It’s a good source, but bear in mind that it doesn’t get you new kinds of basic 
filters, and its strength is in designs that give you good behavior while 
time-varying the parameters. In other words, it’s more important for synth 
filters with frequency controlled by envelope generators than for static 
filters for EQ. But a second-order Butterworth lowpass filter is still a 
second-order Butterworth lowpass filter, whether you get it from Robert’s 
cookbook formulas or a “VA” design.


> On Nov 29, 2023, at 3:07 PM, Jens Johansson <[email protected]> wrote:
> 
> One document I stumbled upon while googling is "The art of VA filter design", 
> by Vadim Zavalishin (it's apparently a book he published for free). If I can 
> even wrap my head around some of that stuff, might it be a recommended good 
> place to start I wonder? 
> 
> Cheers,
> J
> 
> 
> 
> 
> On Wed, Nov 29, 2023 at 11:44 PM Jens Johansson <[email protected] 
> <mailto:[email protected]>> wrote:
>> Thank you for all your great advice so far! It gives some really good hints 
>> which trees to consider barking up, so to speak. Well, I first have to 
>> become multicellular, and evolve eyes, then find the trees, but you get the 
>> idea.
>> 
>> I enclose the full schematic if anyone's curious, my guess was that the 
>> first three filter stages (C1R1, C3R2R3, and C4UA2) could be approximated 
>> with stuff from the cookbook. Just loading the schematic into LTSpice seemed 
>> to confirm this somewhat. Also the stages after this more complex "fifth 
>> order" section (C9R7?, C12C13R11R12C14A) looked like they either are 
>> inconsequential or kind of map onto the more normal filter cookbook 
>> topologies too. The "N" ground is the midpoint between the two power rails 
>> for the opamps. The "ground ground" is the negative rail. 
>> 
>> (It's not the greatest distortion box and also not the worst. Peak 1990 
>> technology. On the other hand my hope is that a simulation could run in a 
>> couple of hundred CPU cycles and not need either bluetooth, AI or a built in 
>> web browser. I was planning to just slap the code on github if I ever got it 
>> working. I would use JSFX because it's so easy to prototype stuff in this 
>> language)
>> 
>> Cheers,
>> J
>> 
>> 
>> 
>> On Wed, Nov 29, 2023 at 10:51 PM brianw <[email protected] 
>> <mailto:[email protected]>> wrote:
>>> On Nov 29, 2023, at 1:49 PM, brianw <[email protected] 
>>> <mailto:[email protected]>> wrote:
>>> > On Nov 29, 2023, at 1:28 PM, robert bristow-johnson 
>>> > <[email protected] <mailto:[email protected]>> wrote:
>>> >> On 11/29/2023 3:09 AM EST Jens Johansson <[email protected] 
>>> >> <mailto:[email protected]>> wrote:
>>> >>> and simulate the below circuit just with simple(r) feed-forward and 
>>> >>> feedback algorithms, or is it so complex that I would have to take the 
>>> >>> step how to learn to deal with those "Wave Digital Filters"?
>>> >> 
>>> >> I am curious about C6.  What does that "4n7" mean?  Is that just a typo 
>>> >> and it's just another 47 nanofarad cap?  And what does R7 connect to?  
>>> >> What is "N"?  Is it another kind of ground?  Or is it some buss 
>>> >> somewhere?
>>> > 
>>> > "4n7" is a shortcut that engineers like to use for "4.7nF" - the standard 
>>> > is to move the unit to where the decimal place belongs, and that saves a 
>>> > digit.
>>> 
>>> It's also the case that a decimal point '.' can easily disappear on a 
>>> printed or hand-written schematic, so placing another symbol in place of 
>>> the decimal makes it much more visible.
>>> 
>>> Brian

Reply via email to