Yates <yates@ieee.org> wrote: > > Hi Randy, > > >I finally got around to picking up the April 4th EET while eating > >supper tonight and who do I see on the inside of the front page? > >Rick's got an article on embedded.com on CIC filters. Cool stuff, > >man. The competition is tough around here! > > EETimes! That's news to me! > (I submitted that article to the Embedded Systems Programming > magazine in May of 2004!) > Is EETimes a hardcopy newspaper that you received in the mail? > If so, I wonder how I can get my hands on a copy.It's a tabloid-sized hardcopy magazine. Tends to focus more on news than "how to" articles. I think you can sign up for a free subscription, similar to with EDN, Electronic Design, etc.. http://www.eetimes.com/ https://www.subscribeeetimes.com/ (I couldn't find your article on-line, but I didn't look too hard.)
Reply by ●April 20, 20052005-04-20
Randy Yates <yates@ieee.org> wrote:>I finally got around to picking up the April 4th EET while eating >supper tonight and who do I see on the inside of the front page? >Rick's got an article on embedded.com on CIC filters. Cool stuff, >man. The competition is tough around here!Yep, just making my way through the article now. Good job Rick. Robert www.numbersusa.com www.americanpatrol.com
Reply by ●April 21, 20052005-04-21
On Wed, 20 Apr 2005 14:18:44 -0700, "Jon Harris" <jon_harrisTIGER@hotmail.com> wrote:>"Rick Lyons" <R.Lyons@_BOGUS_ieee.org> wrote in message >news:42669815.187395703@news.sf.sbcglobal.net... >> On Wed, 20 Apr 2005 03:18:50 GMT, Randy Yates <yates@ieee.org> wrote: >> >> Hi Randy, >> >> >I finally got around to picking up the April 4th EET while eating >> >supper tonight and who do I see on the inside of the front page? >> >Rick's got an article on embedded.com on CIC filters. Cool stuff, >> >man. The competition is tough around here! >> >> EETimes! That's news to me! >> (I submitted that article to the Embedded Systems Programming >> magazine in May of 2004!) >> Is EETimes a hardcopy newspaper that you received in the mail? >> If so, I wonder how I can get my hands on a copy. > >It's a tabloid-sized hardcopy magazine. Tends to focus more on news than "how >to" articles. I think you can sign up for a free subscription, similar to with >EDN, Electronic Design, etc.. >http://www.eetimes.com/ >https://www.subscribeeetimes.com/ > >(I couldn't find your article on-line, but I didn't look too hard.) >Hi Jon, thanks to Randy I found the article at: http://www.eetimes.com/showArticle.jhtml?articleID=160400592 See ya', [-Rick-]
Reply by ●April 21, 20052005-04-21
On Wed, 20 Apr 2005 18:18:49 -0500, r_obert@REMOVE_THIS.hotmail.com wrote:>Randy Yates <yates@ieee.org> wrote: > >>I finally got around to picking up the April 4th EET while eating >>supper tonight and who do I see on the inside of the front page? >>Rick's got an article on embedded.com on CIC filters. Cool stuff, >>man. The competition is tough around here! > >Yep, just making my way through the article now. Good job Rick. > >RobertHi Robert, as Elvis used to say, "Thang yah very mush." (However, Elvis may still be saying that because, while waiting in the grocery store line today, I saw a newspaper that said Elvis is still alive!!! Can it be???) Actually, I've gotta say that our DSP pal Ray Andraka helped me a heck of a lot in understanding (whatever little understanding I have of) these darned CIC filters. One of the most surprising things that I've noticed in DSP is how many complicated issues there are with regard to implementing those simple CIC filters. There's SOooo much to think about! By the way, where is Ray Andraka? If he won the Lottery and is now drinking red wine & living in Southern France, I'm gonna be super ticked off!! See ya Robert, [-Rick-] I heard someone say today, on the radio, that the new Pope is, "God's Rothweiler." Sheece!
Reply by ●April 21, 20052005-04-21
Rick Lyons wrote:> I heard someone say today, on the radio, that the new > Pope is, "God's Rothweiler."I liked the joke that played on NPR about him, and how persuasive he could be: Ratzinger and two other clergy were standing at the Pearly Gates, and one by one are taken in to meet their Maker. The first one goes in, and comes out crying and saying "How could I have been so wrong!". The second one goes in, and comes out crying and saying "How could I have been so wrong!". Ratzinger goes in, and God comes out crying and saying "How could I have been so wrong!". :0) Ciao, Peter K.
Reply by ●April 21, 20052005-04-21
R.Lyons@_BOGUS_ieee.org (Rick Lyons) writes:> On Wed, 20 Apr 2005 03:18:50 GMT, Randy Yates <yates@ieee.org> wrote: > > Hi Randy, > > >I finally got around to picking up the April 4th EET while eating > >supper tonight and who do I see on the inside of the front page? > >Rick's got an article on embedded.com on CIC filters. Cool stuff, > >man. The competition is tough around here! > > EETimes! That's news to me! > (I submitted that article to the Embedded Systems Programming > magazine in May of 2004!) > Is EETimes a hardcopy newspaper that you received in the mail?Yes. Electronic Engineering Times. You mean you've never heard of it? What rock have you been hiding under? How could a Silicon Valley engineer of, what, 30 years never have heard of EETimes? That's like a 12-year-old that's never heard of McDonalds!> If so, I wonder how I can get my hands on a copy.You're welcome to mine if you want it.> >There is at least one thing in this article I don't get, Rick. You > >write: > > > > Although the gain of an Mth-order CIC decimation filter is (NR)M > > individual integrators can experience overflow. (Their gain is > > infinite at DC.) As such, the use of two's complement arithmetic > > resolves this overflow situation just so long as the integrator word > > width accommodates the maximum difference between any two successive > > samples (in other words, the difference causes no more than a single > > overflow). Using the two's complement binary format, with its modular > > wrap-around property, the follow-on comb filter will properly compute > > the correct difference between two successive integrator output > > samples. > > > >If you place the integrator before the comb, and the integrator overflows > >on one sample (without saturating), then how does that mean the comb > >filter will compute the right value? I know that an overflowed, two's > >complement accumulator will come back "in-range" as long as the final > >sum is in range of the accumulator, but if you're processing each output > >one-by-one, you're going to get a bogus value into the comb filter, no? > > > >What am I missing? > > Ah, such a good question. > > Fred Harris' book has a neat example, with data > sequence sample numerical values, of this > whole process. I'll send ya' that example > in the next day or two.I got your email Rick, thanks. Let me chew on it. -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA randy.yates@sonyericsson.com, 919-472-1124
Reply by ●April 21, 20052005-04-21
"Rick Lyons" <R.Lyons@_BOGUS_ieee.org> wrote in message news:42671843.220209515@news.sf.sbcglobal.net... <Snip>> > By the way, where is Ray Andraka? If he won the > Lottery and is now drinking red wine & living in Southern > France, I'm gonna be super ticked off!!From what I know, he's waay too busy with various jobs as well as trying to publish a book at the same time. Rick - I'm sure you could give him a few tips in this dept (namely - if he is (or not) going to move to Southern France and sip on red wine off his book royalties) :-) Cheers Bhaskar
