Simulation can read from FRD files and use only two planes at a time. Names of those planes are "hor" and "ver" in graphs, though they can be actually/geometrically anything as long as angle value is integer.
BUT simulated power & DI responses are not necessarily accurate if combination of selected two planes does not represent power & DI responses of the speaker.
Basic guideline is that you should simulate typical speakers with actual hor and ver planes to get valid power and DI result, but you can use Directivity chart to visualize any other/diagonal plane and fine tune design by that information.
Sorry that I forgot to mention that I can investigate features of dsp application if filter designer program runs on Windows without license and actual/physical dsp gear is not needed and filter designer program shows response accurately.
List of four items to test is for dsp devices which are impossible to configure and test without purchasing device.
As mentioned, most of equalization features in EqAPO are possible to show in VituixCAD. Biggest problem is 2nd order shelving filter with adjustable slope in dB/oct. VituixCAD supports Q-factor for 2nd order shelf. Some conversion table and application instructions are needed to swap between those two. Another challenge is selectable frequency notation of shelving LP/HP: corner or center. VituixCAD supports both, but final choice must be agreed and known by user if I add EqualizerAPO to 'DSP system' list in VituixCAD.
Bessel is phase normalized in EqAPO because only single biquads are available (Q=0.577 with 2nd order Bessel).
Q-factor notation of peak/notch filter is "Not analog RLC" in EqAPO i.e. BW is at peak_dB/2 - just like in miniDSP. This is piece of cake because notation is fixed in EqAPO.
---
We have at least one transfer option left and not discussed. You can export filter gain response of selected driver from VituixCAD Impulse response window as 16-bit PCM mono or 32-bit IEEE mono .wav. Then import IR wav file to EqAPO by adding Advanced filter -> Convolution. Parameters in VituixCAD for example: sample=44100, FFT=32768, taps=16384, IR win=Rectangular or Hanning, 'Center IR' not checked if filter is IIR. Convolver enables also phase linearisation with FIR. This might produce some hassle close to Nyquist frequency. I'll probably cut IR few samples shorter to prevent extra curves.
Last edited:
Blocks would be one possibility to help life of Hypex, EqAPO etc. users who have to play with 2nd order LP/HP sections.
For example block for 8th order low pass with Butterworth, Linkwitz-Riley and Bessel -3dB options:
Block calculates corner frequencies and Q-factor for each section depending on LP frequency and selected option. User can fine tune all parameters individually, but block takes control if Tune block window is re-opened.
This is just a test. This kind of blocks are not yet available in library.
For example block for 8th order low pass with Butterworth, Linkwitz-Riley and Bessel -3dB options:
An externally hosted image should be here but it was not working when we last tested it.
Block calculates corner frequencies and Q-factor for each section depending on LP frequency and selected option. User can fine tune all parameters individually, but block takes control if Tune block window is re-opened.
This is just a test. This kind of blocks are not yet available in library.
The relationship between shelf slope and Q is:
1/Q = sqrt((A + 1/A)*(1/S - 1) + 2)
Audio-EQ-Cookbook.txt
Rev. 2.0.5.3 (2018-10-12)
* Added 'Generic' option to DSP system combo box in Options window. BW of peak/notch filter is between midpoint gain (dB/2) frequencies. Bessel LP/HP is phase-normalized, -3dB frequency depends on order. Frequency parameter of Shelving LP/HP filter is Corner, not midpoint gain (dB/2). Setting is compatible with e.g. EqualizerAPO with some restrictions.
* Added 'Sample rate' combo box to Options window for 'Generic' DSP system.
* Improved active filter calculation and graphs above Nyquist frequency.
* Added 'Generic' option to DSP system combo box in Options window. BW of peak/notch filter is between midpoint gain (dB/2) frequencies. Bessel LP/HP is phase-normalized, -3dB frequency depends on order. Frequency parameter of Shelving LP/HP filter is Corner, not midpoint gain (dB/2). Setting is compatible with e.g. EqualizerAPO with some restrictions.
* Added 'Sample rate' combo box to Options window for 'Generic' DSP system.
* Improved active filter calculation and graphs above Nyquist frequency.
I am talking about comparison before and after applying the filter. In more complex crossovers, it is useful to bypass individual filters. I often use this function in LspCAD.
Sure this feature exists but it's not "by-pass" because it's too limited function.
Passive and active XO components can be electrically Open or Shorted with context menu - right click on component. This applies also to several components at once if group is selected.
An externally hosted image should be here but it was not working when we last tested it.
Drivers can be electrically Open or Shorted, or just acoustically Muted with existing load in the circuit.
An externally hosted image should be here but it was not working when we last tested it.
These features have been in version 2.0 since the beginning.
If this is what you meant, I suggest you read user manual very carefully and try what you find in context menu of each component and graph. Otherwise you could miss whole bunch of essential features.
OT nagging: Generally it is a bit frustrating that someone writes very long e-mails for asking "can I do this and that with VituixCAD?" or complaining missing features if they have been available and documented for months...years. Why should I repeat anything if information is already available.
Last edited:
KS-132s/Scanspeak driver pair - help with phase!
Cripes, what can I say but THANK YOU Kimmosto - totally blown away with the sheer power of this modelling sim, even though I'm only a nuub and learning the basics (no measurements yet).
What's really interesting is once I saw his KS-132s and thought I was back in my living room! I built these from exactly the same drivers when they first come out 2000? but have never been happy with my poor filter design, so I've dusted them down and started using Vituix to bring them back to life.
Used SPL Trace to grab the SPL and Z of both drivers and imported.
Then tried various combinations of 1st, 2nd and third orders (including Optimizer, BRILLIANT tool) and have struggled to achieve both a flat response and aligned phase as these drivers aren't well paired.
Here's a link to some project files, 1st order optimized v8 seems to be the favored, but am I doing something totally wrong, I'm so surprised that 1st order seems to be the flavor for these two?
https://1drv.ms/f/s!AgmKo7c0CB2xjA-2bWZyD--sDiFA
I understand that 'real' enclosures are designed using real measurements, but it's a start. You should see the horrendous filter I did back in 2000, bless!
Any thoughts welcomed.
Cheers
Lea
Cripes, what can I say but THANK YOU Kimmosto - totally blown away with the sheer power of this modelling sim, even though I'm only a nuub and learning the basics (no measurements yet).
What's really interesting is once I saw his KS-132s and thought I was back in my living room! I built these from exactly the same drivers when they first come out 2000? but have never been happy with my poor filter design, so I've dusted them down and started using Vituix to bring them back to life.
Used SPL Trace to grab the SPL and Z of both drivers and imported.
Then tried various combinations of 1st, 2nd and third orders (including Optimizer, BRILLIANT tool) and have struggled to achieve both a flat response and aligned phase as these drivers aren't well paired.
Here's a link to some project files, 1st order optimized v8 seems to be the favored, but am I doing something totally wrong, I'm so surprised that 1st order seems to be the flavor for these two?
https://1drv.ms/f/s!AgmKo7c0CB2xjA-2bWZyD--sDiFA
I understand that 'real' enclosures are designed using real measurements, but it's a start. You should see the horrendous filter I did back in 2000, bless!
Any thoughts welcomed.
Cheers
Lea
Last edited:
Hi Lea,
Your current layout looks pretty good though I would caution you that when you use supplied measurements as opposed to those that you measure yourself, you may be in a for a surprise when/if you obtain your own measurements. There are advantages and disadvantages to different order crossovers.
Be careful of enjoying this too much, however, as it is addictive and, as such, should have a warning label
Jay
Hi Lea,
Your current layout looks pretty good though I would caution you that when you use supplied measurements as opposed to those that you measure yourself, you may be in a for a surprise when/if you obtain your own measurements. There are advantages and disadvantages to different order crossovers.
Be careful of enjoying this too much, however, as it is addictive and, as such, should have a warning label
Jay
Wow thanks Jay, that has really given me a boost, thanks for your positive encouragement.
Indeedy re the measurements, I'm already wondering if Santa might bring me a DATs v2 and Mic measurement system ;-)
As a youngster (20 ish) I was exceptionally fortunate and privileged to meet and be briefly guided by the brilliant Mr Ruark as I have always been a passionate designer, admittedly I flit around electronics technologies, and this year I've rediscovered DIY speakers - last was headphone amps.
Didn't manage to get to bed until 2am last night, think I'm addicted again ;-) GREAT!
Thanks re the support of the filter design too (all credit to Kimmosto and Vituix).
I'm feeding the addiction...
Cheers again
Lea
So of course, now I wanted to compare the KS-132 crossover to my what appears daft 1st order....
Here's the original KS-132 modeled in Virtuix, the same run through the optimizer and then compared against my 1st order. The phase looks better and crossover point much lower which could be good for imaging (further from the bass driver's break up points), or is it? That's the thing with nuubs, I'm not sure how the phase 'should' look, apart from both blue and red lines aligned...
https://1drv.ms/u/s!AgmKo7c0CB2xjBVGdPInydtaWlc5
Any help appreciated, also tell me to shift off this thread if necessary, I wont be offended!
Cheers
Lea
Here's the original KS-132 modeled in Virtuix, the same run through the optimizer and then compared against my 1st order. The phase looks better and crossover point much lower which could be good for imaging (further from the bass driver's break up points), or is it? That's the thing with nuubs, I'm not sure how the phase 'should' look, apart from both blue and red lines aligned...
https://1drv.ms/u/s!AgmKo7c0CB2xjBVGdPInydtaWlc5
Any help appreciated, also tell me to shift off this thread if necessary, I wont be offended!
Cheers
Lea
I would go one step further than this - IMO you simply can't do a good job of designing a crossover in a simulator from manufacturers supplied measurements, because they don't include the effects of your cabinet, in particular the baffle step shift between midrange and bass. Drivers and cabinet work together hand and glove to give the final response.
And frankly, who knows what baffle the manufacturer measured on and how carefully they did it, we might assume it was measured on a very large ("infinite") baffle, but who really knows ? Some manufacturers measurements are also clearly "smoothed" and doctored to make them look better. You want raw measurements, warts and all for the simulator.
Not only does the cabinet affect the frequency response, it also affects the impedance curve of a woofer or midrange driver - and simulating the actual impedance curve of the drivers is critical in passive crossover design as the speaker impedance is effectively one of the components in the passive network.
At the very least if you did try to use manufacturers frequency response measurements you would need to pass them through a baffle step simulator (which vituixcad does include) however the accuracy of such simulations are still limited, and they do not simulate the changes in impedance curve as far as I know.
It's a bit of work but it really is best to learn how to take your own measurements including gated measurements for high frequencies, nearfield measurements for bass, and how to stitch them together properly. It's not easy especially in smaller rooms, but if you can master that you're well on your way to being able to design a good speaker.
I haven't looked at your files but I do have some general advice about the phase response. If you are just tracing manufacturers frequency response plots then using those directly in the simulator then you will never get the phase correct. (other than by accident
)This is because the measurements are only amplitude responses without any phase data. When the amplitude response is imported the "minimum phase" response is calculated for phase data. This assumes that the driver is minimum phase (most are, but some more exotic types dual cone drivers are not) and also does not include any time of flight.
Basically, the phase response that is generated has no information about any acoustic centre offsets between your drivers. Unless they are perfectly time aligned - which is very unlikely unless you have carefully chosen you drivers or stepped the baffle.
So your simulation won't be taking into account the real acoustic centre offset of the drivers so when the simulation says the drivers are in phase they actually won't be, and vica versa. The higher your crossover frequency the more serious this problem will become.
The right way to deal with this problem goes back to making your own measurements I'm afraid. What you would typically do is position the microphone to be equidistant from tweeter and midrange along the design axis of the speaker, then take both your measurements with the microphone in the exact same location, with the exact same configuration in your measurement software. And ideally using dual channel measurement to ensure that relative phase differences between the two drivers are maintained correctly in the measurements.
Once you have those measurements you have measurements whose phase includes any acoustic offsets of the drivers mounted in your cabinet, so when you get the phase to track in the simulation it will track in real life too.
Not much positive to say about KS-132. I had some hopes after purchasing drivers but sound was never "right". Woofer was too large with hard cone, and tweeter too prickly for low crossover point. In the beginning second woofer was broken (too low sensitivity) and I had to buy new pair (which was without Kapton).
Design method without VituixCAD was too simple and inaccurate to handle sound balancing also by power & DI responses. Finally project was just platform for few crossover and time alignment tests, stepped baffle etc. Published version is just some decent, but I can't recommend it.
Please do not trace SPL and Z from datasheets for crossover designing, though measurements are not very easy at home. Measurement gear should be dual channel i.e. USB mic and single channel measurement mode is bad choice.
Design method without VituixCAD was too simple and inaccurate to handle sound balancing also by power & DI responses. Finally project was just platform for few crossover and time alignment tests, stepped baffle etc. Published version is just some decent, but I can't recommend it.
Please do not trace SPL and Z from datasheets for crossover designing, though measurements are not very easy at home. Measurement gear should be dual channel i.e. USB mic and single channel measurement mode is bad choice.
Hi Simon
Just wanted to say a huge thanks for such a detailed response and guidance from clearly an experienced designer, all are really useful. From a novice's point of view I wonder how to go about designing something that whilst maybe not perfect would get us hooked, i.e. how to achieve the best results without the initial investment in measurement hardware - that's where I am right now.
I wonder if a measurement service exists? I know of designers who offer filter designs using the client's full enclosures, but that takes quite some of the fun and learning away.
One last note about driver offset phase, couldn't I physically measure this and add to the z-axis offset in Vituix (already have my cabinets which I constructed all those years ago)?
When I was discussing phase, I was referring to the GD and Phase graph, in that what should I be looking for? I think from memory it's to align both vertical traces?
Maybe time to dig out my books from the loft - or can you recommend one?
Thank you again, I really appreciate your guidance and time.
Lea
Just wanted to say a huge thanks for such a detailed response and guidance from clearly an experienced designer, all are really useful. From a novice's point of view I wonder how to go about designing something that whilst maybe not perfect would get us hooked, i.e. how to achieve the best results without the initial investment in measurement hardware - that's where I am right now.
I wonder if a measurement service exists? I know of designers who offer filter designs using the client's full enclosures, but that takes quite some of the fun and learning away.
One last note about driver offset phase, couldn't I physically measure this and add to the z-axis offset in Vituix (already have my cabinets which I constructed all those years ago)?
When I was discussing phase, I was referring to the GD and Phase graph, in that what should I be looking for? I think from memory it's to align both vertical traces?
Maybe time to dig out my books from the loft - or can you recommend one?
Thank you again, I really appreciate your guidance and time.
Lea
Last edited:
Hi Kimmosto
Thanks for your reply, really appreciated, and from THE designer of the year!
My 'KS-132' see here (I know you'll chuckle) built around year 2000...I actually quite like the sound but my 'un-golden' ears tell me there's a huge dip around xover, so I'm hoping to use your superb Vituix to achieve something enlightening in comparison, this I know will spur me on to purchase said measuring equipment and build something new - or perhaps you can recommend one of your designs?
All in all, for me the limitations are:
No workshop and limited tools to build cabinets
Limited knowledge, but I'm a speedy learner
No measuring equipment.
I guess something like a 10 step process for the novice which excludes measuring (due to initial costs) getting them going on their adventure, would be really useful? BTW, I am working my way through your extensive Vituix documentation, its a seriously wonderful reference.
How about the DATS v2 & OmniMic v2 - they probably wouldn't break my bank?
Your advice is more than welcomed, thank you.
Lea
Thanks for your reply, really appreciated, and from THE designer of the year!
My 'KS-132' see here (I know you'll chuckle) built around year 2000...I actually quite like the sound but my 'un-golden' ears tell me there's a huge dip around xover, so I'm hoping to use your superb Vituix to achieve something enlightening in comparison, this I know will spur me on to purchase said measuring equipment and build something new - or perhaps you can recommend one of your designs?
All in all, for me the limitations are:
No workshop and limited tools to build cabinets
Limited knowledge, but I'm a speedy learner
No measuring equipment.
I guess something like a 10 step process for the novice which excludes measuring (due to initial costs) getting them going on their adventure, would be really useful? BTW, I am working my way through your extensive Vituix documentation, its a seriously wonderful reference.
How about the DATS v2 & OmniMic v2 - they probably wouldn't break my bank?
Your advice is more than welcomed, thank you.
Lea
Last edited:
That is probably not so far from the truth imo too. It's really difficult get balanced and nice sound with too large hard cone woofer and tweeter which is prickly diva such as old Revelator. Another thing is that front baffle of your box and my KS-182 are larger than both originals. That could increase the dip.
Published KS-132 crossover (with 18W...K00 drivers) was mostly kinda improved version of original LDC6 by Vance Dickason. That had much larger and deeper dip at at mid and lower treble.
I just entered original measurement (axial) data to VituixCAD 2.0 and merged woofer data. Overlays are LDC6.
An externally hosted image should be here but it was not working when we last tested it.
I'm quite confident that balance would improve by changing big coil to 2.0 mH and 3.3 ohms resistor to 2.7 ohms. That is small change which won't make it perfect - not even close
OmniMic is USB isn't it? That I cannot recommend due to missing reference channel. Possible constant latency gear won't help if software cannot use signal from reference channel to maintain accurate timing differences between different measurements.
Focusrite Scarlett 2i2 2nd gen and Sonarworks XREF20 is good though not perfect. Mic calibration probably leave some small hump (1 dB) at top octave but that's okay when you know it.
Last edited: