Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[John L Nobile]

Correct me if I'm wrong, but my understanding is that when you use eq, you introduce a phase shift into the signal with analog gear. I've been wondering if this is the case with digital eq's. Crunching numbers as opposed to going through electronics.

[Barry Singleton]

True and yes.

The induced phase shift is not always a bad thing though. 

Barry.

[Ivan Beaver]

Correct me if I'm wrong, but my understanding is that when you use eq, you introduce a phase shift into the signal with analog gear. I've been wondering if this is the case with digital eq's. Crunching numbers as opposed to going through electronics.

If the problem you are correcting is an inherent problem with a driver (and NOT a reflection-timing error etc) then you are most likely ALSO correcting the phase response at that same freq.

So it is not always a bad thing.

Sometimes we use "out of band" eqing in order the help correct the phase response of a particular alignment.

It has little to nothing to do with the actual freq response-but does help correct the phase response.  This type of approach is always in the cut domain-never boosted.

[David Sturzenbecher]

If the problem you are correcting is an inherent problem with a driver (and NOT a reflection-timing error etc) then you are most likely ALSO correcting the phase response at that same freq.

I have seen people use Lake processing more then I have used it myself, so I have little first hand experience. But I am under the impression that Lakes do not adjust the phase for corresponding EQ curves. What is the mentality there?  I can see the validity in having crossover slopes that don't effect phase, but I can't necessarily get my head around the EQ section.

[Ivan Beaver]

I have seen people use Lake processing more then I have used it myself, so I have little first hand experience. But I am under the impression that Lakes do not adjust the phase for corresponding EQ curves. What is the mentality there?  I can see the validity in having crossover slopes that don't effect phase, but I can't necessarily get my head around the EQ section.

There are different types of eq filters IIR (normal ones) and FIR (the fancy ones that introduce time delay with their usage).

I will admit that I know next to nothing about FIR filters-have never used them-so cannot talk with any sort of knowledge about them.  But I would argue that people are often using them to try to "fix" a bad design.

I would argue that most people doing alignments don't even look at phase-know what it means-know what a good phase response looks like etc-or care about it-but they talk about it all the time ;(

I would also argue that if you have to use some of the "wacky" eq curves I have seen advertized, then you have one SERIOUSLY messed up system-and you are only "fixing" it in one seat-and making the others worse.

Is your system THAT messed up it needs THAT much correction?  WOW!  Impressive--but on what level?

But to some people it is all about a SINGLE location.  It should be about ALL locations.

[David Sturzenbecher]

I will admit that I know next to nothing about FIR filters-have never used them-so cannot talk with any sort of knowledge about them. 

Then why even reply? 




Sent from my iPhone using Tapatalk

[John Roberts {JR}]

Then why even reply? 




Sent from my iPhone using Tapatalk

To raise the level of the discussion.

There are different classes of filters and some digital filters do not exhibit the classic  amplitude/phase shift relationshil like analog filters do.

As several have mentioned if you are correcting a naturally occurring amplitude error, it probably also has an associated phase error, so the analog phase shift is often corrective too.

Some digital filters instead of frequency related delay (phase shift) exhibit a fixed delay.  There are applications for both types of filters but this is a little above my pay grade, as well as for most reading this (IMO).

Another reason I prefer to let speaker engineers design the crossovers.

JR

[John L Nobile]

The question I had was more to do with the results of eqing for tonal balance in a system as well as channel eq. The old analog vs digital thing. Something my ears can easily hear. Using eq for phase adjustments is a little beyond me. Sounds like I'd need Smaart or more to do that.

[Mac Kerr]

The question I had was more to do with the results of eqing for tonal balance in a system as well as channel eq. The old analog vs digital thing. Something my ears can easily hear. Using eq for phase adjustments is a little beyond me. Sounds like I'd need Smaart or more to do that.

All analog eqs and most digital eqs cause phase shift. It's how they work. Different filter topologies cause different amounts of phase shift. There are some digital eqs that can do tonal change without phase shift, or phase shift without tonal change that are used for very specific speaker response correction. The downside of these filters are the additional latency they add. Using FIR filters does require instrumentation, it is not for the faint of heart. They are generally used by speaker manufacturers in creating DSP presets for specific speakers or speaker combinations.

Mac

[Tom Danley]

The question I had was more to do with the results of eqing for tonal balance in a system as well as channel eq. The old analog vs digital thing. Something my ears can easily hear. Using eq for phase adjustments is a little beyond me. Sounds like I'd need Smaart or more to do that.

Hi
The phase shift that accompanies "normal" EQ including if done with fir's can exactly offset the phase shift also present in the bump or dip your trying to fix.  When (only when) this bump or dip is produced by a "minimum phase" device such as a single simple driver, the eq correction fixes both the amplitude part and the phase part.  In a minimum phase system, any change in amplitude has a corresponding change in phase response.

The complication is that when you have a bump or dip that is the result of two or more sources or a reflected / delayed version of the sound combining with a direct version, you likely do not have a minimum phase situation.  In this case, conventional eq may make the amplitude look better but will be imposing phase shift that does not fix the original problem and makes that part of it worse.
If the "olden days" the rule of thumb to avoid trying to fix the unfix-able was "only cut the bumps, never try to fill in deep notches"  the reason was deep notches such as one sees with comb filtering are usually a result of two signals adding that had different time origins and so are not minimum phase.

Nowadays too, some fir filters do not have the normal minimum phase behavior and some allow the phase and amplitude to be addressed separately. 
For fun, Google up a program called "re-phase" which allows these to be independent and will generate a correction impulse which you can convolve to fix both.  Some pro speaker controllers allow this correction to be loaded as well as the mini dsp drc box, a cheap way to play with this.
Best,
Tom Danley
 

[ProSoundWeb Community]