| Author |
Message |
moudi
Joined: Oct 07, 2006
Posts: 63
Location: Bern Switzerland
|
 Posted: Mon Jan 08, 2007 4:58 pm Post subject:
LPF and HPF filters |
  |
|
hi folks
i'm back from a 4 week long holiday trip in greece. 
had a great time with a lot of chucking of course!
i came across of two different problems. this thread is dedicated for the first one, second one will follow in another thread 
i had the problem, that changing the Q of a filter with a joystick pad produce some ugly kind of noise (white noise or crackle?).
i made the following little testcase for reproducing it:
| Code: |
SinOsc s => LPF f => dac;
while(true)
{
//set the gains very low.
//so we can be sure nothing overdrives.
s.gain(0.1);
f.gain(0.1);
s.freq(Std.rand2f(300,400));
//with a high filter-frequency,
//the behaviour appears more often.
f.freq(12000);
//if the Q is randomly changed every time,
//a huge crackle (or white noise?) appears sometimes.
f.Q(Std.rand2f(0.1,1));
0.5::second => now;
} |
can anybody confirm this behaviour, or is there a logical reason for this?
and... how can i avoid this?
regards
/moudi |
|
|
Back to top
|
|
 |
kijjaz
Joined: Sep 20, 2004
Posts: 765
Location: bangkok, thailand
Audio files: 4
|
| Posted: Tue Jan 09, 2007 4:43 pm Post subject:
|
|
|
i've just modified your patch..
so that instead of changing Q directly,
i smooth the Q value first (with another LPF)
and assign to f's Q..
it still create that same sound bursts.
i'm looking forward to hearing more about this also.
there is also another issue like this with these butterworth filters..
that has negative frequency.
it can create some kind of super-loud bursts also. |
|
|
Back to top
|
|
|
ge
Joined: Aug 13, 2006
Posts: 108
Location: Palo Alto, CA
|
| Posted: Fri Feb 02, 2007 10:23 pm Post subject:
|
|
|
Sorry for the delay in answering this.
I think the short answer to this is that LPF/HPF/BPF/ResonZ are not meant to be used with small values of Q (<1) paired with high center or cutoff frequency. We'll need to reexamine the code and analyze the filter when things blow up.
Is there a particular reason why such a combination is desired? |
|
|
Back to top
|
|
|
moudi
Joined: Oct 07, 2006
Posts: 63
Location: Bern Switzerland
|
| Posted: Mon Feb 05, 2007 4:35 pm Post subject:
|
|
|
hi ge
| ge wrote: | | Is there a particular reason why such a combination is desired? |
yes i think so. if you want that the filter takes no effect to the sound (like bypassing the sound).
for example i was used to configure a midi controller in a way, that if the knob is turned full on, the frequency of a lpf is that high, that no effect can be heard. so if you also want to control the resonance this way, it's possible that such combinations occurs.
in this specific case, i used the joystick of a gamepad to control the q.
in a neutral position of the joystick i got a 0 value for the q, and therefore this strange noises...
probably a workaround will be to just set the ugen to pass through
regards
/moudi |
|
|
Back to top
|
|
|
ge
Joined: Aug 13, 2006
Posts: 108
Location: Palo Alto, CA
|
| Posted: Wed Feb 07, 2007 10:26 am Post subject:
|
|
|
| Quote: | | yes i think so. if you want that the filter takes no effect to the sound (like bypassing the sound). |
This makes sense. However, the Q on LPF and HPF controls the resonance at the cutoff frequency, while not having as much effect on the rolloff in a global way (for LPF and HPF, it's still going to be around 12 dB/octave). For that reason, perhaps Q is possibly not the most clear name.
| Quote: | | probably a workaround will be to just set the ugen to pass through |
Or possibly crossfade with unfiltered version of the source?
Hope this helps. |
|
|
Back to top
|
|
|
|
Forum index » DIY Hardware and Software » ChucK programming language
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
