Something is screwy with the library's shelving filter

Studiowaves
Studiowaves Member Posts: 453 Advisor

Here's a file showing the frequency response of two filter types. The top is a cascaded 2 pole filter and the bottom is the libraries shelving filter. If you switch between lp and hp it's pretty obvious the slope is starting way to early. I'm assuming the Fc should be down about 3 or so db, but his looks way off. Am I doing somethings wrong?

Note the bottom trace shows Fc. The first slope is a normal shelving control, the second factory slope is the curve at Fc.

Comments

  • errorsmith
    errorsmith Member Posts: 27 Helper

    i think the shelving filter from the lib works correctly.

    the cutoff freq of a shelving filter is in the middle of the transition zone from one plateau to the other. in order to achieve that that the cutoff frequency of the state variable depends on the gain (see mult of w/2 before it goes to the SVF). it needs to do so because the filter slope of the shelving filter at the cutoff freq cannot be steeper than 12 dB/Oct in case of the 2-pole shelving filter. so when you decrease the gain more and more the plateaus move further away from the cutoff freq. try this with other hi or lo shelving filters in your daw or a plugin. that is the normal behaviour. beware that usually gains are limited to say 15dB, which makes this "plateaus are moving away from cutoff" less obvious. but it is there.

    i attached a modified hi shelving. i deactivated the "mult of w/2", and i used a normal Res knob for calculating "2R". i find this more intuitive than 'Slope' when setting up a shelving EQ.

    I guess this is what you are after. You can compare the shelving to the LP. with very high negative gain these are the same.

  • Studiowaves
    Studiowaves Member Posts: 453 Advisor
    edited December 2022

    Thanks, Sweeping the cutoff down doesn't make sense as the cutoff is typically the start of the filter. Mixing instruments is in reality an add and subtract mixture. Changing the cutoff is confusing, This is more like it should be. Nice job!!!

    I do like the variable roll off idea on the edge of the cutoff, that's very useful. Not sure why there are separate inputs for plus or minus shelving, but I may not understand it. It seems a single input for plus and minus db and the variable roll off is all that's necessary. I'm not well educated on filters, would it be difficult to do the same with the high pass while we're on the subject. I hope it's not asking to much.

  • errorsmith
    errorsmith Member Posts: 27 Helper

    this way you can quickly change between hi and lo shelve. just connect gain to 'gain' instead of '-gain' and use the 'LS' out, voila you have 'the same with the high pass'.

    i guess the aim was to have a leaner library with only one 2-pole shelving filter for both cases.

Back To Top