On Jan 28, 10:11�pm, Jerry Avins <j...@ieee.org> wrote:> On 1/28/2012 9:11 PM, maury wrote: > > > > > > > On Jan 28, 5:18 pm, Jerry Avins<j...@ieee.org> �wrote: > >> On 1/28/2012 5:00 PM, maury wrote: > > >>> On Jan 28, 3:21 pm, "Vl4ever"<moikef@n_o_s_p_a_m.gmail.com> � �wrote: > >>>>> On Tue, 17 Jan 2012 11:22:45 -0500, Jerry Avins<j...@ieee.org> � �wrote: > > >>>>>> On 1/17/2012 8:27 AM, Rick Lyons wrote: > >>>>>>> On Sat, 14 Jan 2012 15:06:04 -0600, "Vl4ever" > >>>>>>> <moikef@n_o_s_p_a_m.gmail.com> � � �wrote: > > >>>>>>> � � � [Snipped by Lyons] > > >>>>>>> V14ever, > > >>>>>>> � � �you might help at: > > >>>>http://www.mathworks.com/products/signal/demos.html?file=/products/de... > > >>>>>> Rick, > > >>>>>> Youse is a good guy! Did you let Mathworks know about the bug you found? > > >>>>>> Jerry > > >>>>> Hi Jer, > >>>>> � �No I didn't. �In the past when I thought I found > >>>>> a bug in some commercial software it usually turned > >>>>> to be my 'operator error' instead of a bug. > > >>>>> Maybe I'll experiment further to se4e waht I can learn. > > >>>>> See Ya', > >>>>> [-Rick-] > > >>>> Thank you all for your responses,but unfortunately you have shown me so far > >>>> (except for almost complete 'G' vector description) things I have ran into > >>>> before.SOS and G vectors still remain the same in any form you look at it ( > >>>> whether in *.mat format or in workspace in Matlab.The question is still the > >>>> same-having the layout on the following form : > > >>>> G = > > >>>> � � � 0.0037 > >>>> � � � 0.0037 > >>>> � � � 0.0028 > >>>> � � � 1.0000 > > >>>> SOS = > > >>>> � � � 1.0000 � � � � 0 � -1.0000 � �1.0000 � -1.9887 � �0.9971 > >>>> � � � 1.0000 � � � � 0 � -1.0000 � �1.0000 � -1.9901 � �0.9973 > >>>> � � � 1.0000 � � � � 0 � -1.0000 � �1.0000 � -1.9867 � �0.9945 > >>>> I know now that first three values of G (counting from the top down) are > >>>> multiplicators of its respective row in SOS matrix.The question is what is > >>>> the last value is meant for exactly? > >>>> Thank You, > >>>> � � � � � � Michael.- Hide quoted text - > > >>>> - Show quoted text - > > >>> Return with us now to those early days of yester-year. Specifically, > >>> Jan. 16 where I wrote > > >>> "The last row of the gain matrix is the gain applied to the over > >>> filter. It's all right there in the documentation." > > >>> Well, let's be more detailed. You have all those sections with their > >>> gains, right? Well put all of those section in side a big box. This is > >>> now the overall filter. Thake the output of this filter and apply the > >>> gain of the last row of the gain vector. In other words, the gain is > >>> applied to the overall filter. > > >> So the last entry is the product of the first three? It doesn't look > >> that way to me. What happened? > > >> Jerry > > >> -- > >> Engineering is the art of making what you want from things you can get. > >> �����������������������������������������������������������������������- Hide quoted text - > > >> - Show quoted text - > > > No, Jerry. Let the gains be g1, g2, g3, g4 for each section. Let the > > overall gain be G. Let the sections be s1 ... Then the block dia would > > be > > > s1 g1 ----> �s2 g2----> �s3 g3----> �s4 g4----> �G > > > Whenever I've done IIR filters with fdatool, the overall gain has > > always been 1. > > > Maurice > > Thanks, Maury, but I'm still confused. When cascading amplifiers and > attenuators, the overall gain is the product of the individual sections. > Also, filters usually have frequency-dependent gains, so what do the > numbers really mean? > > Jerry > -- > Engineering is the art of making what you want from things you can get. > �����������������������������������������������������������������������- Hide quoted text - > > - Show quoted text -As I understand it, this is a gain you would apply after going through all sections. I've always converted the fdatool sections into 1 section, especially since I'm usually doing the filtering in a processor.