XSim free crossover designer

Everything in Xsim is limited to 20kHz, no calculations or effects included beyond that (other than tails when doing Hilbert phase derivations).

Hi Bill,

Thank you.

I wish there is a feature in which one can input a real world sound clip waveform and simulate with the given xovers and drivers to see how the output would compare to the input. To me, although the step and impulse responses are very helpful, they are a bit theoretical and it's hard to translate them into real world affects of the real world sound.
 
Hi Bill,

Thank you.

I wish there is a feature in which one can input a real world sound clip waveform and simulate with the given xovers and drivers to see how the output would compare to the input. To me, although the step and impulse responses are very helpful, they are a bit theoretical and it's hard to translate them into real world affects of the real world sound.

Not 100% shure but will mean that should be doable for the design axis curve used in Xsim :)

1. In Xsim export each seperate bandpass as frd file.
2. Import each bandpass frd file per a dedicated transducer driver over in free VituixCad and let them sum with direct parallel wires without any XO parts. Should probably work without using active power amp in XO tab, else add one of those per transducer and if you need to delay any bandpass it can be set per power amp.
3. SPL chart should now look as Xsim did, then push "View" menu and then "Impulse response", fill in settings so rate is same as your audio test tracks will be and export this new IR convolution file.
4. Use a player as Jriver or Foobar to load the IR convolution file and tick it on and off to hear the difference while test track is playing, probably its best procedure to use head phones for listening test because they normal XO-less and then stay clear off add more XO point excess phase on top of what XO summing excess phase is within the convolution file.
 
Not 100% shure but will mean that should be doable for the design axis curve used in Xsim :)

1. In Xsim export each seperate bandpass as frd file.
2. Import each bandpass frd file per a dedicated transducer driver over in free VituixCad and let them sum with direct parallel wires without any XO parts. Should probably work without using active power amp in XO tab, else add one of those per transducer and if you need to delay any bandpass it can be set per power amp.
3. SPL chart should now look as Xsim did, then push "View" menu and then "Impulse response", fill in settings so rate is same as your audio test tracks will be and export this new IR convolution file.
4. Use a player as Jriver or Foobar to load the IR convolution file and tick it on and off to hear the difference while test track is playing, probably its best procedure to use head phones for listening test because they normal XO-less and then stay clear off add more XO point excess phase on top of what XO summing excess phase is within the convolution file.

Thank you. Sounds very intriguing. I'll sure to give it a try.
 
If I give someone all the specs to the drivers and box size and ports. Could someone make up a crossover design for me?
I have all the specs for the Dayton 8” woofers and seas tweeter and box size and ports being used. I have a 2way Electronic crossover for it that the person built for these but I’d rather use passive crossover
 
Last edited:
I am building Marten King's open baffel a 2 way using the FF85WK a 8 ohm driver with a RE of 7.2 . with a high pass xover using a 25uf cap and 4mh inductor. I need to know if this would work with an amp with a 4 ohm output impedience. I downloaded the XSim program but can't find the files needed to plug into it. Do one of you experts have a down and dirty way to find out how much that 4 ohm output resistance would affect the xover point and slope ?
 
Carl and woody, you can get great help here to use this magnificent software, however first there are a few things you need to do.



As I understand speaker crossover design you need to get your drivers in your enclosures and measure the individual driver frequency response and impedance. You also need to measure the combined frequency response of your drivers.


Once you have your zma and frd files the crossover magic can begin!
 
Woody,

go over to multy-way forum and someone might help you. There is free excel spreadsheets (response modeler can simulate OB) and a tracing software to work with manufacturer's graphs. I am not sure what you meant by having amplifier with 4 ohm output impedance. They are usually specified for certain loads.
 
Member
Joined 2008
Paid Member
I must be doing something wrong. I am trying to measure the delay between a tweeter and a woofer. I take three measurements (low, high, both), gate them to get rid of the first reflection and try to fit the curves by manipulating the delay. I can get the curves to match really closely. But I do not understand the results.

1. I must set negative delay on the tweeter horn. Does that mean that I would need to have a deeper horn if I wanted to align the driver with the woofer?

2. The strange thing is it seems I am getting a delay proportional to the measurement distance. I took only two sets of measurements and the closer measurement (ca 1 m)would call for 1.4" difference, at almost double distance I got almost double delay/offset in the same direction.

Could anyone explain why 2. is happening and what is really the physical offset between the drivers? The measurements are gated in REW, all by the same window. Could that be the reason? Ungated measurements do not align that nicely, I guess that's normal due to room interaction. Thanks in advance.
 
Can you show your plots?

I think you need to measure impulse response graphs of both drivers from same mic position and sample time locked between measurements.

Holm Impulse software lets you do this easily. I do not know how REW does this.

When I used this method together with Xsim, reality and sim matched nicely.
 
While optimising curves with actual loudspeaker curves as input (2..3m distance) I am aiming for a sloping down room curve (and perhaps it's in XSIM but I am not aware of it so I ask)...so it would be nice if in XSIM you could load a "Speaker" that is actually a sloped down room-curve (The B&K?). After plotting this in the graph, one could push the "HOLD" button and use that curve as a guideline to map the actual total curve on. Does this exist in one way or the other (or would it be a good feature to add)?
 
While optimising curves with actual loudspeaker curves as input (2..3m distance) I am aiming for a sloping down room curve (and perhaps it's in XSIM but I am not aware of it so I ask)...so it would be nice if in XSIM you could load a "Speaker" that is actually a sloped down room-curve (The B&K?). After plotting this in the graph, one could push the "HOLD" button and use that curve as a guideline to map the actual total curve on. Does this exist in one way or the other (or would it be a good feature to add)?

Hanged on below in three zip folders is frd target files for a 7dB curve (+3,5dB at 30Hz and -3,5dB at 30kHz) and the JBL F. Toole curve for trained listener and final curve is the B&K. One can load them in Xsim "Curves" meny but in they centered on vertical axis around 0dB point its smarter as you say load them on a temporary aux driver where on can offset them to whatever sensitivity number using Xsim dialog "mod sensitivity" and push "Hold" before dismount of aux driver.

attachment.php
 

Attachments

  • Arno Pf_1.png
    Arno Pf_1.png
    12.6 KB · Views: 373
  • 7db_curve.zip
    5.9 KB · Views: 48
  • B_and_K.zip
    180.6 KB · Views: 51
  • JBL_Toole_curve_trained_listener.zip
    10.2 KB · Views: 55
Member
Joined 2008
Paid Member
Here are the source REW measurements - the same box, two distances: LeteckaPosta.cz Sd?lejte Pos?lejte soubory leteckou po?tou - Posl?n? souboru - Zde nahrajte sv?j soubor

And exported MP FRDs are in the attachment. I tried the minimal phase version and the results were still strange. I would expect some differences, but these are dramatic.

The horn is 3D printed prototype and I would like to print a new one with the correct depth. On another boxes I tried REW and its acoustic reference and that worked quite well. I wanted to try the acoustic way.

I definitely need to try the acoustic reference with REW on this box at different distances.
 

Attachments

  • FBT.zip
    29.5 KB · Views: 54
Hanged on below in three zip folders is frd target files for a 7dB curve (+3,5dB at 30Hz and -3,5dB at 30kHz) and the JBL F. Toole curve for trained listener and final curve is the B&K. One can load them in Xsim "Curves" meny but in they centered on vertical axis around 0dB point its smarter as you say load them on a temporary aux driver where on can offset them to whatever sensitivity number using Xsim dialog "mod sensitivity" and push "Hold" before dismount of aux driver.

799913d1575571653-xsim-free-crossover-designer-arno-pf_1-png

Used them....and it's great...! Makes life a lot easier.
Is there also a feature to store/hold several curves in 1 graph?
 
...Is there also a feature to store/hold several curves in 1 graph?

In a way there is in one can in Frequency Response window load one more FRD file via "Curves" meny then "Get File" but one can't offset its vertical axis scale up and down on the fly so you have to prepare that curve to the right SPL area, for example first load curve as a AUX driver and adjust vertical axis to area of interest, now in Frequency Response window pick "File" then "Export system FRD file", and finally load that exported FRD file via menu "Curves"/"Get File".

Probably if you have plenty of screen area its more easy do as in below visual where the three target curves was "Hold" in each their Frequency Response window, see red arrow : )

attachment.php
 

Attachments

  • Arno Pf_3.png
    Arno Pf_3.png
    127.1 KB · Views: 343
Last edited: