Jerry Avins wrote:> Richard Owlett wrote: > > ... > >> Question for all: >> Is there somewhere I can post my 100 point and 1000 point plots ( jpg >> files)? > > > Unless I missed something useful, alt.binaries seems to be entirely > given over to porn. Not pictures, but lists of web sites. I suppose you > can post there anyway, like old times. You can email them to me; I'll > give them a URL of their own. > > JerryThere's an alt.binaries.electronic.schematic group, who's charter should be obvious from the name. How difficult would it be to set up an alt.binaries.dsp group (and hope that DSP doesn't stand for some darned sick perversion that we don't know about)? I know nothing of the inner workings of newsgroups, either technical or political, by the way. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Any DSPers using Scilab [ NEWBIE in trouble ; ]
Started by ●September 4, 2004
Reply by ●September 5, 20042004-09-05
Reply by ●September 5, 20042004-09-05
Richard Owlett wrote:> Question for all: > Is there somewhere I can post my 100 point and 1000 point plots ( jpg > files)?Here's a free image hosting site you can use: http://www.imageshack.us/index.php Images will be deleted if they're not accessed for a year, but obviously that's not a problem for this sort of short-term usage. Ben
Reply by ●September 5, 20042004-09-05
Tim Wescott wrote:> Jerry Avins wrote: > >> Richard Owlett wrote: >> >> ... >> >>> Question for all: >>> Is there somewhere I can post my 100 point and 1000 point plots ( jpg >>> files)? >> >> >> >> Unless I missed something useful, alt.binaries seems to be entirely >> given over to porn. Not pictures, but lists of web sites. I suppose you >> can post there anyway, like old times. You can email them to me; I'll >> give them a URL of their own. >> >> Jerry > > > There's an alt.binaries.electronic.schematic group, who's charter should > be obvious from the name. How difficult would it be to set up an > alt.binaries.dsp group (and hope that DSP doesn't stand for some darned > sick perversion that we don't know about)? I know nothing of the inner > workings of newsgroups, either technical or political, by the way.Tim, I like your subtlety with Darned Sick Perversion. I didn't find alt.binaries.electronic.schematic tracing down Google's tree, so maybe there are more that I missed. It could be that the darned sick perversions caught my attention to the exclusion of everything else, or maybe there's another tree. I'll look again. I don't know about establishing a group any more than you, but if we need one, we can find out. I hope we can just piggyback on something already there. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●September 5, 20042004-09-05
Jerry Avins wrote:> Tim Wescott wrote: > >> Jerry Avins wrote: >> >>> Richard Owlett wrote: >>> >>> ... >>> >>>> Question for all: >>>> Is there somewhere I can post my 100 point and 1000 point plots ( >>>> jpg files)? >>> >>> >>> >>> >>> Unless I missed something useful, alt.binaries seems to be entirely >>> given over to porn. Not pictures, but lists of web sites. I suppose you >>> can post there anyway, like old times. You can email them to me; I'll >>> give them a URL of their own. >>> >>> Jerry >> >> >> >> There's an alt.binaries.electronic.schematic group, who's charter >> should be obvious from the name. How difficult would it be to set up >> an alt.binaries.dsp group (and hope that DSP doesn't stand for some >> darned sick perversion that we don't know about)? I know nothing of >> the inner workings of newsgroups, either technical or political, by >> the way. > > > Tim, > > I like your subtlety with Darned Sick Perversion. I didn't find > alt.binaries.electronic.schematic tracing down Google's tree, so maybe > there are more that I missed. It could be that the darned sick > perversions caught my attention to the exclusion of everything else, or > maybe there's another tree. I'll look again. > > I don't know about establishing a group any more than you, but if we > need one, we can find out. I hope we can just piggyback on something > already there. > > JerryP.S. I looked at most of the alt.binary... groups available from my news server -- there are hundreds of them -- but didn't find it there either. A group available through Google is preferable because not all news servers make all groups available. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●September 5, 20042004-09-05
Jerry Avins wrote:> > Tim, > > I like your subtlety with Darned Sick Perversion. I didn't find > alt.binaries.electronic.schematic tracing down Google's tree, so maybe > there are more that I missed. It could be that the darned sick > perversions caught my attention to the exclusion of everything else, or > maybe there's another tree. I'll look again. > > I don't know about establishing a group any more than you, but if we > need one, we can find out. I hope we can just piggyback on something > already there. > > JerryYou probably couldn't find it because I transposed the name: it's "alt.binaries.schematics.electronic". There's probably little enough need for it at this point that piggybacking on something remotely appropriate would be better. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Reply by ●September 6, 20042004-09-06
On Sat, 04 Sep 2004 18:52:34 -0500, Richard Owlett <rowlett@atlascomm.net> wrote:>Anyone have Scilab 2.7 or 3.0 up and running under WinXP? > >I'm trying to plot magnitude of FFT of rectangular pulse/window. > >My code is > >a(1:899)=1; >a(900:1000)=0; >plot2d( [1:1000], abs(fft(a,-1)), logflag="nl") > >My output does not resemble what Mr. Lyons gives in Figure 3-25c. > >Multiple sources agree with Rick ,) >Hi Richard, I don't "speak" Scilab language, but the Y-axis in my Figure 3-259c) is linear, not logarithmic. Richard, try this: 1) Create a time sequence that has 11 ones followed by 33 zeros. 2) Perform a 44-point DFT on that time-domain sequence. 3) Compute the magnitudes (absolute values) of your frequency-domains samples. 4) If you wish, divide the magnitudes by 11 (I did that just to normalize my Figure 3-25(c) so the the peak of the main lobe was equal to unity.) 5) Plot the freq-domain magnitude-sample sequence. What see should look an awful lot like my Figure 3-25(c). My intention for Figure 3-25(c) was to illustrate that: * The magnitude of the DFT (what I called |X(m)|), of a rectangular time sequence,is a sin(x)/x-like sequence (that's traditionally called a "Dirichlet kernel".) * That |X(m)| has a main lobe (or is it mainlobe?) whose peak value is proportional to the number of unity-valued samples ("K" in the figure) in the original time sequence. * The first positive-frequency null (zero-crossing) of |X(m)| is proportional to the length of the DFT (size of the DFT) *but* inversely proportional to the number of unity-valued samples in the original time sequence. (This inverse relationship is a *very* important issue. It means that the wider the non-zero rectangular time "pulse", the narrower the mainlobe of |X(m)|. That is, the narrower a function is in one domain, the wider will be its transform in the other domain. Likewise, the wider a function is in one domain, the narrower will be its transform in the other domain.) Keep experimenting Richard. That's the best way to learn. [-Rick-]
Reply by ●September 6, 20042004-09-06
Richard Owlett wrote:> Jerry Avins wrote: > >> Richard Owlett wrote: >> >> ... >> >>> Question for all: >>> Is there somewhere I can post my 100 point and 1000 point plots ( jpg >>> files)? >> >> >> >> Unless I missed something useful, alt.binaries seems to be entirely >> given over to porn. Not pictures, but lists of web sites. I suppose you >> can post there anyway, like old times. You can email them to me; I'll >> give them a URL of their own. >> >> Jerry > > > Thanks, I've sent you copies. > At least I think I have. I'm not getting incoming mail but outgoing > apparently works.I got your plots on your second try. You're now famous! Go to http://users.erols.com/jyavins/Owlett.htm ^ -->This must be a capital 'O'!<-- The plots look fine to me. The scalloping you see at the bottom with 1000 points is an artifact of the plotting process. The dots that are connected to make the curve have finite spacing. Not all of the zeros are represented by dots, (In fact, none are and most aren't close.) Don't feel too bad. Rick Lyons once fell into the same trap. It's proof again that computers sometimes lie, so you need to be skeptical Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●September 6, 20042004-09-06
Jerry Avins wrote:> Richard Owlett wrote: > >> Jerry Avins wrote: >> >>> Richard Owlett wrote: >>> >>> ... >>> >>>> Question for all: >>>> Is there somewhere I can post my 100 point and 1000 point plots ( >>>> jpg files)? >>> >>> >>> >>> >>> Unless I missed something useful, alt.binaries seems to be entirely >>> given over to porn. Not pictures, but lists of web sites. I suppose you >>> can post there anyway, like old times. You can email them to me; I'll >>> give them a URL of their own. >>> >>> Jerry >> >> >> >> Thanks, I've sent you copies. >> At least I think I have. I'm not getting incoming mail but outgoing >> apparently works. > > > I got your plots on your second try. You're now famous! Go to > http://users.erols.com/jyavins/Owlett.htm > ^ -->This must be a capital 'O'!<-- > > The plots look fine to me. The scalloping you see at the bottom with > 1000 points is an artifact of the plotting process. The dots that are > connected to make the curve have finite spacing. Not all of the zeros > are represented by dots, (In fact, none are and most aren't close.) > > Don't feel too bad. Rick Lyons once fell into the same trap. It's proof > again that computers sometimes lie, so you need to be skeptical > > JerryOn that it's modeling the fact that you can't sample perfectly evenly. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Reply by ●September 6, 20042004-09-06
Rick Lyons wrote:> On Sat, 04 Sep 2004 18:52:34 -0500, Richard Owlett > <rowlett@atlascomm.net> wrote: > > >>Anyone have Scilab 2.7 or 3.0 up and running under WinXP? >> >>I'm trying to plot magnitude of FFT of rectangular pulse/window. >> >>My code is >> >>a(1:899)=1; >>a(900:1000)=0; >>plot2d( [1:1000], abs(fft(a,-1)), logflag="nl") >> >>My output does not resemble what Mr. Lyons gives in Figure 3-25c. >> >>Multiple sources agree with Rick ,) >> > > > Hi Richard, > > I don't "speak" Scilab language, but the > Y-axis in my Figure 3-259c) is linear, not > logarithmic. >Not *quite* the trap I fell into ;} Originally my project was "Investigate various windowing functions". That led me to Fig 3-16(b) ( a log magnitude plot ) I started experimenting and got "strange" results. Went back to searching book for figure reference so I could post "SEMI"intelligent question. Saw Fig 3-25(c). Missed that it was linear scale. The rest ( as they say ) is history ;} Actually I think this all is a recommendation for your book. [ Not sure how publisher could convey it in an ad. ] The point being, an independent student using your book gets enough information to know when he's doing something wrong. That I didn't know *WHAT* was wrong is irrelevant. I've used toooo many course texts that failed this.> Richard, try this: > > [snip detailed test example ]I'll perform your exact example and some with the same proportions. Based on Jerry Avins' response when he posted the plots I got, I suspect there are interesting times ahead.
Reply by ●September 6, 20042004-09-06
Tim Wescott wrote:> Jerry Avins wrote: > >> Richard Owlett wrote: >> >>> Jerry Avins wrote: >>> >>>> Richard Owlett wrote: >>>> >>>> ... >>>> >>>>> Question for all: >>>>> Is there somewhere I can post my 100 point and 1000 point plots ( >>>>> jpg files)? >>>> >>>> >>>> >>>> >>>> >>>> Unless I missed something useful, alt.binaries seems to be entirely >>>> given over to porn. Not pictures, but lists of web sites. I suppose you >>>> can post there anyway, like old times. You can email them to me; I'll >>>> give them a URL of their own. >>>> >>>> Jerry >>> >>> >>> >>> >>> Thanks, I've sent you copies. >>> At least I think I have. I'm not getting incoming mail but outgoing >>> apparently works. >> >> >> >> I got your plots on your second try. You're now famous! Go to >> http://users.erols.com/jyavins/Owlett.htm >> ^ -->This must be a capital 'O'!<-- >> >> The plots look fine to me. The scalloping you see at the bottom with >> 1000 points is an artifact of the plotting process. The dots that are >> connected to make the curve have finite spacing. Not all of the zeros >> are represented by dots, (In fact, none are and most aren't close.) >> >> Don't feel too bad. Rick Lyons once fell into the same trap. It's proof >> again that computers sometimes lie, so you need to be skeptical >> >> Jerry > > > On that it's modeling the fact that you can't sample perfectly evenly.I don't think that perfectly even sampling is the issue. When plotting a curve by connecting regularly spaced dots, the probability that one of the dots falls on a zero is exceedingly small. (If one actually came up by an accident of fate, how big a screen would you need to represent it along with other data on a log plot? >-)) Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
