elena
Developer
Posts: 17
Posts: 17
|
Post by elena on Nov 9, 2018 12:02:35 GMT 1
Hi, I am working on a special synth of my design but I encountered a minor issue with the builtin Butterworth. Its Freq and Width parameters are FLOAT instead of VOLT, therefore they can be modulated/changed with a very slow update rate. I don't actually need to modulate them with an LFO, just with the synth controls, but as a result of that, everytime I play with the sliders to change these parameters while the synth is sounding, I get horrible noises and spikes (I tried to temporarily solve this by writing out a simple limiter but it is not "lookahead" therefore it fixes the annoyance only partially and anyway I don't like such a remedy - btw, where can I find a 64 bit limiter ??). Using a Slew module does not help because the update rate of these FLOAT parameters remains intrinsically slow.
Unfortunately I could not found any 64 bit bandpass filter with adjustable bandwidth.
The most similar thing I could find is an EQ filter in the TD package, but being it an EQ with adjustable gain I would have to "convert" it to BP by subtracting the original signal, and I have obvious gain inconsistencies... too long to explain...
Any 64bit bandpass filter with frequency and bandwidth adjustable at sample-rate ? Thanks !!!
|
|
Andrew
Developer
Posts: 110
Posts: 110
|
Post by Andrew on Nov 9, 2018 16:58:11 GMT 1
It's simple to do with any bandpass filter, like my TD_SV2 filter in bandpass mode, does require some math to do the scaling though, as having bandwidth described instead of resonance/Q is not very common, even less common for synths. Unfortunately I can't do that math out of my head, and still writing exams, so I'll only be able to assist like next week wednesday if your still stuck. If by chance you need 4 bandpasses in parallel(for building a formant filter), the TD_SVX4 is kinda built for that, the resonance isn't in bandwidth, but they are "zero peak" bandpasses which simply means no matter what the resonance setting, there will be no positive gain at the centre frequency.
|
|
elena
Developer
Posts: 17
Posts: 17
|
Post by elena on Nov 10, 2018 23:43:40 GMT 1
Hi Andrew (btw... thanks for answering my posts always !! You're sooo kind !!! ) Yeah I am working at a formant/vowel synth indeed, you got it ^_^ and I really love the result ! I started this simple little project because I needed some Uh Uh choirs in an arrangment of mine but could not find any I like around ...and obviously I ended up with something ways more versatile than planned LOL Initially I had discarded the option of using TD_SV2 because I did not read any "Q" or "BW" parameter listed, I thought it was fixed; but now that I think better (and that you pointed out that) a resonance without any positive gain as you just said could be the "key" to computing bandwidth. No problem with math, I'm quite experienced with such things LOL... I will experiment with your filter THANKS A LOT ! I simply had overlooked it
|
|
elena
Developer
Posts: 17
Posts: 17
|
Post by elena on Nov 14, 2018 13:22:45 GMT 1
Update: unfortunately only the builtin Butterworth BP does the job I need, and I have to say it does it tremendously well !! In the meanwhile I have tried EXTENSIVELY every filters I could, not just your TD_SV2, but everything including band rejects, EQs, everything... They all are not selective enough, and tend to resonate too much unfortunately. Believe me: I tried them all the ways I could figure out with accurate scaling of gain in function of resonance etc. With all filters I tried my synth sounds too much "synthetic" and tends to "shot" everytime I hit resonating notes (like most similar formant synths around), while with the builtin Butterworth I can achieve voices (soloes and choirs) of an astonishing realism !! A pity that there is not any Butterworth Band Pass with voltage-controlled parameters
|
|