Hi,
just dia a couple of patches to get the profile of filters by testing
them - linear + log versions. I had some questions to see if anyone is
interested in this:
this one is already good enough to get an idea
I think it works precisely, but what's your opinion?
I might do later one patch with both linear and logarythmic versions,
but was too much work for now
settings of env~. I might try out later to see if with other settings it runs faster.
there's a better formula for this
I couldn't find out how to connect frequency values and table indexes. you're welcome to patch it, if you want.
Btw, anyone has good links/references to filter theory courses/papers?
Hope this is useful,
João Pais
Hallo, João Pais hat gesagt: // João Pais wrote:
just dia a couple of patches to get the profile of filters by testing them - linear + log versions. I had some questions to see if anyone is
interested in this:
...
Btw, anyone has good links/references to filter theory courses/papers?
Miller's book is pretty nice. The docs also include filter-graph abstractions to profile filters similar to yours but a bit more powerful in that they also graph the phase response.
Other than that, IMO Julius O. Smith's DSP Guide is the best reference as far as "musical" filters are concered.
Some nitpicks: your patches use some objects not available to me, namely [f+] and [lbang]. I assume they are trivial objects, why didn't you use use the standard [loadbang] and a [f][+] counter? Also the signal sends etc. are global, which makes it impossible to use the objects to compare two filters side by side.
Frank
Miller's book is pretty nice. The docs also include filter-graph
abstractions to profile filters similar to yours but a bit more powerful in that they
also graph the phase response.
do you mean the examples around fig8.12 from figs.zip (filter section)? I
just tried out a couple minutes with them, and they looked made to do the
pictures, and not to serve as a general patch.
Some nitpicks: your patches use some objects not available to me, namely
[f+] and [lbang]. I assume they are trivial objects, why didn't you use use
the standard [loadbang] and a [f][+] counter? Also the signal sends etc. are global, which makes it impossible to use the objects to compare two
filters side by side.
these are my abstractions, you get them in pd-ext/jmmmp. lbang I use while
working on the patch, because it allows me to rebang, and f+ gives me more
options easily than just the "normal" counter. I might try a diferent
approach to the signal sends, involving $0 and copy-paste.
Hallo, João Pais hat gesagt: // João Pais wrote:
Miller's book is pretty nice. The docs also include filter-graph
abstractions to profile filters similar to yours but a bit more powerful in that they also graph the phase response.do you mean the examples around fig8.12 from figs.zip (filter section)? I just tried out a couple minutes with them, and they looked made to do the pictures, and not to serve as a general patch.
Actually I meant H10.measurements.pd in the Pd docs, section 3.audio.examples.
Some nitpicks: your patches use some objects not available to me, namely [f+] and [lbang]. I assume they are trivial objects, why didn't you use use the standard [loadbang] and a [f][+] counter? Also the signal sends etc. are global, which makes it impossible to use the objects to compare two
filters side by side.these are my abstractions, you get them in pd-ext/jmmmp.
It was much faster for me to just replace "lbang" with "loadbang".
lbang I use while working on the patch, because it allows me to rebang,
You probably know this, but it's possible to rebang subpatches (and abstractions, but only all of them or via namecanvas) by sending them a "loadbang" message.
and f+ gives me more options easily than just the "normal" counter.
which you didn't seem to use. :)
I might try a diferent approach to the signal sends, involving $0 and copy-paste.
I'd recommend to use $1, so people can specify their own target names.
Frank