[StanD]

It's because phase shift is likely regarded to be regarded as +/- 90 degrees as that is what can be achieved with caps/inductors. Correct me if I'm wrong

 

[KeithEmo]

Unfortunately you're oversimplifying.
 
Phase shift can be anything - you could have a phase shift of 0.001 degrees, or 27.235 degrees, 90 degrees, or 926 degrees.
 
It just so happens that 90 degrees is a particular phase shift that is useful for certain things

 
The other thing is that, in the real world, phase shift almost always varies with frequency. For example, it is pretty much impossible to design a network out of R's, L's, and C's that will give you the same phase shift at different frequencies. In a simple R-C or L-C network, the phase shift will be different at every frequency, and will vary according to a simple formula. (When you hear of networks that deliver "90 degrees of shift over a wide range of frequencies", what they've really done is to design a network that produces something near 90 degrees of phase shift over a certain range of frequencies. A typical "90 degree all-pass network" in an old-time SQ decoder, for example, was really more like - 90 degrees, +/- 15 degrees, between 100 Hz and 10 kHz, and that's being quite optimistic.)
 
With things like speakers, which have lots of reactive parts, it gets quite a bit messier. Look at some phase plots of actual speakers..... A typical three way speaker may have a phase plot that zig-zags over a range of several HUNDRED degrees of phase, and usually not in a nice straight line or a smooth curve. This is why things like crossovers and room correction get so complicated. You can adjust your sub so that is perfectly in phase with your main speaker's woofer at 64 Hz, but it will still almost certainly not be perfectly in phase at any OTHER frequency. This is why specific situations tend to require lots of fiddling around with crossover frequencies and phase adjustments to find the compromise that provides the best overall results in a given room.
 
 
 
 
Quote:

It's because phase shift is likely regarded to be regarded as +/- 90 degrees as that is what can be achieved with caps/inductors. Correct me if I'm wrong

 

[KeithEmo]

It doesn't - specifically. However, virtually all types of filters (digital or analog) that produce changes in frequency response ALSO produce changes in phase response.
 
You basically have three choices:
 
1) Ignore the phase shift your filters introduce
2) Minimize the phase shift and try for a filter that corrects the frequency response without buggering up anything else
3) Create a specific filter whose frequency response fixes your frequency response issues, but design it such that it's phase response ALSO corrects for phase/time issues.
 
That last one is what Dirac does. Rather than use one specific type of filter, which would have one specific phase response, they use what's called "mixed phase" filters, which is a fancy way of saying that they can choose from a variety of filter types, each of which has different effects on phase. This allows them to analyze each situation, and then design a custom filter that corrects both the frequency response and the timing errors - at least to a significant degree. (Instead of "being stuck" with the phase response of the specific filter type they use, they vary the type of filter they design for each situation so they can USE the differences between them as a tool for correcting phase issues.)
 
If you use a PEQ to correct frequency response, it's sort of pot luck whether the phase response of the filters it uses will make the phase response of the overall system better, worse, or no different. By actively CHOOSING between different types of filter configurations, Dirac can slant that trade-off firmly in favor of correcting the frequency response issues and IMPROVING the phase response. This generally results in a huge overall improvement (although the benefits vary to some degree depending on the situation).
 
Quote:

  What - that does not sound right to me phase and time delay are from speakers not being equidistant and the waves not arriving uniformly and/or speakers being wired backwards inverting the phase . Exactly how does freq response analysis have anything to do with that?

 

[Phishin Phool]

Thanks Keith it is nice to learn something factual and not repeated mis-information and the application of a filter affecting phase certainly makes sense . I have a question - Does having a set of timbre matched speakers help to keep them in phase over varying frequencies or is there still significant variation?

 

[KeithEmo]

That's sort of a tricky question.
 
The fact that the speakers are timbre-matched probably wouldn't signify much one way or the other.
 
However, speakers all of the same model and manufacturer should be quite closely matched. The phase response of a speaker is determined by the drivers, the design and type of crossover used, and even the cabinet tuning has some effect on it. I would figure that speakers of the same model from the same batch would be VERY closely matched, you might or might not see very slight differences over different batches (especially if the particular model is produced for a very long time). Next, speakers of similar models, with similar drivers, and from the same series (like the center and fronts form a certain model line) would also probably be pretty close.
 
Anything beyond that is a bit of a guess since even one manufacturer may decide to use very different crossover designs on different models, and they can be wildly different between manufacturers, even on speakers that are superficially similar in design, and that sound much the same (like three way speakers with dome tweeters and 8" cone woofers).
 
Luckily, it's probably not critically important that different sets of speakers be that closely matched as long as the speakers in each symmetrical pair (front mains, or surrounds) are very close. The biggest effect phase response has is on imaging, and there just isn't much critical imaging between, say, a left front and a left surround. (Things like frequency cancellations due to phase differences will probably be less noticeable than frequency response aberrations caused by room placement and such.)
 
 
Quote:

  Thanks Keith it is nice to learn something factual and not repeated mis-information and the application of a filter affecting phase certainly makes sense . I have a question - Does having a set of timbre matched speakers help to keep them in phase over varying frequencies or is there still significant variation?

 

[bigshot]

The room is going to have MUCH more of an effect on phase of speakers than the speakers.

 

[StanD]

  The room is going to have MUCH more of an effect on phase of speakers than the speakers.

Amen to that. Even the person in the room (not a hall) is an acoustic factor and is going to have an effect, especially when they move around.

 

[Steve Eddy]

Amen to that. Even the person in the room (not a hall) is an acoustic factor and is going to have an effect, especially when they move around.

Move around? If you move around you're obviously not an audiophile. You must sit, motionless with your head locked in the sweet spot. Anything else and you might as well just get a Bose.

se

 

[Phishin Phool]

Thanks - luckily my front soundstage speakers are all make/model matched with sequential serial #'s so hopefully not much variance. I know toe in or out just a bit affects imaging significantly but that is certainly room / dipole influenced. 

 

[Mr Rick]

  Thanks - luckily my front soundstage speakers are all make/model matched with sequential serial #'s so hopefully not much variance. I know toe in or out just a bit affects imaging significantly but that is certainly room / dipole influenced. 

 
Just make sure the relative humidity and barometric pressure are maintained constant, or all your hard work will be for naught.

 

[Phishin Phool]

Lol, actually my home theater room is humidity in climate control however it is not for sonic reasons

 

[analogsurviver]

  Thanks Keith it is nice to learn something factual and not repeated mis-information and the application of a filter affecting phase certainly makes sense . I have a question - Does having a set of timbre matched speakers help to keep them in phase over varying frequencies or is there still significant variation?

Set of timbre matched speakers DOES NOT necessary keep them in phase over varying frequencies - even in case of theorethically perfect match between the speakers.
 
Reason ? Phase response/coherence - or however you might want to call it. Here it is full range single drivers that shine in phase accuracy - full range electrostatics, full range dynamic drivers, from oval SABA, Lowther, etc. I specifically remember measurement for Acoustat full range output transformerless amp powered
ESL vs Bowers & Wilkins 3 way dynamic system - while both were reasonably linear in amplitude, phase response has shown the difference - while the Acoustat did not have exactly the same phase shift for all frequencies within 20-20 kHz, it DID have positive phase across this entire range - (with the possible anomaly at the fundamental system resonance in the bass ) ;
B & W allowed itself FIVE times to go from positive to negative - which sounds yuck with recordings that do take phase seriously. No wonder no one can hear the effect of absolute phase reversal using such designs - as they mangle this important piece of audio information beyond recognition. 
 
You can go and compare measurements from innerfidelity - they publish it as pulse response. While amplitude can be relatively easily made flat, achieving phase coherent output from a multi driver headphone is a major PITA. This is why I never could warm myself for any Shure multi way IEM - I prefer something with less extended response, but "in one piece". Dynamic drivers are getting ever better and although only the very pinnacle of them can ever hope to compare with electrostatics regarding phase fidelity, they are getting better and are trickling down to more reasonable prices. The new HiFiMan HE-1000 ( not heard yet, the first opportunity for me will be likely Hi End Munich in May ) may well be the first dynamic seriously challenging electrostatics . Hopefully, within couple of years the technology may well trickle down to within reach of most serious listeners - like not above 1K per pair.

 

[StanD]

Move around? If you move around you're obviously not an audiophile. You must sit, motionless with your head locked in the sweet spot. Anything else and you might as well just get a Bose.

se

Don't be so down on Bose, they are great at two things.

• ANC (Active Noise Cancellation)
• Marketing

Not necessarily in that order.

 

[SilentFrequency]

I think the main problem with audiophilia is that manufacturers keep coming up with solutions to problems that don't exist. This fancy memory card is a gold plated example. I don't think there is a situation where it would make any audible difference. Jitter... high bitrates... high sampling rates... super audible frequencies... distortion levels three times below the threshold of audibility instead of just two times... these are the sorts of things that people worry about and throw money at. But it's all a complete waste of time, energy and money, because none of this impacts the sound quality of recorded music.

If audiophiles actually focused on improving things they actually *can* hear, instead of the things they *can't*, they might actually be able to achieve better sound quality. But that requires more research and understanding than money. Easier to just trust the salesman and throw money at problems that don't exist. Laziness.

I totally get your point but I think in this case ie SONY which is a totally international brand then a lot of people would put their trust in this brand, rightly or wrongly so Im unsure if laziness is maybe fair to label with?

 

[Steve Eddy]

Don't be so down on Bose, they are great at two things.

1. ANC (Active Noise Cancellation)
2. Marketing

Not necessarily in that order.

Ha!

se