Hi all,
I recently built a baffle and put some drivers in it and was happy with the physical result. I had boomarked some information on how to best measure them on the baffle and produce FRD and ZMA and all the links are either dead or the software is dead and I can't follow a tutorial, etc, because of the project being dead. So I'm back to square one. I realize this is probably simple, but I'm just running into a few issues that I can't seem to break through, so any brain-orientating-help is appreciated as I bone-head through this.
Equipment I have:
MiniDSP Umik-1 microphone (USB)
Dayton UMM-6 microphone (USB)
Dayton DATS V3
USB DAC, Amplifier, etc
Software:
REW
Xim
VituixCAD2
I went back over some notes and some information from others but I ran into some questions and confused myself.
My primary goal:
Measure drivers installed in a baffle I already made and produce FRD and ZMA files from my measurements so that I can work on passive crossovers in software (Xsim & VitruixCad2).
So, that requires that I make pretty good measurements in the first place. And this is my primary hang up.
I know I can make impedance measurements with the DATS V3 in the baffle and produce the ZMA file directly from that without anything extra (I think?). I'm not sure if there's anything else I need to know about or do to the impedance sweep from DATS before using it as a ZMA for crossover building.
Some of my notes:
1 - Driver in the baffle. Microphone centered up on the driver at a distance of 0.5 meters. Reference 1khz sinewave and use voltmeter to get 2V AC at the speaker terminal. Do this with >1m in all directions free of boundaries, etc. Take measurement.
2 - Next driver, same as above.
3 - Now, interferometry. I read about this but couldn't find better (simple) details for me to follow. But basically wire the two drivers in parallel and measure this as well. But, I'm not sure how to set that up properly and what to do with the result. For example, do I wire the tweeter and woofer together in parallel and then place the microphone centered between the two of them at 0.5m distance and do the same 1khz sine wave at 2V AC again? Or is that incorrect? I'm lost on what to do with this 3rd measurement. Measurement 1 is tweeter. Measurement 2 is woofer. Measurement 3 is both of them in parallel? I'm not sure what to do with that part (3). I read something about calibrating the microphone with a separate SPL meter and I'm lost on that part.
4 - In addition to the above, what about if I have 2 woofers and 1 tweeter? How does that change this for measuring the woofer(s) and for the (3) measurement with the tweeter+woofers in parallel?
5 - Any considerations on things if the baffle has a port like bass reflex when producing the response (separate measurements and sum the port plus the woofer)?
6 - Scale this to 1.0m distance and 2.83V. I think I can do this in REW easily (any notes?).
7 - If I use DATS to measure each driver in the baffle, do I have to do anything specific to the results to normalize or scale it, or is it good to go as a ZMA from just that measurement? If the impedance sweep involves a bass reflex cabinet, does this change anything with the ZMA file and do I need to do anything to it?
Thanks for reading and any comments to help out! I appreciate all of your time.
++++++++++++++++++++++++++++++++++++++
Here's the baffles and drivers I made that I'm trying to measure and build the crossover for:
Current 1 meter response with active crossover (DSP). I want to do similar with passive crossover, so I'm trying to re-learn how to properly measure and produce FRD and ZMA so I can start working on that.
Very best,
I recently built a baffle and put some drivers in it and was happy with the physical result. I had boomarked some information on how to best measure them on the baffle and produce FRD and ZMA and all the links are either dead or the software is dead and I can't follow a tutorial, etc, because of the project being dead. So I'm back to square one. I realize this is probably simple, but I'm just running into a few issues that I can't seem to break through, so any brain-orientating-help is appreciated as I bone-head through this.
Equipment I have:
MiniDSP Umik-1 microphone (USB)
Dayton UMM-6 microphone (USB)
Dayton DATS V3
USB DAC, Amplifier, etc
Software:
REW
Xim
VituixCAD2
I went back over some notes and some information from others but I ran into some questions and confused myself.
My primary goal:
Measure drivers installed in a baffle I already made and produce FRD and ZMA files from my measurements so that I can work on passive crossovers in software (Xsim & VitruixCad2).
So, that requires that I make pretty good measurements in the first place. And this is my primary hang up.
I know I can make impedance measurements with the DATS V3 in the baffle and produce the ZMA file directly from that without anything extra (I think?). I'm not sure if there's anything else I need to know about or do to the impedance sweep from DATS before using it as a ZMA for crossover building.
Some of my notes:
1 - Driver in the baffle. Microphone centered up on the driver at a distance of 0.5 meters. Reference 1khz sinewave and use voltmeter to get 2V AC at the speaker terminal. Do this with >1m in all directions free of boundaries, etc. Take measurement.
2 - Next driver, same as above.
3 - Now, interferometry. I read about this but couldn't find better (simple) details for me to follow. But basically wire the two drivers in parallel and measure this as well. But, I'm not sure how to set that up properly and what to do with the result. For example, do I wire the tweeter and woofer together in parallel and then place the microphone centered between the two of them at 0.5m distance and do the same 1khz sine wave at 2V AC again? Or is that incorrect? I'm lost on what to do with this 3rd measurement. Measurement 1 is tweeter. Measurement 2 is woofer. Measurement 3 is both of them in parallel? I'm not sure what to do with that part (3). I read something about calibrating the microphone with a separate SPL meter and I'm lost on that part.
4 - In addition to the above, what about if I have 2 woofers and 1 tweeter? How does that change this for measuring the woofer(s) and for the (3) measurement with the tweeter+woofers in parallel?
5 - Any considerations on things if the baffle has a port like bass reflex when producing the response (separate measurements and sum the port plus the woofer)?
6 - Scale this to 1.0m distance and 2.83V. I think I can do this in REW easily (any notes?).
7 - If I use DATS to measure each driver in the baffle, do I have to do anything specific to the results to normalize or scale it, or is it good to go as a ZMA from just that measurement? If the impedance sweep involves a bass reflex cabinet, does this change anything with the ZMA file and do I need to do anything to it?
Thanks for reading and any comments to help out! I appreciate all of your time.
++++++++++++++++++++++++++++++++++++++
Here's the baffles and drivers I made that I'm trying to measure and build the crossover for:
Current 1 meter response with active crossover (DSP). I want to do similar with passive crossover, so I'm trying to re-learn how to properly measure and produce FRD and ZMA so I can start working on that.
Very best,
The interferometry is one of various workarounds you can use to make up for not producing timed measurements. The best you'd get from this will be simultaneously measuring the two similarly to how they were measured individually. That said, if you measured them each from different locations in the first place, you'll have to work out how those differ from the place you chose for the third measurement, and add that to the delay you're chasing in the simulator.
ZMA shouldn't need scaling.
Accuracy is not a bad thing but your response measurement will be affected by the uniqueness of the enclosure in ways that cannot be expected to be measured from a single location.
ZMA shouldn't need scaling.
Accuracy is not a bad thing but your response measurement will be affected by the uniqueness of the enclosure in ways that cannot be expected to be measured from a single location.
Thanks,
I'm not sure I follow in terms of how I would measure two at the same time, from multiple locations, etc. I think I was reading something that didn't go into detail on this part, so it's foreign to me. Hence all the questions.
Good to know the ZMA won't need scaling. One less thing!
Yes, the idea is that I'd like to measure drivers in the enclosure I build and make crossovers from that data instead of using published T&S of the drivers, so I can get it closer to what it actually will be.
Any thoughts/notes on the questions (notes, but they are questions really) 1 through 7 above?
Thanks!
Very best,
I edited my post to add..
What I'm trying to do here is highlight the purpose. How you do it is up to you as long as you measure the right things.
1. as long as you keep the volume the same between the two, you can move forward. Level tends to be about signal to noise and distortion.. also repeatability. I won't presume to guess in your situation.
6. Why scale, does it serve a purpose for your crossover? Absolute levels won't change the filters. You might see a difference in level by scaling, but you won't change the box diffraction response. You get a better view of that if you move back from the box when measuring.
This is limited by your measurement environment. Closer means lower frequency of reflection free measurement, and it's a compromise.
Let me put that another way, theoretically it's possible to place the mic at one place, take 3 measurements - FR, woofer and both, and use that for working out delay. Then to use two completely separate measurements for response purposes.
What I'm trying to do here is highlight the purpose. How you do it is up to you as long as you measure the right things.
1. as long as you keep the volume the same between the two, you can move forward. Level tends to be about signal to noise and distortion.. also repeatability. I won't presume to guess in your situation.
6. Why scale, does it serve a purpose for your crossover? Absolute levels won't change the filters. You might see a difference in level by scaling, but you won't change the box diffraction response. You get a better view of that if you move back from the box when measuring.
This is limited by your measurement environment. Closer means lower frequency of reflection free measurement, and it's a compromise.
Thanks,
I follow some of that I think. What loses me is the idea of measuring the tweeter, woofer and then both, but both how? This was the comment about measuring the tweeter and woofer in parallel and what that means or if I'm misinterpreting what I read in that. And how I would use that 3rd measurement to work out delay, or if that is just done manually in the crossover software.
I'm definitely trying to not guess, I want to be able to do repeatable methods for several speakers and future projects, hence all the questions.
I'm not sure why on scaling; it was part of what I was reading and spawned questions. It was described as doing 0.5m distance and using 1khz sine wave at 2 volts AC and then scaling it to 1m and 2.83V for the FRD. I assume for simplicity and standard as almost everything is based on 1 meter and 2.83v, etc.
This post spawned most of these questions:
I can't find the source referred to in the post to review/read. So I'm just asking lots of questions.
@xrk971
I'd love to find this process mentioned here by Laub but can't find it. I found his site, but can't find the discussion on this. I've reached out to this poster on DM, but figured I'd ask general forum too in case anyone can help.
Thanks again for your time.
Very best,
Last edited:
The conditions I gave you (regarding S/N and distortion) are so that you might find your own number to represent level. Is there anything I missed?
This can be done with crossover components in place, if you feel that would be better. It should make no difference to the end result as long as the process can be followed.
So you can understand the process, open your simulator to a crossover you've already done, any one will do. Vary the delay for one driver and see how the resulting phase varies the response. This interferometry will just be a matter of working backwards from response to find delay.
Thanks, I'll try this out.
Though before I worry so much about the offset, I'd like to establish and confirm the best way to approaching measuring drivers in their enclosures with the equipment I have in the first place. I'm trying to understand more about the post I linked that refer to this with same equipment and figure out a few things, but so much is missing. 0.5m at 1khz at 2v AC, got that part. But, 20hz to 20khz each driver sweep? I can measure each driver individually. Got that part. But the "in parallel" part I don't understand (ie, is it both drivers at the same time, wired in parallel to each other? Or something else? Do the same 20hz to 20khz sweep like that?). If measuring a woofer that is in an enclosure with a port/vent like bass reflex, do I have to measure the port response and port response and sum them to get the final response to work with? Things like that, which are left out that have me questioning everything so I can understand.
Very best,
Last edited:
Yes.
This is usually below the crossover range so some manage without doing anything here. It's an issue since it's hard to measure that low in a room. If you were outside, one simple measurement would suffice for response.
Thank you, that is very helpful. Slowly getting the puzzle worked out.
I can do ground plane measurements outside for low frequency stuff on good weather days. I already do this for subs anyways.
I was just seeing how to do near field port + woofer measurements, and combining the response in REW to produce the FRD for crossover work but it needs to be corrected for diffraction. So I have to learn how to do that. But yes, its generally lower than I need for woofer + tweeter. That said, it will come up, especially with full range driver builds I do and some woofers that I have in ported cabinets tuned high.
Very best,
Hrm,
Reviewing how to measure things correctly again: the comments before with the 0.5m and 2V at 1khz... there's no mention of gating the impulse response and the woofer isn't done with near field and scaled to the far field (0.5m) measurement and no mention of any diffraction or area for correction, etc. And finally, I'm not seeing how this can be done with multiple woofers and wiring all of them parallel and getting a wildly low impedance?
++++++++++++++++++
Things I'd like to learn to properly measure, drivers in their enclosures for making FRD/ZMA:
1 - Tweeter + Woofer (TM)
2 - Full range or wideband driver in sealed or bass reflex (specifically bass reflex for sure)
3 - Multi-woofer (MTM or even WMTMW, etc)
Very best,
Reviewing how to measure things correctly again: the comments before with the 0.5m and 2V at 1khz... there's no mention of gating the impulse response and the woofer isn't done with near field and scaled to the far field (0.5m) measurement and no mention of any diffraction or area for correction, etc. And finally, I'm not seeing how this can be done with multiple woofers and wiring all of them parallel and getting a wildly low impedance?
++++++++++++++++++
Things I'd like to learn to properly measure, drivers in their enclosures for making FRD/ZMA:
1 - Tweeter + Woofer (TM)
2 - Full range or wideband driver in sealed or bass reflex (specifically bass reflex for sure)
3 - Multi-woofer (MTM or even WMTMW, etc)
Very best,
Last edited:
Best not to run woofers in parallel if you don't have the amp for it.
Are you not planning to take polars, because it changes everything? You can just run the woofers together and measure. As simple as that.
This is not essential for determining delay, I wouldn't worry about it as long as you measure at a good level and keep it the same between measurements.
There are two types of calibration and neither is needed for determining delay. One has to do with level and one has to do with response smoothness. What matters is the relative measurements.
Sorry I'm not sure what you mean by "take polars?"
By running the woofers together, do you mean in series and not parallel? I would imagine distance will matter more when measuring multiple woofers so that one is not crazy closer (higher SPL) than the other?
Very best,
I'm interested in this thread because there is a lot I have to learn about measurements too. But there are a few things, (I think) I understand that I may be able to help clear up.
"Taking polars" is making a bunch of measurements at off axis intervals to create a polar map. This is important if you're interested in directivity.
You don't have to make measurements at any specific level(i.e. 2.83V). Unless you're presenting the data as such, or need to compare it another speaker at that level. What is important is that all the measurements of the drivers for this individual speaker are taken at the same level.
In the absence of a timing reference, measuring the combined drivers can be used to determine physical offset. My understanding of this is to measure the individual drivers and then the combined drivers(all drivers in parallel). Then load all three responses in your simulation software. Next you need to vary the physical offset of one driver(most likely the tweeter) till the combination of individual responses matches the measured combined response.
When I first read about measuring all the drivers in parallel, I too was worried about the resulting low impedance. But I don't believe this is an issue in practice, as the sweep is only a few seconds.
"Taking polars" is making a bunch of measurements at off axis intervals to create a polar map. This is important if you're interested in directivity.
You don't have to make measurements at any specific level(i.e. 2.83V). Unless you're presenting the data as such, or need to compare it another speaker at that level. What is important is that all the measurements of the drivers for this individual speaker are taken at the same level.
In the absence of a timing reference, measuring the combined drivers can be used to determine physical offset. My understanding of this is to measure the individual drivers and then the combined drivers(all drivers in parallel). Then load all three responses in your simulation software. Next you need to vary the physical offset of one driver(most likely the tweeter) till the combination of individual responses matches the measured combined response.
When I first read about measuring all the drivers in parallel, I too was worried about the resulting low impedance. But I don't believe this is an issue in practice, as the sweep is only a few seconds.
Thank you so much, this clarifies many of my questions above nicely.
Ok, so polar mapping off axis. Great, not interested in doing that just yet. I've done some of that before but that was on a completed speaker. Thanks for clarifying that though, that term threw me off in this context.
Hrm, so if I use a random level and just keep it the same level for all far field measurements, then it shouldn't matter then. I guess I get caught up thinking too much about different sensitivities of drivers and incorporating near field response that is normalized and scaled for things that descend in frequency that requires near field to better measure. So if I keep the levels the same for all far field and levels the same for all near field, it should result in what I need without worry about actual sensitivity. I can test that specifically later of course.
The physical offset was definitely throwing me off, the idea of wiring all drivers in parallel with a super tiny impedance sounded like a disaster potentially. So wondered what I would do with low impedance drivers doing that. 4ohm is fine, until you parallel wire 2~3 of them and get some tiny impedance value. That had me wondering what was going on. But if it's low level and a short sweep and doesn't matter, then that's good.
Thanks!
Very best,
See this post and the following messages
Thanks, that is helpful regarding the multi-woofer part!
Very best,
I don't think you would have to necessarily keep the same level for near field measurements. I see what you're saying about scaling, so they line up when you merge them. But I'm not sure if that will work in practice. The levels can be adjusted when you merge them. And from what I've seen on merging, it takes a bit of fudging with to get them to line up.
I would think so too but it seems to be a common approach to this for bass. For many things I wouldn't worry under 400hz so much. But above that it would be nice to not have to do all that to build a good FRD of a driver in an enclosure to work on for crossover.
Very best,
I just remembered I bookmarked a guide to measuring speakers. I need to go back and read through it again one of these days. Post #2 goes into detail on merging responses.
https://www.audiosciencereview.com/...ents-spinoramas-with-rew-and-vituixcad.21860/
https://www.audiosciencereview.com/...ents-spinoramas-with-rew-and-vituixcad.21860/
Thanks, I actually have the course he references from UDEMY by AudioJudgement author. It's a good course. It's just a lot. I was looking for something slightly easier when I noticed XRK's post quoted above and wanted to explore that more.
Very best,