About my question : 

My special interest is in Amateur Radio C.W. (morse code)

operations which require very sharply defined passband

filtering in order to hear the message signal through the

radio band noise.  

I use Analog circuits that do not require a computer interface

or require computer data processing.  

This stand-alone design approach has wide application 

in devices that are used in lab work, with embedded filtering. 

I have explored this question for several years,

and have worked with several commercial analog filters

which work acceptably well,

and have also developed several one of my own.

Perhaps you have developed devices or circuits of your

own use in your research projects.

Reference can be made to Autek, DaTong, Tayloe, Ellis,

Clements, Davies, Cripe, Delyannis-Friend, et al

... basically anything that can be brought up from the internet,

or any device / circuit you have developed,

or has been published by others ( links or papers are OK ).

If you visit my website, be sure to click

(0) the 'AFX Basic Filter' page 

(1) the 'AFX Advanced Filter'  page

(2) the 'AFX Roof Triad Filter'  page. 

(3) the "AFV" variable f(), Dual-Notch Filter page. 

Stretch your brain and read on about Analog ! 

Try to compare Analog Filter results with DSP Filter results. 

Glen A. Ellis ,

retired , University of Tennessee, Medical Units, Memphis, TN.

http://www.geocities.ws/glene77is/index.html

Any practical filter specifications can be mapped to one of the four most common filter functions with certain filter complexity determined basically by the steepness of the filter transitions from the pass band to the stop band.

Then the poles and the zero of the filter can be implemented by analog circuits or digital circuits. The two implementations are equivalent and give more or less the same frequency and phase response. In case of digital implementation, the frequency response is periodic because of the sampling and digitization of the analog signal before entering to the digital filter circuit. So, the analog signal to be digitally filtered must be bandwidth limited to avoid aliasing. In case of audio signals, there is no requirement on the phase linearity of the filter.So, elliptic filters with fewer components can be used to achieve the same complexity.

Best wishes.

Hello Abdelhalim Zekry,

I agree with your answer above but I feel we should mention the domaines

s-domain linear analog and z-domain linear digital

best regards

ERIK LINDBERG

Hello Erik,

I am happy that you commented on my answer and added the most appropriate functional description domain, the S-domain for the analog filters and te Z domain for the digital filters. It is so that the original S-domain description of the filter can be converted to the Z-domain which is a discrete form of the S-domain. One of a very famous transform is the bilinear transform.

Yes my answer may be far from comprehensive You are right Erik. Thank you 

Researchers, 

Thank you Abdelhaliml and Erik. 

I was certain that Erik would write something. 

Gentlemen,

We have strayed from the intended Question. 

Are there

any Analog Active Audio Filters 

that match

any DSP filters  ? 

and thereafter,  Why and Wherefore ? 

as a comparison / contrast  

topically , schematically, and certainly math. 

I am seeking to compare my findings

with other researchers experiences

in this interesting field. 

I have listed links to several examples of my research. 

Here is the result of one of the circuits:

***************************************************************

*** Differential Dual-Notch Filter "AFV-3RL-v4FS-D-vQ":

*** 100 Hz = -3dB     

*** 225 Hz at -12dB

*** 350 Hz = -80dB  

***   7 Hz bandpass spread   per dB attenuation

***   as the Bode shows below. 

*** This is nearly a Vertical Side-Band Skirt

       between -3 dB and -80 dB. 

***************************************************************

In researching, I have found several audio filters ( DaTong ) 

that have near DSP qualities, based on quad State Variable filters 

with Elliptic designs ...  which are component, temperature 

and tuning sensitive.   But,  DaTong is not a Brick-Wall Filter.  

In Fact, 

I question the definition of DSP "Brick Wall" filtering, 

 ( I think it is a biased description ) 

since all my sources simply assume that DSP is "Brick Wall"

and computer aided filtering are the final ultimate method. 

As a "teaser" I offer this for consideration : 

My Analog results are Real-Time

as delivered to the operator's EAR, 

but  

DSP results are always presented as Computer Data ,

never after DAC conversion to the operator's EAR. . 

...

I expect that  if  DSP results

are DAC to natural audio signals driven through earphones,

then there will be substantial changes / degradations

in the resulting final measurements of the DSP signals. 

Dobromir Dobrev has kindly provided a his DSP workup

for my Analog AFX Roofing-Triad-Differential filter,

which is part of my larger project developing

an Analog CW Signal Processor.  

He included bode plots of both the filtered effect

also some interesting plots of the filtered-out-signal.  

Later, he sent gain equations for the  S-domain 

which helped me better understand the math involved

for the three interacting stages in the Roofing filter.   

Dobri's  work is remarkable  and duplicates my basic

early pass-band form shape and the stop-band  down to -48 dB.  

My latter circuits ( linked ) push this down to -98 dB. 

So,

the question remains centered on Analog BandPass Circuits

that may have DSP filtering qualities.

I am hoping that you researchers

can help broaden my awareness

about the great world of Audio Filtering. 

--------------------------------------------------------------------------

On the more technician side of this project,

( being 40 years past school and long retired ) 

my efforts have been to use

(1) fundamental equations

(2) Spice

(3) constructed-model-testing and measurements

to design several different approaches to

Analog BandPass Filters , which I use daily 

in my radio operations on-the-air. 

Hoping for your insights and understanding,

Glen 

http://www.geocities.ws/glene77is/index.html

http://www.geocities.ws/glene77is/GC_ET_AFX_AFV.html

Dear Glen,

You did not find directly the answer of your question. However , i think i  answered your question but may be indirectly. 

Yes, any filter specifications can be implemented by either an analog filter or by a digital filter. It is expected after the implementation, they must have more or less the same performance concerning the gain frequency response.

Assume you give me filter specifications at the given above, at first i determine the prototype low pass filter, then i find the complexity of the filter for certain standard filter function as eg Chebychev. then we get the poles of the filter in the S-domain. Transfer these poles to resonance circuits for the required band pass filter. every pole of the low pass filter gives a resonance circuit. A third order prototype low pass filter has three resonance circuits of corresponding BPF. So,we can implement such resonance circuits using quadratic sections of active filters. 

To design an equivalent digital filter, the prototype low pass filter response is transformed into Z- transfer characteristics of the band pass filter H(Z) using the bi linear transformation as i hinted before. This H(z) can be implemented either by software on a digital signal processing or a general purpose computer or digitally by FPGA, field programmable gate array using VHDL.

The resulting filters will be comparable in performance. provided that they have the same signal to noise ratio.

For more information please refer to the the book in the website: dsp-book.narod.ru/PADFD.pdf

I hope i get some steps closer to your question.

Abdelhalim,

Thank you for the short tutorial.  

It is clearly written and I follow it well.

However, we are maybe aiming at different targets.

I am aware that most of the world

is handling signal processing via digital analysis.

I do not wish to implement a digital solution.

Dobromir Dobrev has kindly provided his DSP workup

with explanation. I have gone as far as I want in that.

I am researching Analog filters.

... Various types of Analog filters.

...

The devices I have developed are to be attached

to the audio output of a radio transceiver.

...

The devices are "piggy-backed' onto a transceiver

in order to enhance the C.W. filtering options.

...

This is for a specific application in human radio operations

using the ear/brain to do analysis of morse code messages.

-----------------------------------------------------------------------------------------------

I apologize,

I need to rethink my question, as it may be circular.

I want to compare the filtering qualities of my own research filters

against a digital filter (which is not yet defined).

(1) Your answer is that

"Anything an Analog Filter Can Do

a DSP Can Do Also."

(2) My question may be

"Anything a DSP Can Do

an Analog Filter Can Do Also ??? "

Each perspective depends on the other perspective.

So, perhaps I already have as much answer as is practical.

Glen

Gentlemen,

just one short comment: If we want to compare the performance capabilities of both analog and digital filter structures, I think we should not exclude "sampled data" realizations (S/C filters) because they are located "between" the analog and the digital domain (analog but not time-continuous). What do you think?

Lutz,

Thank you. Very practical suggestion.  

Yes, the Switched Capacitor Circuit has been considered. 

It has measured noise levels caused by the "switching" 

The Switching Noise increases with Narrowness of PassBand. 

I have a page on my website where a associate has 

used a Spectrum analyzer on a real time S/C circuit 

and has measured the Switching Noise. 

As the S/C passband is Narrowed towards 100 Hz, 

the power-sum of the switching noise become intolerable

in the order of -12 dB below signal. 

....

Our analog approach has a very minor amount

of noise from the OpAmps and MFB filter phase noise,

in the order of -77 dB below signal. 

We use ngSPICE and then construct circuits.

In our working circuits, we compare  

(a) BandWidth

(b) Side-Band Slope characteristics. 

Then,

Compare the Third Harmonic values, and Spread of Passband

down -12dB, -24dB, -48dB, and further. 

Two Bode plots should provide the basc data for this analysis. 

What I expect is that the DSP signal will not be so pristine 

in the real world ... ie, the DSP "Brick-Wall" is a mythe.    

I expect the DAC to introduce Third Harmonic 

and Spread of Passband distortions.  

This will help define the DSP Real World Signal,

and that can be used to compare against the Analog Signal. 

Attached are some of the measurements 

wherein we measured the S/C Switching Noise 

compared with our very quiet Analog Filtering. 

These Analog circuits are from the CALF development 

where we were trying to develop a combination 

of Analog MFB filters enhanced by the Switched-Capacitor circuit. 

That is when we realized that the S/C produced much noise. 

I always welcome your good comments

and try to give them a good reading.

http://www.geocities.ws/glene77is/GC_ET_AFX_NeScaf.html

Dear Glen,

This is what i meant and expressed in my answer. I wrote:

Yes, any filter specifications can be implemented by either an analog filter or by a digital filter. It is expected after the implementation, they must have more or less the same performance concerning the gain frequency response. 

However, concerning the noise immunity the digital filters are more noise immunize. They are insensitivity to enviromental effects and components tolerance. They can be easily altered by software. But on the other side they are subjected to quantization error. more complicated in construction, need A/D AND D/A converters.

They are subjected to aliasing except that their input signal is limited to half the sampling frequency. 

One very important thing is that one can implement phase linear filters using symmetrical FIR filters where there is no analog equivalent except the bessel filter which is base on poles rather than zeros.

Best wishes.

Abdelhalim

Thank you. I respect the answers and suggestions 

given by the expert researchers on this web-site. 

Every point you have presented rings true. 

I believe you are advising me

to utilize the Digital approach. 

I spent much of my career

developing digital analysis techniques

and I certainly agree with you

about the effectiveness of DSP. 

""" the steepness of the filter transitions

from the pass band to the stop band """

are truly major points that I measure. 

However,  as I pursue my own understanding 

of  how  Analog Filters function 

My current project has been, and is now,

(1)   focused on developing Analog Filters, 

(2)   to compare Analog results with Digital results.  

My work is presented at the linked web-pages. 

I want to move forward from this stage 

to do verifying comparisons with DSP methods. 

My last post presented an "experimental method" 

which might help me achieve my goal. 

One question that 'might' be central to my understanding 

related to the measured degradation of the DSP signal 

when it passed through Digital-to-Analog Conversion.  

In this Experimental Process,

I hope to objectively compare the 'before' and the 'after' results. 

I expect the DAC to

(1)   introduce Third Harmonic distortions 

(2)   observe Spread-of-Passband distortions.  

Looking at the DAC output signal I expect 

(1)   to measurable   Increase   in Third harmonic. 

(2)   to observe the original digital Pulse square-wave

as it   Degenerates   into a Gaussian passband    having

    (a) a central peak

    (b) wider-than-original passband

         at -3dB, -12dB, -24dB, -48dB ( for example). 

Your  short and clear tutorials have been helpful in that

they forced me to  re-think, and re-state my 'problem'. 

Estoy un hombre viejo and my ideas change more slowly 

than the mind of a young man.  

Lutz  touched  on my goal  in his phrase :

""" we want to compare the performance capabilities

of both analog and digital filter structures """ 

Now, thanks to your kind help ,

I have a better "Experimental Process"

with which to work. 

A good Teacher pushes the Student

to re-think his own problem and methods.  

Glen 

http://www.geocities.ws/glene77is/

Our Next Project is about

Analog All-Pass Filters used for

Analog Audio Narrow Pass-Band Filtering,

is posted at :

http://www.geocities.ws/glene77is/

http://www.geocities.ws/glene77is/GC_ET_APC_.html

https://www.researchgate.net/post/What_is_a_proper_Analysis_of_this_Variant_All-Pass_Filter_How_does_the_Band-Pass_function_occur4

Original Question was:

What is a proper Analysis

of this Variant All-Pass Filter ?

How does the All-Pass Band-Pass function

occur in these All-Pass circuits ?

Engineering textbooks say that All-Pass Filters

CANNOT be used to generate Pass-Band Filters

nor Low-Pass, nor High-Pass Filters.

However,

we designed Analog All-Pass Audio Filter,

using 12 Lo-Pass and 12 Hi-Pass stages

which produce a Band-Pass f(0)

and Dual-Notches around the f(0).

What is a proper Analysis of this Variant Analog All-Pass Filter ?

How does the Band-Pass function occur ?

**************

Fellows,

Thank you for the encouragement

to congeal this project

into a somewhat formal "paper".

Attached is our "Pre-Paper" to date,

titled "AFX_PROJECT.PDF".

It is not very formally written,

but it is about the major part

of our research project to date.

We revised and date-time stamped the file.

It was uploaded to the RGN

At the Beginning of this "Question"

up at the TOP of this RGN web-page.

I am only an old apprentice

who was caught tinkering

in the Master's Workshop .

Glen Ellis

Attached is our "Pre-Paper" to date August 08, 2021, with frequent updates.

- - - Analog Phase-Filters

for Narrow Band-Pass Audio applications.

Just in case you want to save some time on the digital filter design, here's some code that might help (you just need a 'C' compiler)

Code Bilinear Transform Calculator

Preprint Second-order Bilinear Transform Calculator

Thanks , Mark.

I spent four decades creating and deciphering

relational databases and SQL code.

You and I both believe the digital method works wonderfully well.

However, my goal has always been stated as an Analog Adventure working its way from

(0) Old School Analog Active Filters

into

(1) Analog Phase-Filtering and

(2) Analog Non-Resonant Active Filters.

...

In the few years remaining,

I have released our energies

into exploring where Analog Active Filters

"Coulda, Woulda, Shoulda" have gone

had it not been for the digital revolution.

I have listened to Morse Code

on the Ham Radio bands for 63 years,

and that is part of what drives me forward !

...

Your prior advice and insights

have been very helpful,

and your projects which I have perused

indicate clear thinking and amazing talent.

We only do what we have opportunity and talent

and appreciation to do

... my aim is towards the Analog.

... BTW, as 'credit', I I included your name and

that of Dr. Josef Puncochar in my latest project,

which was just posted, as a pre-paper.

The chapter is named "AFX_pFilter_3_APC.pdf" and it was uploaded with all ten chapters,

long with the full paper.

...

I am still trying to wrap my head around

your SPICE of mechanical phenomena.

Truly, you have a young man's imagination !

...

your old colleague, Glen Ellis

Thanks Glen. I just have an appreciation of the weird, and like to marry unlikely bedfellows.

If you want to see something really weird, try this:

Preprint Abusing the Fermi-Dirac Distribution Equation in SPICE to Mo...

These little toys are so lightweight, you can use them to model stuff as complex as an ALU, and SPICE loves them.

We have developed TWO PROJECTS related to

- - - Analog Phase-Filters in Narrow Band-Pass Audio applications.

...

...

(1) We have developed a "AFX" Basic Band-Pass Narrow Filter Project

project with much tutorial published on this RGN page.

(2) Also, continuing our work,

we have developed "AFC" ...t All-Pass Band-Pass Narrow Filter Project

presented on RGN at :

https://www.researchgate.net/post/We_examine_a_Variant_All-Pass_Band-Pass_Narrow_Filter_circuit

Attached above is our "AFX" "Pre-Paper" current. .

The "AFX" Basic Band-Pass project summarizes previous developments .

We are continuing our experimental research into All-Pass Band-Pass Filters.

The next project is "AFC_All-Pass_Band-Pass_Filter_Project" :

devoted to All-Pass-Filters used to develop radical Band-Pass Circuits.

https://www.researchgate.net/post/We_examine_a_Variant_All-Pass_Band-Pass_Narrow_Filter_circuit

Attached in our Project Paper for our All-Pass Phase-Filter Project

dated Nov 26, 2021 (211126) . More documentation is being added.