low pass filter

does anyone else have a problem with the low pass filter sounding digitally distorted when you get around a setting of 30?? im using jjos 1 v 4.99. any help would be appreciated.

yeah, some samples doesn't sound too good if you filter them down. i noticed that too.

Ive noticed distortion with low cut off points and high resonance settings, i need to compare if its more prominent than in OS1, we are missing the filter attenuation parameter in os2 though.

yea see the funny thing is i have no resonance. its not a super huge deal theres always a work around ... was just wondering if any other users experienced the same thing...

i'm getting digital distortions wothour resonance, too. but only with some samples.

there's a parameter called "ATT" next to "Freq" ans "Reso" that let you lower the signal that goes in the filter to avoid distortion

cool! didn't know what this one was for. but i think it's missing in os2?

could be that the sample was MP3 and has been converted to Wav? once its been mp3 it will always have the properties of a MP3. that could cause the low end to sound "corrupt"

no it has nothing to do with which sample it is. the only requirement is that the sample has enough bass or mid-bass.

There's a bug in the 1k's filter that causes it to clip, even without resonance. I've noticed it in Akai OS as well as JJ OS1. Theoretically it shouldn't happen, since the filter (without resonance applied)is supposed to subtract parts of the signal - thereby reducing the volume of the sample. But it does.

And attenuate sucks as a workaround, because then you have to attenuate everything else to keep the levels relative. All of a sudden you're in mix mode - and you haven't even finished the beat...

I hate filtering kicks on the 1k because of this.

have u tweaked the value 'att' that attenuates the filter depth? if ever i get a sound that just cant get smoothly through the filter - its offen because the effects are too deep for the rather lame filters.....i love 'em but they do distort sometimes......try tipping the 'att' value down 6dB....u will hear a slightly softer cut-off curve but thats the point, it'll save the filter from crunching.

let me know if this helps, like i said, its not ideal but some sounds will be too deep to get the values u are after without attenuation.

pccce

mp3 wrote:There's a bug in the 1k's filter that causes it to clip, even without resonance. I've noticed it in Akai OS as well as JJ OS1. Theoretically it shouldn't happen, since the filter (without resonance applied)is supposed to subtract parts of the signal - thereby reducing the volume of the sample. But it does.

mp3 learn your mathematics. It is entirely possible for you to apply a low pass filter and end up with higher sample points within your sample.

well of course depending on the filter design. so school me. if you don't consider this behavior a bug, then name another filter that will clip under such circumstances.

I will not name another filter... but I will show you some math.

1. Start off with a square wave.
2. Remove any DC offset.
3. Normalize it. (At this point all values should be close to 0dB, if it is not an ideal square wave.)
4. Filter the waveform with a lowpass. You should notice that the lower your cutoff frequency the lower the number of harmonics - the more squiggly the wave gets.

http://en.wikipedia.org/wiki/Image:Synthesis_square.gif

5. Observe clipping.
6. Post a reply saying thank you for schooling me.

Note: If you're going to be condescending, at least try to be correct. You only look like a fool if you're an ass and wrong too.

Ok class in session. Sit down little man.

You picture has about zero to do with filters, or even subtractive synthesis. As a matter of fact, its a demonstration of how to create a square wave from a series of sine waves. Additive synthesis. Not subtractive synthesis. Next time read the fine print.

(By the way, squiggly = higher frequencies. So the your description (clipping as you lower cutoff - due to waveform ripple) makes about zero sense.)

Now, the phenomenon you thought you were describing (and illustrating - even though your description didn't fit your illustration) is called pass-band ripple. Your first assignment is to find a picture of it - since you're so skilled at that.

Of course, pass-band ripple is a reality of any filter. Left unaccounted for (as it seemingly is in the 1k), it can and will cause clipping. Clipping which sounds good in analog, horrible in digital. Any good digital filter designer would account for it - so again, it depends on the filter design - you won't name another digital filter (designed for audio) with the clipping behavior, because you can't, because there are none.

Now, if you design a filter, and you don't account for pass-band ripple, and it clips, then that would be considered a bug in the software. Something that needs to be fixed. Clear? Good, now go away, unless you have something constructive to add.

Anything to add?