I would like to time align the impulses of my next project. I have the ability to, so I want to.
Let me know if my calculations are correct. I will use data from an old project. It is probable that the speaker was not perfectly level on the table when I took these measurements so the actual Z offsets of the drivers might be thrown off a bit. Just let me know if I am calculating this correctly or if there is a flaw in my workflow.
Sound moves 343 mm in 1 ms.
Measurements were made using a loop back as the timing reference.
This is the woofer. -0.043 Ref assumes an offset of -14.7mm since sound will move that far in -0.042 ms
Here is the mid. -.127ms Reference assumes the sound moved 43.65 ms in that time and so its offset is -43.65mm. So when a tone plays through both the woofer and mid, the mid would need to be 28.95mm closer to the mic or listening position to play at the same time as woofer
Here is the tweeter. -0.184 ms Reference. The tweeter is offset -63.11mm back. It would need to move forward 48.4mm to match the woofer. It is a planar tweeter so maybe it does.
To summize: -14.7 for the woofer. -43.65mm for the mid. -63.11 for the tweeter.
If we reset the woofer to 0 Z offset we have -28.95 for the mid and -48.41 for the tweeter.
Time mechanically time align these drivers on their respective baffle the mid has to come forward by 28.95mm and the tweeter 48.41 mm. When a tone is played, they would all play at the exact time if they were offset this amount.
In this particular design the woofer is shallow, the mid cone is deep, and the planar is inset quite a bit. So this does seem to track with what I can see mechanically
Is this correct?
I know VituixCAD pulls this impulse data off of the FRD automatically. Is there a way to view that outside of the impulse tab? All I could figure out with the impulse tab was that you had to import a .wav file. I couldn't really figure out if there was a good way to pull this data out of that software.
Let me know if my calculations are correct. I will use data from an old project. It is probable that the speaker was not perfectly level on the table when I took these measurements so the actual Z offsets of the drivers might be thrown off a bit. Just let me know if I am calculating this correctly or if there is a flaw in my workflow.
Sound moves 343 mm in 1 ms.
Measurements were made using a loop back as the timing reference.
This is the woofer. -0.043 Ref assumes an offset of -14.7mm since sound will move that far in -0.042 ms
Here is the mid. -.127ms Reference assumes the sound moved 43.65 ms in that time and so its offset is -43.65mm. So when a tone plays through both the woofer and mid, the mid would need to be 28.95mm closer to the mic or listening position to play at the same time as woofer
Here is the tweeter. -0.184 ms Reference. The tweeter is offset -63.11mm back. It would need to move forward 48.4mm to match the woofer. It is a planar tweeter so maybe it does.
To summize: -14.7 for the woofer. -43.65mm for the mid. -63.11 for the tweeter.
If we reset the woofer to 0 Z offset we have -28.95 for the mid and -48.41 for the tweeter.
Time mechanically time align these drivers on their respective baffle the mid has to come forward by 28.95mm and the tweeter 48.41 mm. When a tone is played, they would all play at the exact time if they were offset this amount.
In this particular design the woofer is shallow, the mid cone is deep, and the planar is inset quite a bit. So this does seem to track with what I can see mechanically
Is this correct?
I know VituixCAD pulls this impulse data off of the FRD automatically. Is there a way to view that outside of the impulse tab? All I could figure out with the impulse tab was that you had to import a .wav file. I couldn't really figure out if there was a good way to pull this data out of that software.
Last edited:
I can't follow your measurements, but I suspect an error. If this is a flat, and vertical baffle. the tweeter will likely need to be delayed relative to the woofer. The mid may be close to the same as the tweeter. It depends on type, and size.
I've seen software where the tweeter is set as the zero position, and other drivers are relative to it.
I've seen software where the tweeter is set as the zero position, and other drivers are relative to it.
I also suspect and error. I assume this flawed and would like to know how.
Generally, yes, you set the tweeter to 0 position and offset the other two relative to it. This is what I do when using VituixCAD. It was just a little easier to explain basing it off of the woofer. This woofer is quite shallow, the mid is a deep cone, and the tweeter is even deeper so this does make some sense. I will add a picture to the main post. Maybe that will be helpful.
I never even consider z offset. I know it exists, but I rarely use a sim for designing a x-over for my speakers, and I usually slope the baffle. If I'm doing a sim, I make a WAG at the offset. That's usually close enough to get started. Details get worked out later.
I am trying to work toward ironing out details from the get go. Things like this Z offset, using the baffle to actually improve the FR of the system rather than having to deal with it at the end, positioning drivers on the baffle such that they cross more effectively and without so much cancelations off axis in either horizontal or vertical axis, etc.
I am moving away from the "I will fix it in post" method and more towards the "I want to plan this from the beginning, correctly, to give me the best result in the end"
The other thing to note is you need an acoustic reference in the measurements (I found this out the hard way), if you use REW you can either use a second speaker on the second input channel (this speaker is offset and never moves) or you use a loop back form the second channel back into the second mic channel. There are descriptions on how to do this if you search. This ensures that every measurement is aligned at the same point in time i.e. then the signal is sent from the sound card.
The other thing is that if you take all measurement switch out moving the mic position then the offsets ar beaked into then and you don’t have to worry about adding them in Vituix. That is my understanding.
When I did the loop back method my measurements came out pretty close to a bit of Pythagoras and also physical measurements.
The other thing is that if you take all measurement switch out moving the mic position then the offsets ar beaked into then and you don’t have to worry about adding them in Vituix. That is my understanding.
When I did the loop back method my measurements came out pretty close to a bit of Pythagoras and also physical measurements.
These measurements were made using a loop back as the timing reference. Updated main post to mention this
If those measurements show the time of flight difference, you would still need to account for the distance difference caused by the woofer being far below the tweeter, etc.
The time of flight difference is not the same as the z offset. Some geometry is needed to determine that. The software may do it for you.
The time of flight difference is not the same as the z offset. Some geometry is needed to determine that. The software may do it for you.
Last edited:
I don't actually know what that is.
I also feel really dumb right now because I just realized that REW calculates this all for you for each driver
Per usual, I will leave this post up here just in case someone else starts searching for the same thing or in case someone else doesn't know REW has this information from and center for you to see. I do not mind feeling dumb if it helps someone else.
You might as well explain the results. I'm not sure that I see the z offset for the drivers. Maybe point it out for us.
Is that saying that the woofer z is -46mm relative to the tweeter? And what was the mid Z?
Did you tell REW the X, and Y for the drivers? Also the measurement distance? All that would matter, otherwise, it's a WAG.
Is that saying that the woofer z is -46mm relative to the tweeter? And what was the mid Z?
Did you tell REW the X, and Y for the drivers? Also the measurement distance? All that would matter, otherwise, it's a WAG.
This is not always a single number, it can vary by frequency for a given driver.
Didn't you say you used a second channel looped back as your timing reference? Then you don't need the acoustic reference. That is there as a workaround for people that use USB mics.
I did wonder if this was the case.
My thinking is that I take the tweeter and the mid, mount them in a random board, take an impulse measurement at the frequencies of where I plan to cross them, and then I can mechanically align them from there.
Same thing then with the mid to the woofer.
This way they are mechanically aligned very near their crossover frequencies to one another
There's another way. I never made it work. Maybe tried it once. I may miss a detail, or two.
Measure each driver alone.
Measure them together.
Use a sim software and frds made from the measurements. Adjust the offset in the sim until the summed sim result looks like the measured summedresult.
Measure each driver alone.
Measure them together.
Use a sim software and frds made from the measurements. Adjust the offset in the sim until the summed sim result looks like the measured summedresult.
This is kind of what I am going to try. I'll leave the test baffle set up and then I'll 3d print spacers the sizes I think I need. Then I'll measure again. The impulses should line up. I genuinely love this kind of stuff.
And yes, I do realize that the position of my ears throws this all out of the window but it is a better start than it would be without it
In principle it is "trivial" to determine acoustic center offset. I've described it before including last night. In practice I've done it many times to within 1mm repeatably.
In other words, the driver facing upward serves as z-axis "reference" with known acoustic center, namely its physical center. Being the same model drivers, they share common phase. The rest is a matter of using measuring tools.
In other words, the driver facing upward serves as z-axis "reference" with known acoustic center, namely its physical center. Being the same model drivers, they share common phase. The rest is a matter of using measuring tools.
Last edited:
Are you measuring them both on the same baffle and the jsut dropping the mic to the center of the next one? Seems that if you do, REW will tell you z offset right there in your measurement.
I like your method though. You can just move them fore or aft in the sim until you have them where you need them for your XO phase shifts. This makes sense
I think this was the method used by the late Jeff Bagby in his instructions for PCD spreadsheet.
@AllenB agreed, probably not the best description it is either an acoustic (fall back) signal or the loop back electrical signal not both. Thanks for the correction.
