Home Theater Forum and Systems banner

41 - 60 of 64 Posts

·
Registered
Joined
·
91 Posts
Discussion Starter #41
Wow! All is well with these settings, but I am going to guess you forgot to mention also invert the tweeter (in addition to the mid-range)? I had to invert the tweeter also to get the pretty plot below.

In addition, your xo points were off (the one's i was using were 80, 500, 3.8kHz)? I'm going to assume you intent was to leave crossover points the same? If I bump the woofer to 700Hz, distortion goes to 1% at 90dB?

TW = 0 ms (Ref Also inverted)
MR = -0.060 ms (Also reverse the polarity of this driver)
W = +0.800 ms
SW's = +0.000 ms

Left speaker (work of art):
mid and tweeter inv.jpg

Versus Left speaker original (not bad for first try :) ):
Original.jpg


Okay...
I have finished the timing analysis. The following are the changes to the current delay settings in the Xilica. I chose the TW as the reference timing so its delay does not change.
TW = 0 ms (Ref)
MR = -0.060 ms (Also reverse the polarity of this driver)
W = +0.800 ms
SW's = +0.000 ms

The net SPL impact (a minor improvement around 3500 Hz) can be seen below:
View attachment 135890

The XO range SPL reinforcement is shown below:
View attachment 135898

The net impact on overall phase rotation is shown below:
View attachment 135906

The impact on the step response is shown below:
View attachment 135914

View attachment 135922

The overall SPL is not significantly impacted. The phase rotation now follows a more conventional track and thus the step response also looks more conventional for a well timed 4 way system. The 250 Hz room effect is not addressed in these changes.
 

·
Premium Member
Joined
·
1,562 Posts
Yes, XO points the same. I was just referring to the approximate acoustic XO freqs as shown on the driver and overall SPL chart. The effective acoustic XO points often differ from the electrical settings. I was just too lazy to look up the exact electrical settings again.

Reverse the polarity of the TW from the way it was when your measurements were made?? That should not have been necessary based on the data file you sent. The TW driver IR should look like the chart shown below. Are you sure you did not change the TW polarity after you posted that data file? Maybe I accidental inverted it during my initial setup manipulation of the IR locations. I did not go back that far when I double checked the timing changes. I will start from the original data and run through the entire process again tomorrow to triple check everything again.

42 TW IR.jpg
 

·
Registered
Joined
·
91 Posts
Discussion Starter #43
Strange. I also left original timings and offset them as you stated.

For example, I added .8ms to the woofer, subtracted from the existing MR setting, etc... If I don't invert the tweeter I get a nasty dip at 5khz. I think it sounds really good, no time to play with dirac tonight though. Really blends well and sounds like imaging is better.

Yes, XO points the same. I was just referring to the approximate acoustic XO freqs as shown on the driver and overall SPL chart. The effective acoustic XO points often differ from the electrical settings. I was just too lazy to look up the exact electrical settings again
Reverse the polarity of the TW from the way it was when your measurements were made?? That should not have been necessary based on the data file you sent. The TW driver IR should look like the chart shown below. Are you sure you did not change the TW polarity after you posted that data file? Maybe I accidental inverted it during my initial setup manipulation of the IR locations. I did not go back that far when I double checked the timing changes. I will start from the original data and run through the entire process again tomorrow to triple check everything again.

View attachment 135978
 

·
Premium Member
Joined
·
1,562 Posts
You followed the instructions correctly. I reviewed my work again and found no issue, but there is still a difference between our results.

Your measurement, the first chart in Post-41, does appear to be the correct relative polarity between the drivers. It is impossible to be sure how accurate the timing is just by casual inspection however.

I did note that polarity of the TW in that chart is definitely Negative. It flattens out at -180°. That is reversed from convention and my intent. It does properly merge with the MR so that suggest it is reversed from my intent as well. Everything else also appears to also be reversed, so far as I can tell with a casual look. It's possible that reversing the polarity of all the drivers (SW's, W, MR, TW) will result in my intended target response.

[I also noted that the 2nd chart in Post-41 does not agree with the polarity relationship between the MR and TW of the file in Post-33 that I based my analysis on.]

The best action now is to measure the left channel drivers individual again as done in Post-33 using your current delay settings (as per chart 1 in Post-41). This time also include the total left channel measurement (SW's+W+MR+TW). Be sure REW acoustic timing is activated. I can then confirm it matches the target (given that all polarities are reversed).
 

·
Registered
Joined
·
91 Posts
Discussion Starter #45
I feel like I should know the answer to this, but could the factory speaker wiring be wired such that the woofer is opposite polarity of the MR and TW from the factory?

I did not take the drivers out, and just accessed the crossovers directly by removing the bottom of the speaker, trusting the factory +/- labeling on the crossover circuit boards along with the + markings on the wires themselves.

I am reading up on a way to verify polarity with 100% assurance, but I cannot easily remove the drivers to examine the voice coil or cone movement. I will verify with absolute assurance how things are wired before remeasuring.

I feel like this is the problem and why I had the big 5kHz dip before with the Ashly XR1001. I definitely did not change polarity on both channels or anything so I am very confused here. I suppose the Xilica XD4080 could have an issue where sometimes when it boots polarity is not set correctly? I do turn off the Xilica everynight (wish it had a 12V trigger).

Also, does it matter that the 500Hz and 3.8kHz are FIR crossovers, while the 80Hz crossover is a butterworth? I would assume not since REW is adding things together and this should be math a computer could do very accurately.

I did swing the mic out from the listening position to sit down on 2/5/2017 after taking the last set of driver measurements for you, but I swung it back last night after making your changes. It is definitely within an inch or two of where it was. If there was some room issue interfering with things, this could be the issue?

I like your idea of remeasuring everything. Seems like a reasonable and pragmatic way to proceed as phase looks noticeably better to me (but I'm no expert).

I have been taking the all measurements (including the driver measurements) at the MLP with the mic pointed upward.

I also ordered some different amps to try (just for fun has nothing to do with what we are doing here). But since the gain structure will change a third time, I would like to wait until the new amps are installed. I will remove the XLR padding I added, but plan to match the levels from the previous file.

Your measurement, the first chart in Post-41, does appear to be the correct relative polarity between the drivers. It is impossible to be sure how accurate the timing is just by casual inspection however.
 

·
Premium Member
Joined
·
1,562 Posts
I feel like I should know the answer to this, but could the factory speaker wiring be wired such that the woofer is opposite polarity of the MR and TW from the factory?
Sure, the polarities are chosen to best blend the handoff between the drivers. Whether a driver is connected with a Positive or Negative polarity depends on the relative acoustic XO filter slopes and driver offset to the adjoining driver.

I did not take the drivers out, and just accessed the crossovers directly by removing the bottom of the speaker, trusting the factory +/- labeling on the crossover circuit boards along with the + markings on the wires themselves.
I would expect +/- labeling on the XO board and wires to be correct.

I am reading up on a way to verify polarity with 100% assurance, but I cannot easily remove the drivers to examine the voice coil or cone movement. I will verify with absolute assurance how things are wired before remeasuring.
The wired polarity really does not matter now. We chose the needed driver polarity as a result of the measurement analysis and reverses it when needed to get the best handoff. None of my recommendations were concerned with the absolute polarity, only if a driver needed to be reversed in order to achieve the best phase tracking. The exception to that is that is that the convention is to have the TW polarity in positive absolute polarity. That is shown by the phase trace leveling out relatively flat near 0° rather than near 180° near the top of its bandpass range. Reversing a driver polarity is dead simple now as the Xilica would allow that to be done by just checking a box; no concern as to how it was actually connected by wiring.

I feel like this is the problem and why I had the big 5kHz dip before with the Ashly XR1001. I definitely did not change polarity on both channels or anything so I am very confused here. I suppose the Xilica XD4080 could have an issue where sometimes when it boots polarity is not set correctly? I do turn off the Xilica everynight (wish it had a 12V trigger).
I'm confident the Xilica boots correctly each time. It would be an major issue if it didn't.

Also, does it matter that the 500Hz and 3.8kHz are FIR crossovers, while the 80Hz crossover is a butterworth? I would assume not since REW is adding things together and this should be math a computer could do very accurately.
It only makes a difference if there is a change. If the Post-33 measurements were taken with the final XO settings then we would be okay. Any change to the XO settings for IIR slope, freq, or a change to FIR linear phase type will impact the analysis and thus possibly driver polarity and/or delay timing.

I did swing the mic out from the listening position to sit down on 2/5/2017 after taking the last set of driver measurements for you, but I swung it back last night after making your changes. It is definitely within an inch or two of where it was. If there was some room issue interfering with things, this could be the issue?
Exact mic LP location is not a concern. The REW acoustic timing still provides an accurate relative phase relationship between the drivers. The room impact is trivial to this analysis.

I like your idea of remeasuring everything. Seems like a reasonable and pragmatic way to proceed as phase looks noticeably better to me (but I'm no expert).
Good, I think that will resolve the situation. Hopefully the only recommended change will be to reverse the polarity of all drivers.

I have been taking the all measurements (including the driver measurements) at the MLP with the mic pointed upward.
No problem. The phase is not impacted by mic orientation. It is the same at 180° and 0°

I also ordered some different amps to try (just for fun has nothing to do with what we are doing here). But since the gain structure will change a third time, I would like to wait until the new amps are installed. I will remove the XLR padding I added, but plan to match the levels from the previous file.
I will wait for the new data. Just be sure save all the settings and make no changes to the setup after you post the new data.
 

·
Registered
Joined
·
91 Posts
Discussion Starter #47 (Edited)
Yes, that is how I changed polarity last night was with the Xilica XConsole software. My worry was that you have used the left driver data, and I have measured right (or vice-versa), and one of drivers phase is reversed. I had included right and left driver measurements before though, I just looked at them and don't see any large phase deferences between left and right drivers.

The only thing that could be different in Xilica behavior between the two measurements is level being limited in one set of measurements and potentially impacting the results. I'm guessing the limit settings on the TW and MR could probably impact phase. I take the measurements outside the room, so if the limiter settings were too low, I would not know as I can't see the red lights on the front of the Xilica.

I am going to raise the limits just in case. It is confusing as 0dB limit in the software does not line up with the 0dB on the level setting (as far as I know since the red light doesn't come on on the front of the unit).

I had to change the gains a lot when I added XLR pads (which I did all before doing the last driver measurements), but I overlooked raising the limits when I did that. To make matters worse, I also bumped incoming gain too as the Oppo HA-1 wasn't getting loud enough after adding pads and level matching the drivers. Beginning to think this is the issue and I have wasted a lot of your time regrettably.

Gains right now (will change a lot with new low wattage amps) are below. I only installed XLR pads on the weekend, but only on MR and TW drivers since the spare amps I had were huge (600W+ into ohms), and I wanted to eliminate the hiss caused by the Xilica.

New amps (some used Benchmark AHB2 amps) have switchable gain for pro versus consumer gear, much quieter at lower wattage, less IMD, etc...

I think one level in REW was higher than the other, so the higher level must have hit the limiters target (or even the compressor on the incoming channel) and that's why we are seeing strange results?

The wired polarity really does not matter now. We chose the needed driver polarity as a result of the measurement analysis and reverses it when needed to get the best handoff. None of my recommendations were concerned with the absolute polarity, only if a driver needed to be reversed in order to achieve the best phase tracking. The exception to that is that is that the convention is to have the TW polarity in positive absolute polarity. That is shown by the phase trace leveling out relatively flat near 0° rather than near 180° near the top of its bandpass range. Reversing a driver polarity is dead simple now as the Xilica would allow that to be done by just checking a box; no concern as to how it was actually connected by wiring.
 

Attachments

·
Premium Member
Joined
·
1,562 Posts
As you noted the driver polarities in the right channel and left channel in the Post-33 data was identical. I initially confirmed that and also confirmed the mic was reasonably centered. [If the XO's settings in the Xilica was identical for left and right channel as it should be, then that confirms the drivers in both channels were wired identically.]

In that file the SW's timing reference was the right channel per your labeling. I therefore chose the left channel for the analysis as the those drivers timing reference was also the right channel, thus a consistent time reference. The right channel drivers timing reference was the left channel. That means there could be a slight error for SW to W timing since the timing reference channel was different.

Once one channel is timed correctly the other should use the identical timing. There is no need to repeat the analysis. If the mic was not exactly centered the only impact is that the effective LP is moved from the actual mic position to one that is exactly centered and either slightly forward or back from the actual mic distance. This is trivial impact from a practical perspective. Using only one channel also cut my data analysis time by 50%.

Driver gain changes do not impact the driver to driver phase relationship. It will impact the apparent acoustic XO frequency as well as the SPL. This is no concern for driver timing analysis. Timing results will be unchanged. After timing is established the driver levels can still be changed if needed to best prepare for the EQ effort.

I saw no indication that limiting was a factor in the SPL chart, but it is always best to be sure. You are measuring at about 85 dB. We normally recommend 75 dB as good level for this type work. That would provide another 10 dB of headroom to your limiter setting to ease your concern.

Low power and low sensitivity amps on the MR and TW drivers will help with hiss noise. The gain structure of rest of the system is also a major factor.
 

·
Registered
Joined
·
91 Posts
Discussion Starter #49 (Edited)
Thanks for these detailed answers to my questions. I wish there was a step by step textbook on this that explained all the ins and outs. The best I've found on the internet are your posts. I think another beer is in order! Your response was very clear and well thought out.

One last question that has been bothering me: how would one know if the timing was off by 360 degrees? If this were to occur, could you see phase alignment at the crossover, but having timing off?

I think the answer would be you look at each drivers unwrapped phase in REW and make sure they are in line with one another (this is possible since we are using an acoustic reference for timing)? If so, do you have to pick a common point to unwrap at (e.g. the crossover point between two drivers)?

As you noted the driver polarities in the right channel and left channel in the Post-33 data was identical. I initially confirmed that and also confirmed the mic was reasonably centered. [If the XO's settings in the Xilica was identical for left and right channel as it should be, then that confirms the drivers in both channels were wired identically.]

In that file the SW's timing reference was the right channel per your labeling. I therefore chose the left channel for the analysis as the those drivers timing reference was also the right channel, thus a consistent time reference. The right channel drivers timing reference was the left channel. That means there could be a slight error for SW to W timing since the timing reference channel was different.

Once one channel is timed correctly the other should use the identical timing. There is no need to repeat the analysis. If the mic was not exactly centered the only impact is that the effective LP is moved from the actual mic position to one that is exactly centered and either slightly forward or back from the actual mic distance. This is trivial impact from a practical perspective. Using only one channel also cut my data analysis time by 50%.

Driver gain changes do not impact the driver to driver phase relationship. It will impact the apparent acoustic XO frequency as well as the SPL. This is no concern for driver timing analysis. Timing results will be unchanged. After timing is established the driver levels can still be changed if needed to best prepare for the EQ effort.

I saw no indication that limiting was a factor in the SPL chart, but it is always best to be sure. You are measuring at about 85 dB. We normally recommend 75 dB as good level for this type work. That would provide another 10 dB of headroom to your limiter setting to ease your concern.

Low power and low sensitivity amps on the MR and TW drivers will help with hiss noise. The gain structure of rest of the system is also a major factor.
 

·
Premium Member
Joined
·
1,562 Posts
Thanks!!

I wish there was a step by step textbook on this that explained all the ins and outs.
A detailed example of the overall process is long and complicated and there are numerous paths. Even stepping through a detailed example of my preferred method is more than I want to at this time. Maybe I will do that sometime. The practical value of tuning to this level is probably more in the charts than the actual sound quality benefits. There are probably not too many DIY'ers that want to spend excessive time fine tuning the system rather than listening to it.

One last question that has been bothering me: how would one know if the timing was off by 360 degrees? If this were to occur, could you see phase alignment at the crossover, but having timing off?

I think the answer would be you look at each drivers unwrapped phase in REW and make sure they are in line with one another (this is possible since we are using an acoustic reference for timing)? If so, do you have to pick a common point to unwrap at (e.g. the crossover point between two drivers)?
It is clear your efforts have been pretty effective so you are getting your own method pinned down. I think the difference on one XO we discussed was probably a simple confusion as to the settings that were active at the time of the measurement. Possibly the wrong file was posted.

Below are some short thoughts to your questions in case it helps. If you need more clarification, let me know.

1. I avoid using unwrapped phase as I found it easy to get confused due to room modes and strong reflections. Your heavily treated room provides relatively clean phase trace so... Possibly others have good results using unwrapped phase. I just found it problematic for most all work. I haven't really tried it with FDW now available as I am comfortable using wrapped phase charts.

2. Regarding IIR XO filters:
> The initial rise of the 2 Impulses should initially be set to the same time (initial rise, not the peaks). [They also need to be near 0 ms for their phase to be easily read, but only their relative positions are important in to how they actually complement each other.] From there, the minimum offset that provides a crossing at the acoustic XO frequency and the closest phase tracking is the best solution. That solution assures the same arrival time of the XO frequency from both drivers. Moving relative timing further away in either direction from that solution we can find other crossing points that represent the arrival of the sound from the 2 drivers being 1, 2, 3,... cycles delayed from each other. If we invert one Impulse we find other solutions at 1/2, 1-1/2, 2-1/2,... cycles away. Again the closest phase tracking represents the closest timing across the XO range.
> If the 2 acoustic filter slopes ideally follow a low-pass/high-pass LR24 XO shape then phase will track directly on top of each other. The same is true for other complementary filter shapes. To the extent that the 2 acoustic XO filters deviates from being complimentary the best solution involves a phase crossing at the XO frequency and the one and shows the minimum diversion above and below the XO point.

3. Regarding FIR linear phase filters:
> If ideal FIR XO's were implemented then the resulting large positive impulse peaks would be aligned at the same time.
> You applied FIR filters to the upper 2 XOs. Note that Post-33 shows a very shallow slope on the HPF of the 770 Hz SO (acoustic XO). There is significant overlap the W to MR as a result. I don't know if this was a limitation of the taps available or ??. It didn't appear to me that this XO was done following the normal best practices for linear phase XO filters. As a result, I see this as a mixed phase XO and a somewhat less attractive option than just targeting an LR-48 IIR response. Regardless, once the XO filters are chosen and in place, there is still one best solution for delay timing and that is what was recommended above.
 

·
Registered
Joined
·
91 Posts
Discussion Starter #51
Yes, this is a limitation on the number of taps and how they are assigned. So it is beneficial to use LR-48 for LF HP and LF-MR XO both?

Or would it be worth while to do LF HP as LR-48dB and spread the available taps out between TW and MR channels (this is how the FIR slope is controlled, the Xilica automatically determines the slope after you assign taps to each channel pair). In this scenario, I'd assign most taps to MR so low pass FIR slope closely matched Linkwitz Riley 48dB/octave and use the remaining taps for the TW channels (e.g. left MR/right MR is a channel pair of channels 3&4, channels 5&6 and TWs, etc...). Otherwise, I am essentially wasting taps unless you think I should do a brick wall for MR & TW crossover?

3. Regarding FIR linear phase filters:
> If ideal FIR XO's were implemented then the resulting large positive impulse peaks would be aligned at the same time.
> You applied FIR filters to the upper 2 XOs. Note that Post-33 shows a very shallow slope on the HPF of the 770 Hz SO (acoustic XO). There is significant overlap the W to MR as a result. I don't know if this was a limitation of the taps available or ??. It didn't appear to me that this XO was done following the normal best practices for linear phase XO filters. As a result, I see this as a mixed phase XO and a somewhat less attractive option than just targeting an LR-48 IIR response. Regardless, once the XO filters are chosen and in place, there is still one best solution for delay timing and that is what was recommended above.
 

·
Premium Member
Joined
·
1,562 Posts
I mislead you with that comment. No XO changes are necessary. I only meant that the acoustic W/MR XO looked a little less than optimized in how smooth and symmetrical it is. It's not a classic LR24 or LR48. It's common to spend much more time optimizing the acoustic XO than setting the delay timing. The attached chart shows the acoustic XO's for the Post-33 data.

I have limited expertise and no experience setting up linear phase XO's. For guidance there, I urge you read the available how-to guides. I think MiniDSP.com has some guidance and the rePhase thread on diyAudio forum is comprehensive and it contains other authoritative links as well.

I was mostly intending to point out that the roll-off slopes of W/MR linear phase XO is not very conventional. I also suggested that there was extended overlap of the MR in the W-MR XO. On reviewing that thought again, I now see the initial drop of the MR is about 10 dB and very steep. It then levels out significantly before falling normally. I was hinting that some attention to the XO settings may be worthwhile and I speculated the possibly a simple LR24 minimum phase XO there may be an improvement from the current settings. It is possibly not as significant as I was originally thinking.

The points are:
> The XO design can be optimized as well.
> With proper XO design either minimum phase or linear phase XO's can provide similar results.
> Once you have designed and applied the XO scheme there is a best timing setting for the drivers.
> Given Post-33 as the XO settings then the delays suggested previously provide very good phase tracking through the XO ranges. There is no significant improvement available in that regard.

Leave the current XO settings, improve the FIR XO setting, or try IIR XO settings to try to achieve a more idealized acoustic XO target if you like.
Whatever choice you make, I can then help confirm the delay settings you choose.

Symmetry.jpg
 

·
Registered
Joined
·
91 Posts
Discussion Starter #53 (Edited)
Ended up saving taps by using 48dB/Octave butterworth for all filters except MF to HF at 3.8kHz.

Left driver measurements (taken tonight 2/15/2017): NEW LINK: https://drive.google.com/open?id=0B1J0a4OV_WGLczRRc0VYU091czQ

Also dated the file this time as I was confused. I ended up moving my room around many times, but never could avoid the 248Hz issue, so I moved subs and main speakers to maximize imaging, screen and seating. I also lowered my shelves I made so I can fit my center speaker and also see my entire screen.

Here are the time delays (note I left 16ms of padding in case it is needed later when offsetting the delay to match your settings).

Sorry it took so long. I took my room all apart to fit a floor model couch, but found I couldn't fit the couch into my door due to the hallway :(

I mislead you with that comment. No XO changes are necessary. I only meant that the acoustic W/MR XO looked a little less than optimized in how smooth and symmetrical it is. It's not a classic LR24 or LR48. It's common to spend much more time optimizing the acoustic XO than setting the delay timing. The attached chart shows the acoustic XO's for the Post-33 data.
 

Attachments

·
Premium Member
Joined
·
1,562 Posts
The best timing solution is as follows:
1. Reverse the current polarity of both the Subs and the LF Drivers
2. Delay changes from the current settings:
  • TW = 0.00 ms (ref)
  • MR = +0.07 ms
  • LF = +0.90 ms
  • Subs = +1.80 ms
Data
> Acoustic XO's:
  • MR-TW XO (~3400-4250 Hz):
  • LF-MR XO (~300-1200 Hz):
  • Sub-LF XO (~40-160 Hz):
> Final Charts:
53 SPL Support.jpg

53 Phase Rotation.jpg

53 MR-TW XO Phase Tracking.jpg

53 LF-MR XO Phase Tracking.jpg

53 Sub-LF XO Phase Tracking.jpg
 

·
Registered
Joined
·
91 Posts
Discussion Starter #55
How does this one look for MR and TW? I ended up inverting TW and MR (incidentally this polarity means TW and MR polarity will NOT be inverted in the Xilica settings). Then I set TW impulse peak to 0ms (and set IR window delay to 0 for TW verifying wrapped TW phase was mostly horizontal) and modified MR as shown (without changing IR window delay for MR).

After that I compared TW and MR impulses to the originals measured in my previous post from last night.

With proper XO design either minimum phase or linear phase XO's can provide similar results.
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter #56 (Edited)
Just saw this as I was posting my independent test from this morning. Looks like since in mine I had "Inverted TW and MR. Add MR delay of +213us. Decrease TW delay by -146us" that I might be getting this as the different here would be MR +67us or ~70us = .07ms! How about that; I think I finally got the same optimized result on you.

I suspect if I were to continue with MR and LF I would see the need to re-invert MR and TW? I have no idea how you did this so fast!?! I spend two hours this morning just getting the relative offset for TW and MR.

EDIT: I can see how inverting MR and TW like I did doesn't give as constant of a phase slope one octave below 3.8kHz. I still like yours better. It is a work of art!

The best timing solution is as follows:
1. Reverse the current polarity of both the Subs and the LF Drivers
2. Delay changes from the current settings:
  • TW = 0.00 ms (ref)
  • MR = +0.07 ms
  • LF = +0.90 ms
  • Subs = +1.80 ms
 

Attachments

·
Premium Member
Joined
·
1,562 Posts
We both found the same relative delay and polarity needed for MR-TW XO. Well done! Since all driver polarities are reversed between us, we both have found the same relative driver to driver solution (this assume the delays for the other 2 XO's are also the same). Only the overall polarity of the total system is reversed and there is of no practical consequence to sound quality as to which overall polarity settings you decide use.
 

·
Registered
Joined
·
91 Posts
Discussion Starter #58
Looks great. Only a small dip in phase around 3.8kHz that wasn't predicted. I think the small dip might be due to the Xilica rounding to the nearest centimeter when I enter the values.

I did end up swapping TW and MR to normal polarity (had them reversed in the Xilica from before), which if I understand correctly had the same impact as swapping Sub and LF polarity.

If this looks good, I'm going to subtract 18ms from each delay and run dirac.

Thanks again!


We both found the same relative delay and polarity needed for MR-TW XO. Well done! Since all driver polarities are reversed between us, we both have found the same relative driver to driver solution (this assume the delays for the other 2 XO's are also the same). Only the overall polarity of the total system is reversed and there is of no practical consequence to sound quality as to which overall polarity settings you decide use.
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter #59
Here it is with a small additional delay of .032ms added to the TW (so as to not impact the MR-LF crossover.
 

Attachments

41 - 60 of 64 Posts
Top