Announcement

Collapse
No announcement yet.

Effect of phase through XO region?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Effect of phase through XO region?


  • #2
    Re: Effect of phase through XO region?

    Just posting the screen shots of both.
    Attached Files

    Comment


    • #3
      Re: Effect of phase through XO region?

      it does track a bit better on the first.
      " To me, the soundstage presentation is more about phase and distortion and less about size. However, when you talk about bass extension, there's no replacement for displacement". Tyger23. 4.2015

      Quote Originally Posted by hongrn. Oct 2014
      Do you realize that being an American is like winning the biggest jackpot ever??

      http://www.midwestaudioclub.com/spot...owell-simpson/
      http://s413.photobucket.com/albums/pp216/arlis/

      Comment


      • #4
        Re: Effect of phase through XO region?

        The farther away from the crossover point, the less phase tracking matters IMHO. When the tweeter is -40 or 50 db from the woofer, are you going to even hear it, much less detect phase problems?

        I would worry more about the 5K region than the phase.
        Don't listen to me - I have not sold any $150,000 speakers.

        Comment


        • #5
          Re: Effect of phase through XO region?

          I'm no crossover guru, but here's my "practical" take on it.

          The actual phase magnitude doesn't matter (ie. if it's 180 or 90 or 35). It's all about the phase for the two drivers being the same magnitude through the crossover region. A simple thought experiment for this would be: if the drivers are slightly out of phase, what would happen? Well, one leads the other for a given sound - but both drivers are audible. This means that if a 1ms attack (of a snare or something) occurs and the drivers are slightly out of phase, the effective sound you hear might be 1.5 ms long. Knowing this, I would say whichever crossover has the closer tracking throughout the region is better, even if it doesn't wrap at 180 right through the crossover.

          One way to quantify the better tracking (in addition to reverse null magnitude) would be to use one of the phases as a reference line (let's say the tweeter) and do a sum of squares analysis through the region. So if your crossover is 1500 Hz, and the region you are curious about is 1000-2500 Hz, you would subtract the woofers phase from the tweeters phase, square the result, then add up all the squares. The crossover that gives you the lowest Sum of Squares would have the statistically better phase tracking. This method would be favored to better phase tracking above the crossover point, due to the logarithmic scale. You could employ some sort of weighting function to account for this though. If you are interested, I could work out the math later today - and maybe even put together a spreadsheet that imports your frds/zmas and does the calculation for you.

          Now, someone should chime in and tell me if this is the right way to think about it - but from a physics standpoint I'm pretty sure this is what is happening. Also - I just came up with that sum of squares thing right now - has anyone done something like that before? It seems like it would be a nice way to quantify phase tracking, if also used with reverse null magnitude.

          EDIT: Just read johnny's post - you could weight the sum of squares technique by relative magnitude of the drivers, which would automatically account for what he said. IE phase at 1000 Hz is less important than phase at 1500 Hz in the example above.

          Comment


          • #6
            Re: Effect of phase through XO region?

            Originally posted by tculverhouse View Post
            I'm no crossover guru, but here's my "practical" take on it.

            The actual phase magnitude doesn't matter (ie. if it's 180 or 90 or 35). It's all about the phase for the two drivers being the same magnitude through the crossover region. A simple thought experiment for this would be: if the drivers are slightly out of phase, what would happen? Well, one leads the other for a given sound - but both drivers are audible. This means that if a 1ms attack (of a snare or something) occurs and the drivers are slightly out of phase, the effective sound you hear might be 1.5 ms long. Knowing this, I would say whichever crossover has the closer tracking throughout the region is better, even if it doesn't wrap at 180 right through the crossover.

            One way to quantify the better tracking (in addition to reverse null magnitude) would be to use one of the phases as a reference line (let's say the tweeter) and do a sum of squares analysis through the region. So if your crossover is 1500 Hz, and the region you are curious about is 1000-2500 Hz, you would subtract the woofers phase from the tweeters phase, square the result, then add up all the squares. The crossover that gives you the lowest Sum of Squares would have the statistically better phase tracking. This method would be favored to better phase tracking above the crossover point, due to the logarithmic scale. You could employ some sort of weighting function to account for this though. If you are interested, I could work out the math later today - and maybe even put together a spreadsheet that imports your frds/zmas and does the calculation for you.

            Now, someone should chime in and tell me if this is the right way to think about it - but from a physics standpoint I'm pretty sure this is what is happening. Also - I just came up with that sum of squares thing right now - has anyone done something like that before? It seems like it would be a nice way to quantify phase tracking, if also used with reverse null magnitude.

            EDIT: Just read johnny's post - you could weight the sum of squares technique by relative magnitude of the drivers, which would automatically account for what he said. IE phase at 1000 Hz is less important than phase at 1500 Hz in the example above.
            Finding the MSV of phase miss-alignment would be an easy way to put it to a single quantity, but isn't it easy to judge just by inspection?

            The notion I got from the thread I just had about phase alignment and power response left me thinking that phase alignment need not be perfect as long as they are summing as desired, both on and off-axis, thus acheiving whatever targeted acoustic roll-off.

            As for the drivers being slightly out of phase corresponding to different delays, that's not quite true. AFIK - The derivative (rate of change) of phase is group delay, meaning the two drivers would have different delays for a given frequency range if the slope of their Xover region phases didn't match.

            Comment


            • #7
              Re: Effect of phase through XO region?

              If my driver phases are tracking well together through the crossover area and there is a deep, well defined reverse null then I don't worry at all about where the phases wrap (180) with respect to the crossover point. Here is an example:

              Craig

              The lowest possible F3 box alignment is not always the best alignment.

              Designing and building speaker projects are like playing with adult Lego Blocks for me.

              Comment


              • #8
                Re: Effect of phase through XO region?

                What is phase wrap? Is that represented by the vertical line in phase, for example in PCD? The point at which that vertical line appears is meaningless, no?

                Or is it something else?

                Comment


                • #9
                  Re: Effect of phase through XO region?

                  Originally posted by PWR RYD View Post
                  If my driver phases are tracking well together through the crossover area and there is a deep, well defined reverse null then I don't worry at all about where the phases wrap (180) with respect to the crossover point. Here is an example:

                  That's a sh!t-hot job there

                  रेतुर्न तो थे स्रोत
                  return to the source
                  leviathan system thread
                  deadhorse thread
                  shockwave build thread

                  instagram :: greywarden_13

                  in war, victory . . . in peace, vigilance . . . in death, sacrifice.

                  Comment


                  • #10
                    Re: Effect of phase through XO region?

                    Thanks for all the info. Might take me a bit to digest it all, but it sounds like the better phase tracking is prefered over the better reverse null.

                    Originally posted by johnnyrichards View Post
                    I would worry more about the 5K region than the phase.
                    Yeah not sure what that's all about. I think that is just bad baffle placement on my part. It dissappears and smooths out a bit off axis... Says my measurements anyway.

                    Comment


                    • #11
                      Re: Effect of phase through XO region?

                      Phase wrap, or the location of the vertical line in phase seen in PCD, is 100% meaningless. Phase is "wrapped" for plotting purposes to keep the scale at -180 to +180 degrees. IRL phase doesn't "wrap"

                      Comment


                      • #12
                        reverse null, it is a null, right?

                        Comment


                        • #13
                          Re: Effect of phase through XO region?

                          Originally posted by jclin4 View Post
                          reverse null, it is a null, right?
                          Right - the remedy is to hook one of the drivers up in reverse polarity (ie. change it's phase by 180 degrees)

                          Comment


                          • #14
                            Re: Effect of phase through XO region?

                            If I mess around with the values in LspCAD I find a wider null gives better phase tracking than a deeper null. For example a wide 25dB null could have good phase tracking, so I adjust the values to get a 60dB null and although it's deeper, it's also not a wide as the 25dB null. When I check the phase tracking it's slightly out.

                            If I then try the same crossover values with some older measurements of the same speaker, the results are slightly different again. I don't know if it's because the drivers needed more running-in or if it's something i'm doing wrong :(

                            I've only designed one speaker, so don't have much experience and my technical knowledge isn't that great either.

                            Comment


                            • #15
                              Re: Effect of phase through XO region?

                              I think another thing to consider is just a few degrees off-axis can have an impact on reverse null and reverse tracking. If your phase and reverse null changes between measurement sessions I would tend to suspect that your mic was off a degree or two, or positioned a few inches one way or the other.

                              Here is a side project I have been playing with for a few weeks.

                              On axis, 1M:


                              Off axis, 15 degrees horizontal (simulation mode):


                              Off axis, 3 degrees vertical (simulation mode):


                              I mentioned a while back that phase is nothing more than a point in space. IMHO, I think we worry too much about what is going on in the design axis WRT phase, when we should be at least as concerned with what is happening on a few other axes.
                              Don't listen to me - I have not sold any $150,000 speakers.

                              Comment

                              Working...
                              X