Home Theater Forum and Systems banner
1 - 6 of 6 Posts

·
Registered
Joined
·
3 Posts
Discussion Starter · #1 ·
Besides the difference in interpretation between EQ device implementation in REW, Q mostly, are all the PEQ's zero phase-shift throughout their operating curve? If not, how do they vary?

I'm noticing that any time there are filters that overlap each other, it seems to disturb the character of the sound over and above what just the EQ sum of the overlaps would suggest.

I know you should not necessarily take AutoEQ suggestions literally as there are many extra's of a very tight Q. You can eliminate quite a few of those, but almost always you still end up with a few filters that overlap each other, at least partially. Not necessary overlap that is all positive, or all negative, but little 'trimmer' ranges to smooth out correct peaks made by two adjacent equalizers both operating positive or negative. Anyway, where there are these overlaps it is often quite audible as kind of a phasing/hollow sound in that particular range, most especially in midrange and upper-midrange areas.

In just trying different EQ definitions, it *seems* that the FBQ2496 is less likely to be as sensitive to this, especially compared to GENERIC. So I'm wondering about the specs and comparative performance of the various emulations.

This is in a studio environment with B&W Nautilus 801N monitors, which are pretty revealing of this sort of thing. No subs. Any tips?

Thanks.

--Bill
 

·
Super Moderator
Joined
·
9,309 Posts
First, welcome to the Forum, Bill!

Frankly I can’t tell from your post if you’re talking based merely on filter modeling in the REW program, or based on actual listening experiences with actual equalizers set up from REW’s filter recommendations.

Just judging from the title of your post, “Various EQ’s performance in REW,” it seems you’re confusing the fact that REW is merely a measurement platform and as such has absolutely nothing to do with the “performance” of the equalizer in use, whatever that may be: You can’t reasonably expect the same result plugging REW’s filter recommendations into an Alesis vs. a Pultec.

This seems painfully obvious, which makes me think perhaps I’m missing something?

Regards,
Wayne
 

·
Registered
Joined
·
3 Posts
Discussion Starter · #3 ·
Thanks Wayne.

I am exporting the Filters Impulse Response and using it in J Rivers Media Center's convolution engine. A house curve is employed similar to your sharp knee method, though tweaked a bit, and not quite as much downward slope above 2k as the BK curve. It's a full range system, 2 channel.

So I am in essence using the result of the internal EQ in REW by itself via convolution, thus my questions about the type of EQ they are, and possible behavioral differences one to another.

Does that help clarify things?

--Bill
 

·
Premium Member
Joined
·
6,711 Posts
All of the equalisers REW emulates use IIR filters, their phase shift is linked to their amplitude response - one way to see that more directly is to re-import an exported filter response so you can see the amplitude and phase of the response in the SPL and Phase graph. The level will be high on import, you can rescale that using the controls on the Impulse graph or by subtracting an offset using the controls on the SPL graph.

REW currently generates more high Q filters than it should using the Generic setting (a side effect of trying too hard to avoid increasing dips), beta 22 addresses that, I'll release it in the next few days.
 

·
Registered
Joined
·
3 Posts
Discussion Starter · #5 ·
Great information, John. Thanks.

So it would seem that regardless of Q setting, the phase shift would be directly related to the amplitude of the eq curve at a specific frequency. If that is absolute, then phase shift + or - would be directly summed just like amplitude to get the resulting value at the given frequency, overlapped curves or not. Correct? Even if the eq center points weren't consistent.

Taking the question further, would it be safe to say that in reality, the most these filters can do is approximate the characteristics of the equivalent adjustment in specific devices (listed in hardware options or not), and that the actual results as heard could be quite different than the graph might suggest.

Do you have any eventual plans to more precisely emulate the true characteristics of the hardware listed in the EQ choices? Something along the lines of certain VST type plugins that emulate Pultec, Neve, SSL and other board EQs? Or is that just too far beyond the scope of REW?

Thanks for a great program!

--Bill
 

·
Premium Member
Joined
·
6,711 Posts
The emulations are precise.

The phase shift of an IIR EQ filter depends the amount of gain or cut and the Q, the higher the Q the narrower the spread over which the phase changes and the greater the gain or cut the larger the maximum excursion of the phase shift, but far enough either side of the filter the shift returns to zero. Phase shift from multiple filters does simply add. Here are a few examples to illustrate, phase is the lower trace in each case and shown on the right hand axis. The first 3 are boost filters, cut filters look the same but mirrored, per the last example - the sum of the last 2 filters would have flat amplitude and zero phase shift, although that isn't always true - it depends on the definition of Q used for the filter. Note that even a 9 dB filter has a fairly modest phase shift, peaking at less than 30 degrees. A 20 dB filter would have peak phase shift of about 55 degrees.

Text Line Pattern Plot Design

Text Line Plot Slope Pattern

Text Line Plot Pattern Diagram

Text Line Plot Pattern Design
 
1 - 6 of 6 Posts
Top