[Tyson]

Hi all, I'm running my GR Research V2's in active mode, using a miniDSP 2x8 as my crossover unit, and a tube amp for the tweeters, an identical tube amp for the mids, and a crown K2 for the bass.  I've been doing some more measuring and tweaking and I'm pretty darn happy with the Frequency Response I've been able to dial in. 

My question is on the phase, shown in the dotted line below.  My crossover points are at 220hz and at 2100 hz, using LR24 for all points.  And I'm using HOLM Impulse with an Earthworks M23 calibrated mic for all measurements. 

Looking at the phase, I know that the steep angles are caused by my use of steep filters.  My question is this - are there any tricks or tips I can use to try to improve what's there right now?  Or is this as good as it gets and I just have to live with it?

[sts9fan]

My question is, does phase matter if you have good frequency response?  I am thinking no but have been pondering this lately.

[S Clark]

"My question is, does phase matter if you have good frequency response?"
Yes, especially when you consider off axis response.  You can tweak a crossover to get a very good response, but phase issues can cause a 6 dB drop or even more as you move as little as a few inches, literally changing as you sit up straight, or sit a foot or so from your sweet spot.  That would be an extreme case, but I've seen it when working on a line array project.  (and I hope to never do another one  )

Tyson, what kind of graphs do you get when reversing polarity? Do you get a good deep cancellation? If so, it is an indication of decent phase alignment.  Hopefully some of the pros will come along and give you a more detailed answer. 

[*Scotty*]

Tyson, you get 360 degrees of phase rotation with the 24dB/oct filter, this puts each driver back in phase at the crossover frequency which is why all your drivers are wired in phase with one another.
 The phase shift comes with the territory and is unavoidable. The other potential problem is phase errors due a misalignment of the acoustical centers of the drivers. This has a greater chance of causing frequency response perturbations at the midrange to tweeter crossover point due to the shortness of wavelengths involved at these frequencies. I can't tell how smoothing has been applied to your frequency response measurement from the supplied graph or how much of a problem you might have. Measurements at 30 and 60 degrees off-axis might reveal potential diffraction related problems.
 Your impulse response might be cleaned up a little bit by applying some felt around the tweeter if the hash after the initial pulse is due diffraction problems. McMaster-Carr on the web is a good place to source this material from.
Scotty

[Tyson]

It's a coax, and the tweeter is located at the back throat of the midrange unit, so their acoustical centers are aligned.  But it also means that I can't put foam around the tweeter

If I reverse phase on the tweeter or on the bass woofers, I get a deep null at the crossover points.

Also, the measurement I posted above was taken at 30 degrees off axis (which is how I listen to them).  Measuring at 60 degrees or even directly on-axis doesn't affect the FR very much.

[Freo-1]

My question is, does phase matter if you have good frequency response?  I am thinking no but have been pondering this lately.

Don't they go hand in hand?  If the phase is out, then frequency response will be out as well, would it not?
 
Have you asked the boffins at the DIY Audio site?   They have helped me out on amp issues in the past.

[*Scotty*]

Tyson it looks like you get what you get. Here is a link to the review of the Dunlavey SC VI speaker.
http://www.stereophile.com/content/dunlavy-audio-labs-signature-sc-vi-loudspeaker-measurements
The measurements show the phase coherence of the design. In this case phase and amplitude are equivalent. A deviation in amplitude will show up as problem with phase accuracy in Dunlavey's design.
Have you done a step response measurement of your loudspeaker? It would be interesting to see what it looks like based on your impulse response results.
Scotty

[Davey]

Don't they go hand in hand?  If the phase is out, then frequency response will be out as well, would it not?

Absolutely not.

The hifi industry is full of speaker systems that exhibit flat (or nearly flat) amplitude response on a measuring axis, but have multiple "wraps" of the phase response.  Any time you implement a crossover for a multi-way speaker system the resultant phase response will not be linear (flat.)  Only in a few special cases (DSP, first-order analog acoustic) can/could that be achieved, and then only at a single measurement point in space.

The phase response indicated by Tyson's measurement is a result of many things.  I don't know if this phase response is indicative of the stock V2 phase response Mr. Richie has designed.  It would be if the amplitude slopes Tyson generated with the miniDSP matched those of the stock system.  In any case, the V2 is not a linear-phase system so the non-flat phase response is inherent.

Cheers,

Dave.

[Freo-1]

Let's try asking a different way: 
When integrating a subwoofer, adjusting the phase clearly adjusts the output relative to the overall signal received at and around that frequency.   That effectively adjusts the bass output heard with regard to overall frequency response. 

[Davey]

Is that a question or a statement?  (I think the OP's (Tyson) original question has already been answered.)

The point here is that changes in phase response CAN affect amplitude response....but not always.  An all-pass filter is a good example a person might see utilized in speaker design work.

http://en.wikipedia.org/wiki/All-pass_filter

Cheers,

Dave.

[Freo-1]

Is that a question or a statement?  (I think the OP's (Tyson) original question has already been answered.)

The point here is that changes in phase response CAN affect amplitude response....but not always.  An all-pass filter is a good example a person might see utilized in speaker design work.

http://en.wikipedia.org/wiki/All-pass_filter

Cheers,

Dave.

Thanks.
 
That is what I was after.