WinPCD Results Different Than Xsim/VituixCAD

[Impious]

My results from WinPCD and wildly different than Xsim or VituixCAD. But the results from Xsim and VituixCAD are extremely similar. Soooo....either I missed moving something over correctly in BOTH Xsim and VituixCAD, or I'm missing something that's wrong in WinPCD. But for the life of my I can't figure out what's different between them.

I've verified I'm using the same measurement files in all three programs. I've got the Z offset in all 3, and I believe for the correct value. I measured on the tweeter axis for both drivers so I don't need an X or Y offset. I've double checked the circuits and from what I can tell they all match.

Attached are results from all three programs, with FR both in phase and with polarity reversed.

What am I missing here?

[Impious]

Max attachments was 5 so here are the VituixCAD snaps

[Wolf]

Looks like in PCD you shorted the woofer coil. The other 2 you did not.
Wolf

[Impious]

Looks like in PCD you shorted the woofer coil. The other 2 you did not.
Wolf

I thought about that after I posted. So I switched the coil location in WinPCD after I posted. I get the same result.

[dcibel]

In PCD, check horizontal and vertical off-axis angle are set to zero, hidden in your screenshot.
​

In VituixCAD, it won't changer your results to any meaningful effect, but for basic single axis, single channel measurement, it's preferable to enter delay value in driver tab. Offsets in crossover schematic used for physical offsets, leave at 0,0,0 if measurements completed with stationary USB mic.

[Impious]

Got the issue resolved. Z offset should be entered positive in VituixCAD & Xsim. I entered it as a negative value as it's entered in WinPCD.

Such a Duh! moment

[dlr]

Got the issue resolved. Z offset should be entered positive in VituixCAD & Xsim. I entered it as a negative value as it's entered in WinPCD.

Such a Duh! moment

That was going to be my first question. WinPCD, as in the PCD spreadsheet use absolute 3-D space positioning. I took this approach after having spent years working with CALSOD that uses the same scheme. No knowledge of the mic (design point) location is required to position the drivers and mic, the 3-D point positions are representative of precisely where they are located in 3-D space, relative to the origin (0,0,0), whether or not any particular driver is set to that point. All positional references are in relation to the origin, no delays to the mic. If you were to plot them on a 3-D graph, you would be able to visually see their positions based on the (x,y,z) with the z value placing them either in front of the origin plane (positive z), on the plane through the origin (usually considered to be the baffle front) or behind the origin plane (negative z). Early on I convinced Jeff to implement the 3-D scheme as used in CALSOD as it seemed more intuitive to me, having used CALSOD starting in 1995. Jeff had long experience with it as well. This makes all points in front of the origin positive. It also made it easier to code, especially when calculating off-axis positions.

Xxim (and apparently) VituixCAD invert this. That is, positive z values place a driver farther away from the mic/design point. All programs compensate for the excess-delay introduced whatever scheme is used.

I'll add some explanation for those not familiar with WinPCD/PCD since I think there is occasional confusion on this issue between the apps.

Driver and mic (x,y,z) values are all relative to (0,0,0). If the mic is set at (0,0,1) that places it one meter in front of the (0,0,0) location, no matter the positioning of the speaker. Tilted or stepped baffles are the same. The other drivers are all relative to (0,0,0). Their relative acoustic offset, when determined appropriately, will correctly position them relative to the mic point. If the tweeter is used as reference (0,0,0) then all drivers and mic are relative to the tweeter being at the origin. This is most often the case if measurements are made on the tweeter axis with the mic 1m in front of the tweeter.

If you use WinPCD and run the power response calculations, you'll be able to see this in action. The graph on the right of the window displays the mic/design point moving in real time with the calculation. It is 2-D, so at first appears a bit odd. However, it can be visualized as creating the surface of a sphere with each movement, the so-called Spinorama as seen from the front. If you could rotate it as a 3-D graph you would see the front half of a sphere. All changes in off-axis points is moving the design/mic point relative to the origin, (0,0,0).

There is one caveat. A tweeter almost always has some z-offset from its faceplate front. That places its acoustic center a small amount behind the faceplate. When the faceplate is mounted flush to a baffle, its acoustic center is actually a small bit behind the front baffle. But as long as all other drivers have had their relative acoustic offset determined appropriately, all off-axis calculations are still correct, just relative to the small tweeter offset.

dlr

[Wolf]

Don't forget PCD was also like the other 2 with positive Z until Jeff changed it in a later edition to reflect as you just described as voted by the masses. As long as you use one of the 2 most recent versions of PCD, or WinPCD, then Z is negative.
This has been brought up before as someone had and older version.
Wolf

[wogg]

So I've done some learning from this thread! Of course, only after I pulled the trigger on parts for the crossover design using incorrect offsets derived from WinPCD entered into VituixCAD so I'm going to use this as an experiment and learning experience.

Here's the design result using the WinPCD derived offsets, using a tweeter + mid and mid + woofer measurement to align the summation between the drivers in the software when the Z offset is adjusted. For this the X and Y dimensions were entered as measured on the baffle, X and Y values positive to the right and up X,Y and negative to the left and down.



So based on this thread, I decided to derive the offset in Vituix directly using the same method, Mid and tweeter alignment:



And Mid to woofer alignment:



That landed at very different offset settings.



Process to do this:

• Load overlay for mid+tweet measurement
• Mute woofer
• Adjust Z until the shape hits
• Clear overlay
• Mute tweeter and enable woofer
• Load overlay for mid+woofer measurement
• Adjust Z until the shape hits

That has led my perfectly sweet crossover design to look like this now:



That's now just... not good.

Since I have the parts en route, here's what I'll do. I'm going to prototype (non-destructively) one of these together and take close to the same on axis measurements. We'll see which sim it looks like, and if we're all correct, it'll look closer to this last chart with the big ol'; 7dB broad dip at 4kHz and nearly no phase alignment.

Then I can go to my very long history as a PE customer and their excellent customer service to get the right parts

[dcibel]

For single axis design using usb mic at single point in space, use aux menu time align tool to determine delay value, enter delay into driver tab, leave dimensions in crossover at 0,0,0. Result will be same as xsim.

In vituixcad, driver coordinates in crossover are meant for physical offsets only when 3D (off-axis) response data is included, and measurement is 2-channel at driver axis.

[wogg]

For single axis design using usb mic at single point in space, use aux menu time align tool to determine delay value, enter delay into driver tab, leave dimensions in crossover at 0,0,0. Result will be same as xsim.

In vituixcad, driver coordinates in crossover are meant for physical offsets only when 3D (off-axis) response data is included, and measurement is 2-channel at driver axis.

So what if I'm doing a bit of both? My intent was single point on axis to lock in offsets and do full response design using merged and min phase woofer close mic and port data. Then take that same design validated against horizontal polar measurements.

Also... in the single axis scenario, min phase or no?

[dlr]

If I understand you, then minimum-phase, absolutely, if you are doing to go with single point only and intend to model off-axis from them. That's the only way to get valid off-axis modeling with single point.

[dcibel]

So what if I'm doing a bit of both? My intent was single point on axis to lock in offsets and do full response design using merged and min phase woofer close mic and port data. Then take that same design validated against horizontal polar measurements.

Also... in the single axis scenario, min phase or no?

No no no...Delay determined by 3-measurement process with USB mic applies to that measurement only. You are essentially "locked in" to the mic location for design. Once you move the mic or rotate the speaker the delay is invalid. acoustic centre is a moving target off-axis, hence the requirement for 2-channel measurement setup to capture timing effortlessly without many extra steps with each angle.

You can measure the complete speaker polar data with USB mic, but for combining multiple measurements from multiple drivers, a timing reference should be pre-requisite.

With USB mic the only process I could suggest is "acoustic timing reference" with REW, but it is much more painful than simple 2-channel electrical loopback, since acoustic reference distance to mic must be maintained at all times, moving mic to measure different driver on it's axis becomes difficult, better to lift speaker while mic and acoustic reference remain stationary. Repeatability is low if you have to come back and re-measure anything.

Min phase is mostly pointless. Only use is for merging nearfield and port response, because mic will be moved for each measurement, min phase provides alignment for this for best merge result. For splicing near and far field data, use aux-time align tool to determine delay with far field measurement, then merge tool. Min phase / HBT is only destructive in that process.

If it helps, I attempted and capture the benefit of 2-channel measurement systems for speaker design in this document I put together recently:
https://drive.google.com/file/d/1orY...usp=share_link

I also have plans to write another document on why minimum phase is mostly pointless and can be forgotten. Maybe something I'll try and put together over my xmas time off, we'll see.

[dlr]

Minimum-phase data is essential if software that can model off-axis data is to be used. You cannot accurately model off-axis data without it. There is nothing destructive about HBT generated phase, it is, in fact, required to generate minimum-phase data which is itself required to model (not use measured) off-axis response.

If you are going to use software that uses a large number of off-axis measurements directly, that's a separate issue, but a blanket statement that minimum-phase "can be forgotten" is not accurate considering most design software that can model off-axis. It's not helpful to those who are going to use software that can work with single point measurements. My comments will be targeted to that scheme.

Two-channel measurement systems are the best option if you want to use measured phase and use it directly, but it's still good only for a single point in space (the measurement point). Those measurements cannot be used to model the off-axis response. Measuring each driver on its axis does add some complexity and in my opinion is best avoided except for those with some experience with that.

However, measured phase is not even a requirement to get a valid set of measurements to use. I've pointed this out many times over the years. HBT generated minimum-phase is what allows this to work. The three-measurement scheme together with spliced close-mic responses and HBT-phase of the resulting driver measurements will provide an accurate relative acoustic offset that is then good for both on-axis and off-axis modeling. Measurement systems that may be difficult to work with to ensure accurate timing between measurements can be more difficult, but measured phase isn't a requirement for most software. I can take any measurement system of any kind, single or two channel, and get exactly the same results in design software for single point that can then be used to model off-axis. The key is getting good SPL files with good low end splices (even better if step response is also incorporated) that are then used to get the HBT phase. There are limitations, of course, in the off-axis, but the most important aspect is the first arrival (on-axis) with power response second.

Power response will be roughly independent of the baffle other than the step response area. For a particular baffle, calculated power response will not have much influence due to above-step diffraction because the baffle doesn't change the integrated output response of the system to any great degree. The diffraction will affect individual off-axis points to a great degree, but not power, so modeling power response is a fairly accurate method. Driver center-to-center, the crossover used and driver directionality have the most impact on power. Granted that driver directionality will be an assumption as well, but it can be reasonably estimated.

In brief, minimum-phase data is essential in many situations. It depends on the software being used and the degree to which it will be used.

For those readers unfamiliar with the three-measurement scheme, I have an article at my site that describes it, Finding Relative Acoustic Offset Empirically. As wogg alluded to, it's a built-in function in WinPCD.

dlr​

[dcibel]

Two-channel measurement systems are the best option

Got that part right ;)

Mininum phase is not required for "modelling" off-axis data for anything you've described. Only delay between measured data is needed, which doesn't need to be minimum phase. Measured phase from simple impulse response peak detection may not be minimum phase, but somewhere close, and it does not need to be modified via HBT for accurate delay / phase interaction. HBT can include user error from poorly chosen slopes causing error. min phase / HBT steps can be skipped with same results, just different delay value.

Why "model" off-axis at all and leave the mic on the shelf? At best you have a model of crossover interaction off-axis, maybe piston model for the driver? What about diffraction, what about driver breakup, what about waveguides and horns and other complex arrangements? Accurate simulation comes from​ minimizing calculated assumptions.

However, measured phase is not even a requirement to get a valid set of measurements to use. I've pointed this out many times over the years. HBT generated minimum-phase is what allows this to work.

"3 measurement scheme" with mic at tweeter axis is problematic, one of those measurements is already off-axis, and simulation is locked at mic measurement distance. 1m measurement distance is not the same as 2-3m listening distance, okay for smaller speaker but not for large speakers, with some large tower with woofer down near the floor you can be capturing SPL error simply from the distance differences from mic to driver, that doesn't translate at real listening distance.

Better method would be to measure each driver straight on, not only will on-axis be actually on-axis for each driver, but constant mic distance ensures correct SPL as well. But now how to capture the delay with USB mic? 3 measurement process no longer works, so more works is needed, or timing reference (2-channel). Continue with off-axis measurements and now 1m measurement can be represented at 3m listening distance with just some response interpolation and SPL/distance calculation.

Anyway, it's clear that you are biased towards use of your own software, and I'm biased towards measuring systems that include timing reference and use of VituixCAD for design. We can move on from this discussion, it's clear that neither of us is going to convince the other of a "better way". Don't get me wrong, I started DIY using a Omnimic, following Jeff's instructions, 3 measurement process etc. but have moved on from that having learned more along the way with each new design adventure.

As far as 2-channel measurements go, I don't understand why there is so much resistance to include a timing reference in the measurements to capture accurate phase relations between multiple measurements and drivers. It's not like a cheap USB audio interface and XLR mic is all that more expensive than a USB mic. It's like some DIY spirit of needing to prove that you can get by with a paper clip and a rubber band while the correct tool is just an arms length away.

Btw, if you've been paying attention, the 3 measurement process is automated in VituixCAD in aux menu - time align, just load the 3 measurements and hit solve button to find delay.

[wogg]

Holy crap you guys are both super helpful... I'll have to come back and digest this again later but here's the bottom line.

• I didn't even think of transitioning to 2 channel measurement. I updated to an EMM-6 and have a ton of high quality mic channels at my disposal. Not resistant, just sheer routine. I attacked it like I did with my old setup.
• There are still lots of already built in tools in both PCD and Vituix that I'm not using

By far the biggest revelation is use 2 channels dummy! My parts will sit around until I re-do the measurements and start the crossover design again with a much better data set. This will all end up in a classic Wogg style over-written article later, with props to you both dlr and dcibel for helping me in the right direction.

Cross link to the design I'm working on: Anarchies

[DeZZar]

Two channel timing reference for the win. I also grabbed a reliable DB meter some time back to calibrate SPL.
I have a little process I go through before measurements to set 2.83volt output precisely, setup timing reference loopback and SPL level setting - reliable repeatability!!