Hello everyone,

I’ve heard a lot about the importance of good phase characteristics in an equalizer, but I’ve never heard anyone say how phase characteristics - good or bad - are sonically manifested.

I have a TDM 1/3-octave graphic EQ that I recently noticed something unusual. I was doing sound for a church retreat with a simple system of a baby Mackie board and JBL Eons (their stuff, but I usually pack my EQs no matter how simple the gig – just in case). I had a particular slider pegged as potential feedback location (something in the 250-400Hz range, if I recall) and I noticed that I could move the slider from maximum cut to zero with virtually no sonic change in the timbre of the voice of the woman who was speaking. (I was directly in front of one of the speakers, maybe 40-50 ft. away.)


Yes, there was “some” change, but it was very subtle. If I hadn’t been listening for it I very well might not have noticed all. I’ve typically seen this only with parametric EQs with filters set for very narrow 1/6 – 1/8 octave bandwidth – certainly not in other 1/3 octave equalizers I’ve used (none of which are better than the TDM).

So I’m wondering, is an example of good phasing characteristics, or does this EQ simply have tight and well-controlled filters?

In any event, if someone could let me know how to audibly determine an equalizer’s phase performance I’d sure like to know.

Thanks,
Wayne A. Pflughaupt


Reply posted by jack arnott on June 16, 2002

Hello Wayne:
These EQ's do have a very tight pattern. They are not like the Ashly which will give a very wide "V" when pulled out.

One thing, did you check if your range was set to 6 or 12 db cut/boost. It is a bit counter intuitive with the EQ in out switch.

Jack


Reply posted by TDuoos on June 14, 2002

Hi Wayne

I personally have had lil gremlins in my dual TDM 31's and live less than 100 miles from the factory... A great looking house with Tim’s TDM on the ground floor... way up in the hills of Hillsboro, OR.

I have seen piles of dual slider pc boards loaded with defective sliders on there way to scrap... And...found out that the early duals used a .5amp fuse that usually blew when least expected so if yours saz .5a....up it to 1amp per Tim at TDM.

Also make sure the pc spacer post at or near the two large standing power sup caps on the power sup/in-out board is making GOOD ground connect to the external chassis ... without it your can pick up some very good RF.

Call Tim and have him send you a couple of replacement slider pots.... We also have used other TDM products for the last ten years....email me if you want to chat.

Good luck.

TD


Reply posted by michael on June 13, 2002

Wayne,

You certainly have the right idea.

You can test this with Smaart Pro computer program. It provides both freq. and phase. Two inputs and compare a pink noise source and secondly, send it through the EQ. Match up the amplitude and use the transfer function of the program to compare. When you move a slider, the Smaart Pro should show you the change in freq. and in another window, the change in phase.

I did this with a mixer and kept printed records of it and I found this process to be effective. Try what you like and document it.

You will require:
1. a pc with stereo input
2. a pink noise generator
3. Smaart Pro software

plus, when I did it, I printed the results on paper and I didn't realise I could take "screen shots" which would have been much prettier and in color too


Reply posted by Bink on June 13, 2002

Wayne -

Maybe the EQ filter is faulty. Of the several unwell EQs I've seen in my years once the problem was that a single filter failed to cut at all -- even though its boost function was still operational.

You can check this with pink noise. Does the filter affect pink? If so, then JR's assessment was right and the person's voice had precious little power in that 1/3rd octave range.

-Bink


Reply posted by John Roberts on June 13, 2002

Perhaps that vocal didn't have much content in that frequency band...

No signal... no phase shift.

JR


Reply posted by Marty McCann on June 13, 2002

On the other hand since it was a church situation, if indeed there was a coloration perceived and the filter did not seem to offer much help when introducing a cut, you may have been in a situation where the actual placement circumstances of the loudspeaker is exciting a mode within the room itself.

This often occurs in churches where they have chosen to recess the loudspeaker into some sort of cavity (or tone chamber), but do not seal in the remaining opening around the enclosure. The combination of box, cavity, and opening (or port) can create a Helmholtz resonator effect, and esssentially color or bark acoustically. I say this because of the frequency range that you reported (250 Hz -400Hz). As a matter of fact, Don Davis himself used to refer to this condition as "The Dog House Effect."


Reply posted by Wayne A. Pflughaupt on June 14, 2002

Many thanks to all who replied. Michael, your response is intriguing – unfortunately I don’t have the proper PC or software.

I guess I shouldn’t have brought up the experience with the TDM – maybe that wasn’t a phase situation at all. It seemed only to distract from my real “quest.”

So can anyone tell me how to sonically recognize good or bad phasing characteristics in an EQ?

Regards,
Wayne


Reply posted by Andy Peters on June 14, 2002

Smart-ass answer: listen to them?

Seriously: different EQs have different phase responses because their filter circuits differ. Filter Q and rolloff determine the phase characteristics. I'm sure if I wasn't lazy, and I had my computer and MathCad and SPICE up and running, I could do some plots of varying filters and show the phase responses, because I don't remember which filter topologies have "more" phase shift than others (tho' due to minimum phase, one would expect the more complex filters to have more complex phase response).

Remember that phase is a frequency-dependent time delay, and if that phase is not linear, you get all sorts of weird time-domain effects -- ringing on non-sinusoidal signals, for example. (This is one reason why early CD players were considered to have a harsh, "digital" sound -- because the anti-aliasing filters were so sharp, phase weirdness at the top end would cause the filters to ring.)

Fun test: put a square wave into a filter and see what the edges look like on a 'scope.

Filter shape is one reason why KT EQs sound different from, say, a dbx or BSS unit. As to which is "better," that's in the ear of the beholder. Also, which manufacturer B may use the exact same filter topology as Manufacturer K, if B uses components with less-stringent tolerances, the filter response could vary widely from what one expects. Also, some filter topologies are less sensitive to component tolerances than others.

--a


Reply posted by Tom Danley on June 14, 2002

I would say that generally all filter shapes have an exactly corresponding phase shift (related to the shape) with an exception being some DSP and all pass filters.

If one has a loudspeaker who's response rolls off or has a bump, these undesired "filters" also impose a particular phase shift, correcting that with an EQ also corrects the phase as most electroacoustic "things" are "minimum phase" (meaning there is a specific phase change for a change in amplitude).

If one wanted to EQ a cancellation notch caused by two different acoustic paths (one delayed RE the other), one finds that this "filter" does not have the normal phase change with amplitude and if the two paths have equal amplitude, the notch is infinitely deep as well. Using an EQ to put a big peak there does fill in the sides some but while having little effect on the response, does add a large amount of phase shift at the EQ point. Since the phase shift is not corrected by the phase error of the problem filter, it is left in the signal.

Bottom line is that what you hear so far as phase usually depends a lot more on what kind of problem you are trying to fix rather than on what type of circuit was used to produce the filter. Non minimum phase things like multiple acoustic paths and so on cannot be fixed with EQ.

One does have to wonder that since loudspeakers generally do not preserve phase, how much does that obscure the electronic phase effects.

Cheers,

Tom Danley