Sorry for delay,
You know, I am all confused and I am starting to think I can't make this
change ;) despite the fact that technically everything is trivial.
On 08/18, Julius Smith wrote:
>
> Nice! I like the generalization of delay interpolation while keeping
> backward compatibility. I would only suggest a more foreground name for
> fb_xxx, such as fb_comb_common,
I'd prefer __fb_comb() but I won't insist,
> suggesting that it's a function shared by
> the other feedback-comb variations, and that it's a "helper" function not
> normally called by itself
Why not?
> (no mem protection, e.g.).
sorry, I don't understand this part...
> Then I would also add
>
> fb_f4acomb(maxdel) = fb_comb_common(de.fdelay4a(maxdel)) : mem;
Hmm... the extra ": mem" again? for what?
In any case I am not sure this makes sense... We also have fdelay[1-3]a,
we have fdelaylti/fdelayltv, we can have more "delay" helpers, I don't
think we should add the fb_comb variant for every dop.
> (I still follow the "define it when you really need it" rule myself.)
Exactly! see above...
--------------------------------------------------------------------------
But the main problem is comments. Let's start from the very beginning,
currently we have
// #### Reference
// <https://ccrma.stanford.edu/~jos/pasp/Feedback_Comb_Filters.html>
fb_comb(maxdel,N,b0,aN) = (+ <: de.delay(maxdel,N-1),_) ~ *(-aN) :
!,*(b0) : mem;
and as I have already mentioned the purpose of ": mem" is not documented.
But there is another thing I didn't notice before: the link above describes
the difference equation as
y(n) = b0*x(n) - aN*y(n-N);
while this code (without the extra/undocumented mem) implements
y(n) = b0*x(n) + aN*y(n-N);
(just in case... of course, both versions make sense! my only concern is
the documentation).
So. If I do (as we discussed before)
fb_xxx(dop,N,b0,aN) = + ~ -aN * dop(N-1) : *(b0);
fb_comb(maxdel) = fb_xxx(de.delay(maxdel)) : mem;
How can I document fb_xxx() ?
Should I try to fix the comment above fb_comb/fb_fcomb ? If yes, how?
--------------------------------------------------------------------------
Now. If I do
// note the -aN => aN change
fb_xxx(dop,N,b0,aN) = + ~ aN * dop(N-1) : *(b0);
then this version can refer to ~jos/pasp/Feedback_Comb_Filters.html,
just the comment should explain that "dop" possibly means the fractional
delay.
But then fb_comb() should be defined as
fb_comb(maxdel,N,b0,aN) = fb_xxx(de.delay(maxdel),N,b0,-aN) : mem;
not really nice from the code reading perspective. And it is still not clear
what should I do with the comments above fb_comb/fb_fcomb.
--------------------------------------------------------------------------
In short. sorry for the spam, let me repeat I am all confused.
And... sorry for the spam, I am going to ask another question ;)
Oleg.
_______________________________________________
Faudiostream-users mailing list
Faudiostream-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/faudiostream-users
