SPL Trace is initially designed for datasheets etc. without phase data in SPL or Z so minimum phase is calculated automatically. But you can try to trace and export phase curve as SPL to separate file and copy-paste data from dBSPL column to Phase[Deg] column of the file with traced SPL e.g. with Excel.
Phase wrapping from -180 to +180 deg could cause some problems. It should be single step to avoid problems while unwrapping. 1/12 oct. smoothing should be off while exporting to maintain as steep steps as possible.
Generally SPL Tracing is not recommended action.
Rev. 2.0.20.0 (2019-08-20)
Improvements in 3D geometry calculation of driver and speaker/mic rotation. This is mostly preparatory work for possible off-axis response interpolation and plane/azimuth angle (phi) rotation.
Improvements in 3D geometry calculation of driver and speaker/mic rotation. This is mostly preparatory work for possible off-axis response interpolation and plane/azimuth angle (phi) rotation.
I have a question on Vituix.... Under the Crossovers Tab, while designing a crossover circuit, you can adjust the X, Y, Z physical parameters for the drivers. X & Y do little to nothing on the frequency response of your circuit unless placed far away, as expected, but Z will certainly alter your phase, and thus significantly impact your crossover design for even small amounts. I could not fully understand the Help file on this.....Where is Z = 0? Is this with tweeter and woofer at front of cabinet? Is it at the Coil Windings of both the tweeter and Woofer, so that most typical tweeters would need to be recessed a couple of cm in order to have Z = 0? Or is it based on the front, where if I have the tweeter windings lined up with the spider of the woofer, I would need to input a negative number? I imagine placing the tweeter farther back in the cabinet is a negative number for Z, right?
But what constitutes Z = 0?
“Front baffle surface?” So, if you have a waveguide that positions your tweeter back a little, you enter a negative number, correct?
Thank you for responding so quickly. I read various places in the manual dealing with this, but was uncertain of what it means. I just want to get it all correct, since my first attempt at a crossover was not the sound of the curve I expected to hear.
z=0 corresponds to the point where you referenced the microphone distance. I like to use the baffle front as z=0.
Example 1: if you measured the SPL with the microphone at a distance of 100cm from the baffle front, you'd tell Vituix that the microphone distance was 100cm and z=0 corresponds to the baffle front.
Example 2: if you set the microphone at a distance of 100cm from the dust cap, and the dust cap is 5cm behind the baffle front, I would tell Vituix the microphone distance was 95cm to have z=0 at the baffle front.
Example 1: if you measured the SPL with the microphone at a distance of 100cm from the baffle front, you'd tell Vituix that the microphone distance was 100cm and z=0 corresponds to the baffle front.
Example 2: if you set the microphone at a distance of 100cm from the dust cap, and the dust cap is 5cm behind the baffle front, I would tell Vituix the microphone distance was 95cm to have z=0 at the baffle front.
I am really not trying to be dense here, but....first off, at this point I am not measuring, I am designing and predicting while building a crossover in software. If you mean hypothetically placing a microphone, then at what place is the measurement taken on the drivers for a basic two-way, woofer/tweeter setup? A theoretical mic to a baffle would be the baseline of zero, but where, exactly, on the speakers do you measure the offset from zero? At the front surface of the baffle - so that if you set a tweeter back into a waveguide it would be a negative number? What about the woofer? Flange mounting at the baffle, center surface of the cone, or back at the coil and spider?
In some respects, it doesn't matter if the zero point is the mic or the baffle or the back of the woofer or the magnet. It could be the wall. The point is that one reference point is chosen, and you set the number for the other one. But when determining that number for two drivers as distinctly different in shape and size as a woofer and a dome tweeter, how do you determine the point on the other driver in relation to the first?
I tend to think that it would be the wire coil within the speaker. That would be the only point that makes sense across two different driver types, but I really don't know.
I just notice a huge difference in phase - and thus dramatic differences in the crossover when you adjust the Z number even as little as 10mm. There is some noticeable difference at even 1 or 2 mm. So, the exact point on the tweeter and woofer that you are referring to when you say that it is offset by 33mm is a critical measurement.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
As far as I can tell, the idea with the Z number is to tell the software how these two things are lined up, basically in relation to each other, so the software can deal with any phase shift this placement causes. So, if my reference is the front edge of the enclosure (and drivers are surface mounted, then the point equals the back edge of the mounting surround of the woofer. Woofer Z=0, and my tweeter is also at Z = 0. Or, If i go by the front edge of the driver, woofer Z= +6.5 and tweeter Z=+3. If I go by where the coil of wire is inside each speaker, my woofer is Z= -18 and Tweeter Z= -1. If I go by the back of the magnet, woofer Z= -75 and tweeter is Z= -21. When a couple of millimeters causes slight phase cancellations at the crossover point and noticeable dips in the curve, the point on the speaker you reference matters significantly, and I just don't know what point that should be. And different points of reference mean different numbers. These example show a difference of +3 to -54. That is a huge effect on your crossover in this software.
Last edited:
...oh, also, am I backwards on my negative and positive? If it is the distance from mic to enclosure, and then recessed into the cabinet, it stands to reason that would be a positive number instead of negative as I have it.
(My brain had been viewing this from the point of view of the enclosure, so recessing would be negative. From a mic, I was estimate positive. Right?
(My brain had been viewing this from the point of view of the enclosure, so recessing would be negative. From a mic, I was estimate positive. Right?
Hi SSmith3 think sounds you stress a bit about it, maybe take a break or day off from subject before going back at it : )
Some pointers:
Guess VituixCad doesn't care if its baffle or if its woofer transducer acoustic center or if its tweeter transducer acoustic center that is the reference point at Z=0, its the difference distance if any in real world build on design axis that matters and if that number is say 33mm it doesn't matter if you delay one (positive number for Z axis in a longer distance to microphone) or make the other in advance (negative number for Z axis for a shorter distance to microphone). A note is if one use DSP for filtering we can forget about the problem because DSP can offset the physical acoustical timing error, problem is only there if using passive filters and transducers acoustical center is not design to be physical offset so they fire in sync relative to design axis. In real world the longer the distance to speaker is the little less in milimeters the error become also listening height will change the physical distance.
When you simulate a design probably using datasheet responses remember that response include whatever baffle step and difraction that manufacture used as test baffle dimensions and be shure for a real world speaker builds other dimensions that the response will be more or less much different in response curve, about datasheet resonses one can threat them as Z=0 as long as physical build take care the acoustic centers is alligned or one use DSP, and if not that is possible you should set some estimated numbers into Z axis real world mitchmatch.
EDIT: By thew way have you looked up the measurement preparation guide at this link https://kimmosaunisto.net/Software/VituixCAD/VituixCAD Measurement Preparations.pdf
Some pointers:
Guess VituixCad doesn't care if its baffle or if its woofer transducer acoustic center or if its tweeter transducer acoustic center that is the reference point at Z=0, its the difference distance if any in real world build on design axis that matters and if that number is say 33mm it doesn't matter if you delay one (positive number for Z axis in a longer distance to microphone) or make the other in advance (negative number for Z axis for a shorter distance to microphone). A note is if one use DSP for filtering we can forget about the problem because DSP can offset the physical acoustical timing error, problem is only there if using passive filters and transducers acoustical center is not design to be physical offset so they fire in sync relative to design axis. In real world the longer the distance to speaker is the little less in milimeters the error become also listening height will change the physical distance.
When you simulate a design probably using datasheet responses remember that response include whatever baffle step and difraction that manufacture used as test baffle dimensions and be shure for a real world speaker builds other dimensions that the response will be more or less much different in response curve, about datasheet resonses one can threat them as Z=0 as long as physical build take care the acoustic centers is alligned or one use DSP, and if not that is possible you should set some estimated numbers into Z axis real world mitchmatch.
EDIT: By thew way have you looked up the measurement preparation guide at this link https://kimmosaunisto.net/Software/VituixCAD/VituixCAD Measurement Preparations.pdf
Last edited:
Thanks, Byrtt.
It’s just that it seems like a simple question, and the answers seem inadequate or vague. And more importantly, the difference at the crossover frequency of the 33mm difference between the two main potential ways I could see this Z measurement being defined makes a huge dip in my response curve, with a drop of 12-15dB. It isn’t an inconsequential number to the results, and I don’t know which number to use. It’s practically like dropping in a notch filter.
If the Z number can cause that dramatic of a notch in my crossover design, I want to make sure I am getting the right number in there and adjust before I buy new caps and inductors for my passive crossover.
But you are certainly right that I would be better off measuring the response than using predicted data. I may do that. At least then, whatever effect the Z number would have should be included in the recorded test data - not to mention the other acoustics of my specific enclosure.
Thanks.
But it would still be nice to know what physical part of the speaker is the point of reference to measure Z. But I can’t seem to get a simple answer on that question. You refer to “acoustic center”. Where, physically, is the acoustic center on a dome tweeter and a woofer that would be the points you use to determine Z?
But you DID give me a simple answer for + vs - on Z. Thanks for that.
It’s just that it seems like a simple question, and the answers seem inadequate or vague. And more importantly, the difference at the crossover frequency of the 33mm difference between the two main potential ways I could see this Z measurement being defined makes a huge dip in my response curve, with a drop of 12-15dB. It isn’t an inconsequential number to the results, and I don’t know which number to use. It’s practically like dropping in a notch filter.
If the Z number can cause that dramatic of a notch in my crossover design, I want to make sure I am getting the right number in there and adjust before I buy new caps and inductors for my passive crossover.
But you are certainly right that I would be better off measuring the response than using predicted data. I may do that. At least then, whatever effect the Z number would have should be included in the recorded test data - not to mention the other acoustics of my specific enclosure.
Thanks.
But it would still be nice to know what physical part of the speaker is the point of reference to measure Z. But I can’t seem to get a simple answer on that question. You refer to “acoustic center”. Where, physically, is the acoustic center on a dome tweeter and a woofer that would be the points you use to determine Z?
But you DID give me a simple answer for + vs - on Z. Thanks for that.
Welcome Ssmith3 and yes seriously before ordering any components get some measurements of system in situation : )
About real acoustic center for a dome tweeter and a woofer is probably real complicated from unit to unit, myself don't care so much because i use DSP and can offset to whatever numbers and get documentation via step response or scope in measurement program, but have read some pointers about real acoustic center over time and as far as i remember its a physical point somewhere in front of center of voicecoil where the energy will start as electric signal, so probably you could estimate a point a little in front of voicecoil for your drivers and then use that numbers and their physical position on baffle to get to a close estimate to use in VituixCad simulation. Also if i remember correct real point of acoustic center can sometimes be frequency dependable onto some transducers, if that is true and two physical different drivers is alligned and sums say a 1kHz sinus perfect in cycle, then they will drift a bit in cycle at say 2kHz, for such a case of allignment for a sinus wave guess one should look for allignment of acoustic center is optimized at same frequency as one intend to slope the XO.
Last edited:
Typically with unidirectional i.e. boxes speakers, difference between straight non-stepped front baffle surface and acoustic center of each driver should NOT be adjusted with Z mm parameter. Z should be 0 mm for all drivers. Adjusting with Z is legacy error in DIY scene inherited from old-fashioned tools which were not designed for off-axis response simulation. Correct parameter is Delay [us] of driver which handles timing properly also to rear sector (90...180 deg): positive delay stays positive also to 180 deg off-axis.
Using of Z mm parameter for timing adjustment is meant for stepped baffle.
Z works also with dipole radiators because positive Z becomes negative when mic/ear is rotated to 180 deg off-axis.
Using of Z mm parameter for timing adjustment is meant for stepped baffle.
Z works also with dipole radiators because positive Z becomes negative when mic/ear is rotated to 180 deg off-axis.
Basically DSP does not change anything compared to passive XO if you design with simulation wanting to simulate+visualize also power & DI and other off-axis. Both need measured timing (compared to rotation center) for each driver model, and setting of geometry parameters X,Y,Z,R,T for each driver instance.
Most of my latest designs are DSP with IIR, FIR or mixed. Only way to design with possibility to make controlled changes even in remote is making "everything by the book".
I see. I was in this boat, too. I took me a while to realise that Vituix is not based on the idea of "forward simulation of everything".
Think of Vituix as a data processing tool. You feed it with measured data (which is referenced to whatever z=0 you used in the measurement), and then Vituix processes this data (by considering the x-over filters and geometrical offsets between the drivers). The output is referenced to the z=0 in your measurements.
Thanks guidance and imagine work by the book get rocking good results : )
Will say in my own two way active DSP system i can be a bit lazy and jump over all that work because a 10F/8424 is the system dictator for dispersion in its sloped with 96dB/oct linear phase XO covering 240Hz and up, below one woofer sits within 1/8 wavelenght and cover the rest. System is on wall and guess nothing one can do about mid tweters inherent dispersion in begin to measure all the way around, what helped me enorm for realistic good room sound is set speakers exactly mounted position into room (you know relative distance numbers to bounderys and listening position etc) into one of Jeff Bagby's spreadsheet and use that curve in reverse into DSP, that got me close to exelent in room sound not to speak of with open door in nearby rooms one can hear that correction made wonders so speaker response is kind of tuned to room.
Rev. 2.0.21.0 (2019-08-26)
* Acoustical frequency responses interpolated within measured off-axis directions including diagonal angles.
* Acoustical frequency responses interpolated within measured off-axis directions including diagonal angles.
So, I made a custom waveguide, mimicking the Buchardt s400’s. This sets the tweeter back into the enclosure by 33mm. I attempted to design it so that the depth of the waveguide put the voice coil on the tweeter drive at the same Z measurement as the voicecoil on the woofer. But, if the front of the baffle is actually the reference point, then I am -33. Or +33? The difference between -33, 0, and +33 makes quite a difference and a noticeable dip in the sound at the crossover frequency. This is the source of my issue. I really don’t know if when accounting for that waveguide, if I am -33, 0, or +33mm.
But in the end, I am really trying to set up to take real measurements instead of using theoretical specs, as I am sure that will give me the best results.
Rev. 2.0.21.1 (2019-08-27)
* Axial response is selected if off-axis directions are not measured, though driver's location, rotation or tilt would require at least single off-axis measurement.
Logically at least single off-axis response is needed to locate driver other than origin X,Z=0,0 or rotate or tilt. I suppose missing off-axis measurement data is still standard so I made workaround to rev. 2.0.21.1: if only one frequency response is loaded, that response is selected for all locations, turning and directions. Axial result could be okay if X,Y offset is not too long compared to directivity of radiator (below XO frequency).
* Axial response is selected if off-axis directions are not measured, though driver's location, rotation or tilt would require at least single off-axis measurement.
Logically at least single off-axis response is needed to locate driver other than origin X,Z=0,0 or rotate or tilt. I suppose missing off-axis measurement data is still standard so I made workaround to rev. 2.0.21.1: if only one frequency response is loaded, that response is selected for all locations, turning and directions. Axial result could be okay if X,Y offset is not too long compared to directivity of radiator (below XO frequency).