## [Audacity-devel] Chebyshev filter

 [Audacity-devel] Chebyshev filter From: Steve the Fiddle - 2013-12-31 15:35:55 ```In the Wikipedia article about Elliptic Filters (http://en.wikipedia.org/wiki/Elliptic_filter) it says: "As the ripple in the stopband approaches zero, the filter becomes a type I Chebyshev filter. As the ripple in the passband approaches zero, the filter becomes a type II Chebyshev filter" That makes sense when thinking of "ripple" as the amount of 'wiggle' in the passband/stopband, but that is not how the term "stopband ripple" is used in the new "Scientific Filter" effect. The "Stopband Ripple" in the Scientific Filter appears to be the minimum attenuation in the stop band, not the amount of "ripple". The Wikipedia article on Chebyshev Filters (http://en.wikipedia.org/wiki/Chebyshev_filter#Type_II_Chebyshev_filters) refers to "stopband attenuation y in decibels". I presume that this is what we are calling "Stopband Ripple" in the Scientific Filter? Steve ```

 [Audacity-devel] Chebyshev filter From: Steve the Fiddle - 2013-12-31 15:35:55 ```In the Wikipedia article about Elliptic Filters (http://en.wikipedia.org/wiki/Elliptic_filter) it says: "As the ripple in the stopband approaches zero, the filter becomes a type I Chebyshev filter. As the ripple in the passband approaches zero, the filter becomes a type II Chebyshev filter" That makes sense when thinking of "ripple" as the amount of 'wiggle' in the passband/stopband, but that is not how the term "stopband ripple" is used in the new "Scientific Filter" effect. The "Stopband Ripple" in the Scientific Filter appears to be the minimum attenuation in the stop band, not the amount of "ripple". The Wikipedia article on Chebyshev Filters (http://en.wikipedia.org/wiki/Chebyshev_filter#Type_II_Chebyshev_filters) refers to "stopband attenuation y in decibels". I presume that this is what we are calling "Stopband Ripple" in the Scientific Filter? Steve ```
 Re: [Audacity-devel] Chebyshev filter From: Norm C - 2014-01-02 00:27:32 ```Stevethefiddle wrote > ...zero, the filter becomes a type II Chebyshev filter" > > That makes sense when thinking of "ripple" as the amount of 'wiggle' > in the passband/stopband, but that is not how the term "stopband > ripple" is used in the new "Scientific Filter" effect. It is, although to be completely accurate it should be specified as the negative of what it is currently specified as. Eg. for a stopband ripple of "20 dB", the stopband response varies between 0 (-inf dB) and 0.1 (-20 dB). If you think it's clearer to specify it as a negative quantity, I can change it accordingly - probably the best thing would be to put a negative sign in the non-editable portion just to the left of the edit box though, since it can never go positive. Norm -- View this message in context: http://audacity.238276.n2.nabble.com/Chebyshev-filter-tp7561009p7561026.html Sent from the audacity-devel mailing list archive at Nabble.com. ```
 Re: [Audacity-devel] Chebyshev filter From: David Bailes - 2014-01-02 17:18:48 Attachments: Message as HTML ```On Thu, Jan 2, 2014 at 12:27 AM, Norm C wrote: > Stevethefiddle wrote > > ...zero, the filter becomes a type II Chebyshev filter" > > > > That makes sense when thinking of "ripple" as the amount of 'wiggle' > > in the passband/stopband, but that is not how the term "stopband > > ripple" is used in the new "Scientific Filter" effect. > > It is, although to be completely accurate it should be specified as the > negative of what it is currently specified as. Eg. for a stopband ripple of > "20 dB", the stopband response varies between 0 (-inf dB) and 0.1 (-20 dB). > > If you think it's clearer to specify it as a negative quantity, I can > change > it accordingly - probably the best thing would be to put a negative sign in > the non-editable portion just to the left of the edit box though, since it > can never go positive. > I don't think that hard coding the negative sign is the best way to do this. If someone has to enter a negative number, then I think that they should have to type in a negative number. In addition, if the negative sign is hard coded, then code will have to be added/changed so that screen readers read the number correctly. David. > > ```
 Re: [Audacity-devel] Chebyshev filter From: Steve the Fiddle - 2014-01-02 18:49:05 ```I'm still not clear about this. For a Chebyshev type 2 filter, is "stopband ripple" the same as "minimum stopband attenuation" or are they different things? Wikipedia seems to be saying that they are different things but the Scientific Filter interface seems to be saying they are one and the same. Can someone clarify this point? Steve On 2 January 2014 17:18, David Bailes wrote: > On Thu, Jan 2, 2014 at 12:27 AM, Norm C wrote: >> >> Stevethefiddle wrote >> > ...zero, the filter becomes a type II Chebyshev filter" >> > >> > That makes sense when thinking of "ripple" as the amount of 'wiggle' >> > in the passband/stopband, but that is not how the term "stopband >> > ripple" is used in the new "Scientific Filter" effect. >> >> It is, although to be completely accurate it should be specified as the >> negative of what it is currently specified as. Eg. for a stopband ripple >> of >> "20 dB", the stopband response varies between 0 (-inf dB) and 0.1 (-20 >> dB). >> >> If you think it's clearer to specify it as a negative quantity, I can >> change >> it accordingly - probably the best thing would be to put a negative sign >> in >> the non-editable portion just to the left of the edit box though, since it >> can never go positive. > > > I don't think that hard coding the negative sign is the best way to do this. > If someone has to enter a negative number, then I think that they should > have to type in a negative number. > In addition, if the negative sign is hard coded, then code will have to be > added/changed so that screen readers read the number correctly. > > David. >> >> > > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > audacity-devel mailing list > audacity-devel@... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > ```
 Re: [Audacity-devel] Chebyshev filter From: - 2014-01-02 19:23:27 ```Steve, Actually, the word "ripple" makes sense only for passband. For stopband ripple is always infinite, since it is between minus infinite and some gain represented in decibels as a negative value. You can see this looking at the gain plot of Chebychev II. I suggest using attenuation, instead of gain, i.e., "maximum passband attenuation" and "minimum stopband attenuation". They are both positive for any filter, be it Butterworth, Chebychev I or Chebychev II or any other, such as elliptic por Bessel. This is customary in filter theory. It is unnecessary to use the word ripple. People who are knowledgeable about filter theory will naturally associate ripple with maximum passband attenuation. Newbies will consider it as some magic word. Regards, Federico > I'm still not clear about this. > For a Chebyshev type 2 filter, is "stopband ripple" the same as > "minimum stopband attenuation" or are they different things? > Wikipedia seems to be saying that they are different things but the > Scientific Filter interface seems to be saying they are one and the > same. Can someone clarify this point? > > Steve > > On 2 January 2014 17:18, David Bailes wrote: >> On Thu, Jan 2, 2014 at 12:27 AM, Norm C wrote: >>> >>> Stevethefiddle wrote >>> > ...zero, the filter becomes a type II Chebyshev filter" >>> > >>> > That makes sense when thinking of "ripple" as the amount of 'wiggle' >>> > in the passband/stopband, but that is not how the term "stopband >>> > ripple" is used in the new "Scientific Filter" effect. >>> >>> It is, although to be completely accurate it should be specified as the >>> negative of what it is currently specified as. Eg. for a stopband >>> ripple >>> of >>> "20 dB", the stopband response varies between 0 (-inf dB) and 0.1 (-20 >>> dB). >>> >>> If you think it's clearer to specify it as a negative quantity, I can >>> change >>> it accordingly - probably the best thing would be to put a negative >>> sign >>> in >>> the non-editable portion just to the left of the edit box though, since >>> it >>> can never go positive. >> >> >> I don't think that hard coding the negative sign is the best way to do >> this. >> If someone has to enter a negative number, then I think that they >> should >> have to type in a negative number. >> In addition, if the negative sign is hard coded, then code will have to >> be >> added/changed so that screen readers read the number correctly. >> >> David. >>> >>> >> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into >> your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >> AppDynamics >> Pro! >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> audacity-devel mailing list >> audacity-devel@... >> https://lists.sourceforge.net/lists/listinfo/audacity-devel >> > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics > Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > audacity-devel mailing list > audacity-devel@... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > ```
 Re: [Audacity-devel] Chebyshev filter From: Steve the Fiddle - 2014-01-02 21:34:22 ```On 2 January 2014 19:23, wrote: > Steve, > > Actually, the word "ripple" makes sense only for passband. For stopband > ripple is always infinite, since it is between minus infinite and some > gain represented in decibels as a negative value. You can see this looking > at the gain plot of Chebychev II. Thanks Federico, that certainly brings some clarification. Looking at the diagrams on Wikipedia: http://en.wikipedia.org/wiki/Elliptic_filter#Comparison_with_other_linear_filters it looks like, for Chebyshev type 2 filters, the minimum attenuation is "1 - ripple" on a linear scale. Is that correct? Sorry if these questions appear noobish, but as this is being put forward as Scientific we need to be precise with the documentation. Also, can you suggest some practical use cases when Chebyshev filters might be wanted in digital audio processing rather than a Butterworth filter? Being that we are not concerned about 'component count' the usual description (for analogue Chebyshev filters) about a slightly sharper cutoff seems hardly relevant, but I think we need to say something about why a Chebyshev filter might be needed. Thanks Steve > > I suggest using attenuation, instead of gain, i.e., "maximum passband > attenuation" and "minimum stopband attenuation". They are both positive > for any filter, be it Butterworth, Chebychev I or Chebychev II or any > other, such as elliptic por Bessel. This is customary in filter theory. It > is unnecessary to use the word ripple. People who are knowledgeable about > filter theory will naturally associate ripple with maximum passband > attenuation. Newbies will consider it as some magic word. > > Regards, > > Federico > > >> I'm still not clear about this. >> For a Chebyshev type 2 filter, is "stopband ripple" the same as >> "minimum stopband attenuation" or are they different things? >> Wikipedia seems to be saying that they are different things but the >> Scientific Filter interface seems to be saying they are one and the >> same. Can someone clarify this point? >> >> Steve >> >> On 2 January 2014 17:18, David Bailes wrote: >>> On Thu, Jan 2, 2014 at 12:27 AM, Norm C wrote: >>>> >>>> Stevethefiddle wrote >>>> > ...zero, the filter becomes a type II Chebyshev filter" >>>> > >>>> > That makes sense when thinking of "ripple" as the amount of 'wiggle' >>>> > in the passband/stopband, but that is not how the term "stopband >>>> > ripple" is used in the new "Scientific Filter" effect. >>>> >>>> It is, although to be completely accurate it should be specified as the >>>> negative of what it is currently specified as. Eg. for a stopband >>>> ripple >>>> of >>>> "20 dB", the stopband response varies between 0 (-inf dB) and 0.1 (-20 >>>> dB). >>>> >>>> If you think it's clearer to specify it as a negative quantity, I can >>>> change >>>> it accordingly - probably the best thing would be to put a negative >>>> sign >>>> in >>>> the non-editable portion just to the left of the edit box though, since >>>> it >>>> can never go positive. >>> >>> >>> I don't think that hard coding the negative sign is the best way to do >>> this. >>> If someone has to enter a negative number, then I think that they >>> should >>> have to type in a negative number. >>> In addition, if the negative sign is hard coded, then code will have to >>> be >>> added/changed so that screen readers read the number correctly. >>> >>> David. >>>> >>>> >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into >>> your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>> AppDynamics >>> Pro! >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> audacity-devel mailing list >>> audacity-devel@... >>> https://lists.sourceforge.net/lists/listinfo/audacity-devel >>> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >> Pro! >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> audacity-devel mailing list >> audacity-devel@... >> https://lists.sourceforge.net/lists/listinfo/audacity-devel >> > > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > audacity-devel mailing list > audacity-devel@... > https://lists.sourceforge.net/lists/listinfo/audacity-devel ```
 Re: [Audacity-devel] Chebyshev filter From: Federico Miyara - 2014-01-03 00:38:23 ```Steve, > Looking at the diagrams on Wikipedia: > http://en.wikipedia.org/wiki/Elliptic_filter#Comparison_with_other_linear_filters > it looks like, for Chebyshev type 2 filters, the minimum attenuation > is "1 - ripple" on a linear scale. Is that correct? Not exactly. On a linear scale, attenuation is the reciprocal of gain. In the Wikipedia example, you have a linear ripple of 0.1 in the stopband, so maximum stopband gain is 0.1. Minimum stopband attenuation is, thus, 1/0.1 = 10, and not 1 - 0.1 = 0.9. This means any component in the stopband is attenuated at least by a factor of 10. For instance, a tone of amplitude 0.3 is attenuated to 0.03 or less, depending on its frequency: in a fortunate case it would be reduced to zero. On a log scale, linear 0 gain will translate into minus infinite dB gain, while in terms of attenuation it would be infinite dB. Since log(1/x) = -log(x), gain and attenuation in dB are one the opposite of the other. Thus, when linear gain goes to zero, linear attenuation goes to infinite, so logarithmic gain goes to minus infinite while log attenuation goes to infinite. When linear gain is 0.1, log gain is -20 dB and log attenuation is + 20 dB. So type II Chebyshev stopband attenuation goes from 20 dB to infinite dB once and again, being 20 dB the minimum stopband attenuation > Sorry if these questions appear noobish, but as this is being put > forward as Scientific we need to be precise with the documentation. > Also, can you suggest some practical use cases when Chebyshev filters > might be wanted in digital audio processing rather than a Butterworth > filter? Being that we are not concerned about 'component count' the > usual description (for analogue Chebyshev filters) about a slightly > sharper cutoff seems hardly relevant, but I think we need to say > something about why a Chebyshev filter might be needed. This is a very good question. Chebyshev filters have been used in analog audio circuits because they require far less components to attain the same general specs than Butterworth. The tradeoff is that they have worse phase response and transient response (ringing). They are useful, for instance, as analog antialias filters Probably I wouldn't actually use Chebyshev filters in audio processing because an FFT filter does a better work, but they are valuable for experimental work, for instance to test how that kind of filter will affect different types of signals prior to hardware implementation. As a didactic tool it is also quite interesting. Consider that many teachers use Audacity as a class software tool. Regards, Federico ```
 Re: [Audacity-devel] Chebyshev filter From: Martyn Shaw - 2014-01-03 01:35:34 ```+1 Martyn On 03/01/2014 00:37, Federico Miyara wrote: > > Steve, > >> Looking at the diagrams on Wikipedia: >> http://en.wikipedia.org/wiki/Elliptic_filter#Comparison_with_other_linear_filters >> it looks like, for Chebyshev type 2 filters, the minimum attenuation >> is "1 - ripple" on a linear scale. Is that correct? > > Not exactly. > > On a linear scale, attenuation is the reciprocal of gain. In the > Wikipedia example, you have a linear ripple of 0.1 in the stopband, so > maximum stopband gain is 0.1. Minimum stopband attenuation is, thus, > 1/0.1 = 10, and not 1 - 0.1 = 0.9. This means any component in the > stopband is attenuated at least by a factor of 10. For instance, a tone > of amplitude 0.3 is attenuated to 0.03 or less, depending on its > frequency: in a fortunate case it would be reduced to zero. > > On a log scale, linear 0 gain will translate into minus infinite dB > gain, while in terms of attenuation it would be infinite dB. Since > log(1/x) = -log(x), gain and attenuation in dB are one the opposite of > the other. Thus, when linear gain goes to zero, linear attenuation goes > to infinite, so logarithmic gain goes to minus infinite while log > attenuation goes to infinite. > > When linear gain is 0.1, log gain is -20 dB and log attenuation is + 20 dB. > > So type II Chebyshev stopband attenuation goes from 20 dB to infinite dB > once and again, being 20 dB the minimum stopband attenuation > >> Sorry if these questions appear noobish, but as this is being put >> forward as Scientific we need to be precise with the documentation. >> Also, can you suggest some practical use cases when Chebyshev filters >> might be wanted in digital audio processing rather than a Butterworth >> filter? Being that we are not concerned about 'component count' the >> usual description (for analogue Chebyshev filters) about a slightly >> sharper cutoff seems hardly relevant, but I think we need to say >> something about why a Chebyshev filter might be needed. > > This is a very good question. Chebyshev filters have been used in analog > audio circuits because they require far less components to attain the > same general specs than Butterworth. The tradeoff is that they have > worse phase response and transient response (ringing). They are useful, > for instance, as analog antialias filters > > Probably I wouldn't actually use Chebyshev filters in audio processing > because an FFT filter does a better work, but they are valuable for > experimental work, for instance to test how that kind of filter will > affect different types of signals prior to hardware implementation. As a > didactic tool it is also quite interesting. Consider that many teachers > use Audacity as a class software tool. > > Regards, > > Federico > > > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > audacity-devel mailing list > audacity-devel@... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > ```
 Re: [Audacity-devel] Chebyshev filter From: Martyn Shaw - 2014-01-03 01:00:52 ```Hi I agree with Federico that the terms "maximum passband attenuation" and "minimum stopband attenuation" are good ones to use. And an attenuation of '1dB' means 'make it 1dB smaller', so no problem with the minus sign. So I committed that. I didn't like the way that the controls were still there but 'grayed' out for the different filters, so I made the ones that are relevant visible and the rest invisible. I committed that as well but left the old code in comments in case people don't like it. I think we should change the names of quite a number of variables, but left that for now. On 02/01/2014 21:34, Steve the Fiddle wrote: > On 2 January 2014 19:23, wrote: >> Steve, >> >> Actually, the word "ripple" makes sense only for passband. For stopband >> ripple is always infinite, since it is between minus infinite and some >> gain represented in decibels as a negative value. You can see this looking >> at the gain plot of Chebychev II. > > Thanks Federico, that certainly brings some clarification. > > Looking at the diagrams on Wikipedia: > http://en.wikipedia.org/wiki/Elliptic_filter#Comparison_with_other_linear_filters > it looks like, for Chebyshev type 2 filters, the minimum attenuation > is "1 - ripple" on a linear scale. Is that correct? Well, removing the word 'ripple' from the interface makes this a moot point really. But on those diagrams, it looks like in the stopband the filter goes down to about 0.1. That's -20dB of gain or 20dB of attenuation. On a linear scale that's a gain of (a factor of) 0.1 or an attenuation of (a factor of) 10. > Sorry if these questions appear noobish, but as this is being put > forward as Scientific we need to be precise with the documentation. No problem :-). I'd like documentation accurate too. > Also, can you suggest some practical use cases when Chebyshev filters > might be wanted in digital audio processing rather than a Butterworth > filter? Being that we are not concerned about 'component count' the Well, 'component count' equates to 'machine cycles' in our digital realm, really. > usual description (for analogue Chebyshev filters) about a slightly > sharper cutoff seems hardly relevant, but I think we need to say > something about why a Chebyshev filter might be needed. It's a classic! This effect give us the 'traditional' IIR filters that are derived from analogue filters. The Equalization effect does 'traditional' FIR filter design, so we have both. We currently don't have more up-to-date methods of making digital filters, and I'm guessing that there are many, but I haven't been following the literature. I am still referring to a 1975 edition of O&S http://www.amazon.co.uk/Digital-Signal-Processing-Alan-Oppenheim/dp/0132146355 ! TTFN Martyn > Thanks > Steve > >> >> I suggest using attenuation, instead of gain, i.e., "maximum passband >> attenuation" and "minimum stopband attenuation". They are both positive >> for any filter, be it Butterworth, Chebychev I or Chebychev II or any >> other, such as elliptic por Bessel. This is customary in filter theory. It >> is unnecessary to use the word ripple. People who are knowledgeable about >> filter theory will naturally associate ripple with maximum passband >> attenuation. Newbies will consider it as some magic word. >> >> Regards, >> >> Federico >> >> >>> I'm still not clear about this. >>> For a Chebyshev type 2 filter, is "stopband ripple" the same as >>> "minimum stopband attenuation" or are they different things? >>> Wikipedia seems to be saying that they are different things but the >>> Scientific Filter interface seems to be saying they are one and the >>> same. Can someone clarify this point? >>> >>> Steve >>> >>> On 2 January 2014 17:18, David Bailes wrote: >>>> On Thu, Jan 2, 2014 at 12:27 AM, Norm C wrote: >>>>> >>>>> Stevethefiddle wrote >>>>>> ...zero, the filter becomes a type II Chebyshev filter" >>>>>> >>>>>> That makes sense when thinking of "ripple" as the amount of 'wiggle' >>>>>> in the passband/stopband, but that is not how the term "stopband >>>>>> ripple" is used in the new "Scientific Filter" effect. >>>>> >>>>> It is, although to be completely accurate it should be specified as the >>>>> negative of what it is currently specified as. Eg. for a stopband >>>>> ripple >>>>> of >>>>> "20 dB", the stopband response varies between 0 (-inf dB) and 0.1 (-20 >>>>> dB). >>>>> >>>>> If you think it's clearer to specify it as a negative quantity, I can >>>>> change >>>>> it accordingly - probably the best thing would be to put a negative >>>>> sign >>>>> in >>>>> the non-editable portion just to the left of the edit box though, since >>>>> it >>>>> can never go positive. >>>> >>>> >>>> I don't think that hard coding the negative sign is the best way to do >>>> this. >>>> If someone has to enter a negative number, then I think that they >>>> should >>>> have to type in a negative number. >>>> In addition, if the negative sign is hard coded, then code will have to >>>> be >>>> added/changed so that screen readers read the number correctly. >>>> >>>> David. >>>>> >>>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Rapidly troubleshoot problems before they affect your business. Most IT >>>> organizations don't have a clear picture of how application performance >>>> affects their revenue. With AppDynamics, you get 100% visibility into >>>> your >>>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of >>>> AppDynamics >>>> Pro! >>>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> audacity-devel mailing list >>>> audacity-devel@... >>>> https://lists.sourceforge.net/lists/listinfo/audacity-devel >>>> >>> >>> ------------------------------------------------------------------------------ >>> Rapidly troubleshoot problems before they affect your business. Most IT >>> organizations don't have a clear picture of how application performance >>> affects their revenue. With AppDynamics, you get 100% visibility into your >>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics >>> Pro! >>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> audacity-devel mailing list >>> audacity-devel@... >>> https://lists.sourceforge.net/lists/listinfo/audacity-devel >>> >> >> >> >> ------------------------------------------------------------------------------ >> Rapidly troubleshoot problems before they affect your business. Most IT >> organizations don't have a clear picture of how application performance >> affects their revenue. With AppDynamics, you get 100% visibility into your >> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! >> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk >> _______________________________________________ >> audacity-devel mailing list >> audacity-devel@... >> https://lists.sourceforge.net/lists/listinfo/audacity-devel > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > audacity-devel mailing list > audacity-devel@... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > ```
 Re: [Audacity-devel] Chebyshev filter From: Martyn Shaw - 2014-01-02 01:47:10 ```Hi Steve I see what you mean. In the section of the interface where we currently have: Passband Ripple: [ ] dB Stopband Ripple: [ ] dB we should really have text and boxes relevant to the type of filter selected: Butterworth: nothing Chebyshev type 1: Passband ripple, stopband cutoff Chebyshev type 2: Passband cutoff, stopband ripple rather than just making the boxes un-editable in ScienFilterDialog::EnableDisableRippleCtl I have had a go at doing this, but not been successful so far. TTFN Martyn On 31/12/2013 15:35, Steve the Fiddle wrote: > In the Wikipedia article about Elliptic Filters > (http://en.wikipedia.org/wiki/Elliptic_filter) it says: > "As the ripple in the stopband approaches zero, the filter becomes a > type I Chebyshev filter. As the ripple in the passband approaches > zero, the filter becomes a type II Chebyshev filter" > > That makes sense when thinking of "ripple" as the amount of 'wiggle' > in the passband/stopband, but that is not how the term "stopband > ripple" is used in the new "Scientific Filter" effect. > > The "Stopband Ripple" in the Scientific Filter appears to be the > minimum attenuation in the stop band, not the amount of "ripple". > > The Wikipedia article on Chebyshev Filters > (http://en.wikipedia.org/wiki/Chebyshev_filter#Type_II_Chebyshev_filters) > refers to "stopband attenuation y in decibels". I presume that this is > what we are calling "Stopband Ripple" in the Scientific Filter? > > Steve > > ------------------------------------------------------------------------------ > Rapidly troubleshoot problems before they affect your business. Most IT > organizations don't have a clear picture of how application performance > affects their revenue. With AppDynamics, you get 100% visibility into your > Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! > http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk > _______________________________________________ > audacity-devel mailing list > audacity-devel@... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > ```