Question regarding Faust for DSP

On Sat, 26 Jul 2014 21:18:37 -0700, Mauritz Jameson wrote:

>>First, it will help your hit rate on this sort of thing if you SAY that
>>it's a homework problem.
> 
> It's not. It's for work. The problem was written up by a professor, so I
> guess that's why it's written the way it is. I'm not looking for a
> solution.
> Just some help interpreting the problem, because I'm not familiar with
> Faust syntax. I've looked up the Faust manual, but that didn't get me
> anywhere.
> 
> So I'm looking for somebody who would be kind enough to help me
> interpret the problem and 'translate' it into layman's terms or at least
> something that doesn't require the need to study the Faust manual.

Is your (or perhaps your boss's) goal to become proficient in Faust, or 
to lay your hands on some 7th- and 9th-order allpass filter designs?

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

> Is your (or perhaps your boss's) goal to become proficient in Faust, or 
> to lay your hands on some 7th- and 9th-order allpass filter designs?

It's not my goal (or my boss's) that I become proficient in Faust. I have been asked to implement some filters. I just don't get the specification/requirements. So I'm asking for somebody (who is familiar with Faust) to help me understand the requirements. 

To summarize:

- Boss: Implement a Zyquinox 
- Employee: Huh? What is that? Please explain.
- Boss: Implement a Zyquinox 
- Employee: Ok boss. Let me look it up.
- Internet: A Zyquinox is a Xapidodo
- Employee: Let me find somebody who knows what a Zyquinox is
- Comp.dsp: A zyquinox is just a low pass filter
- Employee: Ahhh...Now I understand what I need to do

On Sun, 27 Jul 2014 11:09:50 -0700, Mauritz Jameson wrote:

>> Is your (or perhaps your boss's) goal to become proficient in Faust, or
>> to lay your hands on some 7th- and 9th-order allpass filter designs?
> 
> It's not my goal (or my boss's) that I become proficient in Faust. I
> have been asked to implement some filters. I just don't get the
> specification/requirements. So I'm asking for somebody (who is familiar
> with Faust) to help me understand the requirements.
> 
> To summarize:
> 
> - Boss: Implement a Zyquinox 
> - Employee: Huh? What is that? Please explain.
> - Boss: Implement a Zyquinox
> - Employee: Ok boss. Let me look it up.
> - Internet: A Zyquinox is a Xapidodo
> - Employee: Let me find somebody who knows what a Zyquinox is
> - Comp.dsp: A zyquinox is just a low pass filter
> - Employee: Ahhh...Now I understand what I need to do

If all you need is a low-pass filter, why do you need Faust and why are 
you looking at code for it that appears to implement an all-pass filter?

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

>If all you need is a low-pass filter, why do you need Faust and why are
>you looking at code for it that appears to implement an all-pass filter? 

Are you serious?

It you bother to respond to my posts, and you're able to provide some insight, why not help a fellow DSP guy instead of wasting time on detours with irrelevant questions. 

I really don't get the attitude in this newsgroup. If somebody asked a DSP question and I was able to help, I would provide that help regardless of whether or not the posed question had to do with homework, personal interest or professional work. 

And a link to the Faust tutorial? Really? Been there - done that.

Anyway - thanks...

On Sun, 27 Jul 2014 13:03:12 -0700, Mauritz Jameson wrote:

>>If all you need is a low-pass filter, why do you need Faust and why are
>>you looking at code for it that appears to implement an all-pass filter?
> 
> Are you serious?
> 
> It you bother to respond to my posts, and you're able to provide some
> insight, why not help a fellow DSP guy instead of wasting time on
> detours with irrelevant questions.
> 
> I really don't get the attitude in this newsgroup. If somebody asked a
> DSP question and I was able to help, I would provide that help
> regardless of whether or not the posed question had to do with homework,
> personal interest or professional work.
> 
> And a link to the Faust tutorial? Really? Been there - done that.
> 
> Anyway - thanks...

No one here knows Faust.  So we can't help you with Faust.

You have said you don't want to know Faust, you just want to design a 
filter.  So I'm trying to find out what kind of filter you need, so that 
I can help you design it.

Just about all the regulars here can help you design a filter, using 
tools that are common in the DSP community rather than some bizarre and 
esoteric one-off.  As a side-benefit, you may end up with skills on a 
tool that'll actually get you a job if it shows up on your resume.

So yes, I'm seriously interested in helping you to meet your real goal, 
just as soon as I know what it is.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

On Sun, 27 Jul 2014 13:03:12 -0700 (PDT), Mauritz Jameson
<mjames2393@gmail.com> wrote:

>>If all you need is a low-pass filter, why do you need Faust and why are
>>you looking at code for it that appears to implement an all-pass filter? 
>
>Are you serious?

It's a serious question that would help anyone who really wants to
help you to do so.   If you're serious about getting help, you might
consider that.

>It you bother to respond to my posts, and you're able to provide some insight, why not help a fellow DSP guy instead of wasting time on detours with irrelevant questions. 

If you really want some help from people who can help you, don't waste
their time with attitude and irrelevant information.   Answer their
questions so that they can help you.

Assuming you're serious about getting help.

>I really don't get the attitude in this newsgroup. If somebody asked a DSP question and I was able to help, I would provide that help regardless of whether or not the posed question had to do with homework, personal interest or professional work. 

I really don't get the attitude of people who complain about the
people trying to help them in this newsgroup.   They ask a DSP
question in a way that makes it difficult to answer the question in a
useful way or makes it unclear what they're really asking, and they
expect others to read their mind and drop all of their own homework,
personal interests, or professional obligations on their behalf.

So far your behavior appears to be seeking spoonfeeding, which
generally doesn't go over well here since dsp is a pretty deep topic
with a lot of nuances.   If you appear unwilling to do the background
work, then the likelihood of any provided help actually doing any good
is low, so many who could help don't bother or keep asking questions
until

a) the OP clarifies the question in a way that people can provide
useful answers

and/or b) it becomes clear whether the OP can actually put any
provided help to productive use.

I think some people are still evaluating a and b in this case.

>And a link to the Faust tutorial? Really? Been there - done that.

Then focus your questions to what you don't yet understand.   You
asked several times for help with "Faust" but with no specifics.
Don't be surprised that somebody points you to a Faust tutorial.

>Anyway - thanks...


Eric Jacobsen
Anchor Hill Communications
http://www.anchorhill.com

Ok. Here are some more specific questions:

Assuming that the following is a way to specify filters using Faust syntax, how do I translate the following into digital difference equations OR which page(s) in the Faust manual should I read to be able to translate the following into digital difference equations?


1. highpass_plus_lowpass(1,fc) = _;

2. highpass_plus_lowpass(3,fc) = tf2s(1,-1,1,1,1,w1) with { w1 = 2*PI*fc; };

3. highpass_minus_lowpass(3,fc) = tf1s(-1,1,1,w1) with { w1 = 2*PI*fc; };

4. highpass_plus_lowpass(5,fc) = tf2s(1,-a11,1,a11,1,w1)

with { 
  a11 = 1.618033988749895;
  w1 = 2*PI*fc; 
};

5. highpass_minus_lowpass(5,fc) = tf1s(1,-1,1,w1) : tf2s(1,-a12,1,a12,1,w1)
with { 
  a12 = 0.618033988749895;
  w1 = 2*PI*fc; 
};

6. highpass_plus_lowpass(N,fc) = switch_odd_even(N%2,N,fc) with {
  switch_odd_even(0,N,fc) = highpass_plus_lowpass_even(N,fc);
  switch_odd_even(1,N,fc) = highpass_plus_lowpass_odd(N,fc);
};

7. highpass_minus_lowpass(N,fc) = switch_odd_even(N%2,N,fc) with {
  switch_odd_even(0,N,fc) = highpass_minus_lowpass_even(N,fc);
  switch_odd_even(1,N,fc) = highpass_minus_lowpass_odd(N,fc);
};

8. highpass_plus_lowpass_even(N,fc) = highpass(N,fc) + lowpass(N,fc);

9. highpass_minus_lowpass_even(N,fc) = highpass(N,fc) - lowpass(N,fc);

10. highpass_plus_lowpass_odd(N,fc) = highpass(N,fc) + lowpass(N,fc);

11. highpass_minus_lowpass_odd(N,fc) = highpass(N,fc) - lowpass(N,fc);


I have been asked to rewrite [10] and [11], as for orders 3 and 5 in [2]-[5], to eliminate pole-zero cancellations. Does this mean that I have to rewrite [10] and [11] in the same way as it was done for the 3rd and 5th order filters in [2]-[5] ???

On 7/27/14 4:03 PM, Mauritz Jameson wrote:
>> If all you need is a low-pass filter, why do you need Faust and why are
>> you looking at code for it that appears to implement an all-pass filter?
>
> Are you serious?
>
> It you bother to respond to my posts, and you're able to provide some insight, why not help a fellow DSP guy instead of wasting time on detours with irrelevant questions.

whoa! dude!  it's

1. hardly irrelevant
2. germane to the issue of efficiently helping you design 9th-order APFs
4. germane to the issue of whether or not we want to bother helping you 
or not.  and to be forthright, if *you're* forthright and if the message 
from that is that we're helping you with homework or for paid work, we 
might say for the former with "learn-by-doing" and to the latter with 
"have your boss send me an email and i can discuss with him or her how 
much it might cost if they contract with me to do this specific job."

you see, we're not against giving someone new starting in the business 
of writing DSP code.  and we're not against helping some EE student grok 
some systems/signal/LTI/DSP concept or even a specific how-to-do 
problem.  i have, myself, answered many questions here on comp.dsp with 
far more detail to *help* the OP than many people would bother.  i'm 
doing that with this specific answer.

but if a poster makes me feel like it, i might just be thinking "why 
bother?"

see 4 points (i can count!) to refute your "wasting time on detours with 
irrelevant guestions" assertion.

> Anyway - thanks...
>

well, for *what*?



-- 

r b-j                  rbj@audioimagination.com

"Imagination is more important than knowledge."

On 7/27/14 9:18 PM, Mauritz Jameson wrote:
> Ok. Here are some more specific questions:
>
> Assuming that the following is a way to specify filters using Faust syntax, how do I translate the following into digital difference equations OR which page(s) in the Faust manual should I read to be able to translate the following into digital difference equations?
>
>
> 1. highpass_plus_lowpass(1,fc) = _;
>

what'sa "_"?

> 2. highpass_plus_lowpass(3,fc) = tf2s(1,-1,1,1,1,w1) with { w1 = 2*PI*fc; };
>

what'sa "tf2s"?


-- 

r b-j                  rbj@audioimagination.com

"Imagination is more important than knowledge."

On Sun, 27 Jul 2014 18:18:42 -0700, Mauritz Jameson wrote:

> Ok. Here are some more specific questions:
> 
> Assuming that the following is a way to specify filters using Faust
> syntax, how do I translate the following into digital difference
> equations OR which page(s) in the Faust manual should I read to be able
> to translate the following into digital difference equations?
> 
> 
> 1. highpass_plus_lowpass(1,fc) = _;
> 
> 2. highpass_plus_lowpass(3,fc) = tf2s(1,-1,1,1,1,w1) with { w1 =
> 2*PI*fc; };
> 
> 3. highpass_minus_lowpass(3,fc) = tf1s(-1,1,1,w1) with { w1 = 2*PI*fc;
> };
> 
> 4. highpass_plus_lowpass(5,fc) = tf2s(1,-a11,1,a11,1,w1)
> 
> with {
>   a11 = 1.618033988749895;
>   w1 = 2*PI*fc;
> };
> 
> 5. highpass_minus_lowpass(5,fc) = tf1s(1,-1,1,w1) :
> tf2s(1,-a12,1,a12,1,w1)
> with {
>   a12 = 0.618033988749895;
>   w1 = 2*PI*fc;
> };
> 
> 6. highpass_plus_lowpass(N,fc) = switch_odd_even(N%2,N,fc) with {
>   switch_odd_even(0,N,fc) = highpass_plus_lowpass_even(N,fc);
>   switch_odd_even(1,N,fc) = highpass_plus_lowpass_odd(N,fc);
> };
> 
> 7. highpass_minus_lowpass(N,fc) = switch_odd_even(N%2,N,fc) with {
>   switch_odd_even(0,N,fc) = highpass_minus_lowpass_even(N,fc);
>   switch_odd_even(1,N,fc) = highpass_minus_lowpass_odd(N,fc);
> };
> 
> 8. highpass_plus_lowpass_even(N,fc) = highpass(N,fc) + lowpass(N,fc);
> 
> 9. highpass_minus_lowpass_even(N,fc) = highpass(N,fc) - lowpass(N,fc);
> 
> 10. highpass_plus_lowpass_odd(N,fc) = highpass(N,fc) + lowpass(N,fc);
> 
> 11. highpass_minus_lowpass_odd(N,fc) = highpass(N,fc) - lowpass(N,fc);
> 
> 
> I have been asked to rewrite [10] and [11], as for orders 3 and 5 in
> [2]-[5], to eliminate pole-zero cancellations. Does this mean that I
> have to rewrite [10] and [11] in the same way as it was done for the 3rd
> and 5th order filters in [2]-[5] ???

I really think that if you need to stick with Faust (which you have said 
you don't), that this is not the right group for it.  If someone here was 
facile with finding filters in Faust, they would have come forward.  The 
resounding silence on the Faustian aspects of your question is telling me 
that it's not a tool you should use with any expectation of getting help 
-- and it's not like the regulars here aren't good with DSP-related 
design.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com