Gearslutz.com - View Single Post

Ben B · 25th June 2009

I'll admit that I haven't read through this entire thread, so forgive me if this is a redundant post.

I've never tried to null one EQ against another, but this thread made me curious. My "go-to" plug-in EQ is the Filterbank E6 for most general duties. I like it due to its flexibility (e.g. the ability to adjust peak, slope, and dip in the shelving bands, the LPF/HPF features, and other standard parameters in the peaking bands).

So, out of curiosity, yesterday I decided to see if I could get it to null against the DigiRack EQIII. I routed pink noise to two aux inputs, with the E6 inserted on one and the EQIII on the other. I also inserted a Trim plug-in on both auxes, so that I could easily reverse the polarity of either one.

At this point, reversing the polarity of either channel produced a negative infinity null in PT, although I hadn't made any EQ adjustments yet. PT reports 0 samples of latency for both plugs, so all is good here.

Next, I set both EQs to give an equivalent boost at 1kHz, with a Q of 1.0. Reversing the polarity of either aux at this point did not produce a null, although I was not suprised by this. I assumed that it would require some more rigorous matching of center frequency, gain, and Q values to produce the desired null.

So at this point, I began making very fine adjustments, using the smallest amount of control resolution possible on both plug-ins. When making these adjustments, it is easy to hear when you're getting more cancellation or less cancellation, and likewise it is easily visualized on a meter.

Despite my best efforts, I could only produce a "null" (misnomer) of about -54dB compared to the combined level of both auxes mixed together "in-phase."

So in this case, something is going on which prevents a complete null. The only thing that comes to mind is possible differences in response curves in the "skirts" of the peaking bands, below the 3dB down points from where the Q values are obtained.

Can anyone else think of something I might have overlooked?

Thanks!

-Ben B

25th June 2009	#336
Ben B Lives for gear Joined: Aug 2007 Location: USA Posts: 1,336	I'll admit that I haven't read through this entire thread, so forgive me if this is a redundant post. I've never tried to null one EQ against another, but this thread made me curious. My "go-to" plug-in EQ is the Filterbank E6 for most general duties. I like it due to its flexibility (e.g. the ability to adjust peak, slope, and dip in the shelving bands, the LPF/HPF features, and other standard parameters in the peaking bands). So, out of curiosity, yesterday I decided to see if I could get it to null against the DigiRack EQIII. I routed pink noise to two aux inputs, with the E6 inserted on one and the EQIII on the other. I also inserted a Trim plug-in on both auxes, so that I could easily reverse the polarity of either one. At this point, reversing the polarity of either channel produced a negative infinity null in PT, although I hadn't made any EQ adjustments yet. PT reports 0 samples of latency for both plugs, so all is good here. Next, I set both EQs to give an equivalent boost at 1kHz, with a Q of 1.0. Reversing the polarity of either aux at this point did not produce a null, although I was not suprised by this. I assumed that it would require some more rigorous matching of center frequency, gain, and Q values to produce the desired null. So at this point, I began making very fine adjustments, using the smallest amount of control resolution possible on both plug-ins. When making these adjustments, it is easy to hear when you're getting more cancellation or less cancellation, and likewise it is easily visualized on a meter. Despite my best efforts, I could only produce a "null" (misnomer) of about -54dB compared to the combined level of both auxes mixed together "in-phase." So in this case, something is going on which prevents a complete null. The only thing that comes to mind is possible differences in response curves in the "skirts" of the peaking bands, below the 3dB down points from where the Q values are obtained. Can anyone else think of something I might have overlooked? Thanks! -Ben B