• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

DIY speakers with Waveguide

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
Does someone have a guide how to this with REW and UMIK-1?

There is a 2-part thread on the forum about getting started with REW, if that's your concern.
For a more in-depth and broader treatment, check out Joe D'Appolito's book on loudspeaker measurement.
It's a bit dated, but the content is good. I got a copy from the local library and ended up buying a used version.
 

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
One needs different kind of measurements for loudspeaker design or room mode analysis.
UMIK or any usb-mic is not recommended for speaker design/measurements, because of variations in delay. There are sideways to handle that, with some extra work.

VituixCAD has recommended procedures at homepage https://kimmosaunisto.net/
 
Last edited:

Biblob

Addicted to Fun and Learning
Joined
Sep 13, 2018
Messages
633
Likes
600
Hmm, I didn't know this, thanks for letting me know.
I do have an Asus U7, would the ADC be good enough for this purpose? If so, I'll buy a EMM-6 as well.
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
UMIK or any usb-mic is not recommended for speaker design/measurements, because of variations in delay. There are sideways to handle that, with some extra work.

Can you elaborate on the nature of the problem(s) using a UMIK-1 with REW for measurements, as well as the workarounds "with some extra work"?

EDIT: your link cautions
Note! Single channel measurement systems such as USB microphones with latency variations by default are not recommended for speaker engineering due to timing and phase variations and normalizations. REW should not be used with single channel connection or mode for far field measurements because timing is normalized by the program. Single channel connection and mode is acceptable for near field measurements only.

Let's just say I'm skeptical, but open to guidance.
Perhaps @JohnPM can comment?
 
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
Thank you. It's my first ever DIY too, so I feel flattered. Have much to learn still.
Still don't know how I'm gonna measure accuately enough to design the XO. I want to cross at about 1,6 khz though.

Does someone have a guide how to do this with REW and UMIK-1?

So, the website audiojudgement has this course:
https://www.udemy.com/course/acoust...asurements/?referralCode=0682837C54A3EBBADCAC
I've never taken it but it's 10 dollars and seems to cover what you need pretty comprehensively. I'd email the guy and ask what equipment he uses.

If you're doing a passive crossover you'll need to measure the impedance of each driver as well. By far the easiest way to do this is with a 'DATS' sold by parts express, which is $110. REW can also measure impedance if you're comfortable making your own weird cable for the task:
https://www.roomeqwizard.com/help/help_en-GB/html/impedancemeasurement.html
 
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
Can you elaborate on the nature of the problem(s) using a UMIK-1 with REW for measurements, as well as the workarounds "with some extra work"?

EDIT: your link cautions


Let's just say I'm skeptical, but open to guidance.
Perhaps @JohnPM can comment?
The workaround I believe Juhazi is referring to is finding the Z offset of the drivers being tested by means of a method such as this:
http://www.box.net/shared/ouxjjsx0m8bs00cil5iq
Which I have tried to do countless times and it's never provided satisfactory results. Meanwhile the semi dual channel measurement with ARTA gives perfect results every time.
 

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
Sorry, I cant help personally with measurement technique. I have never even tried to make such , to design/simulate a passive xo system - I am way too dumb and lazy to learn that!.

I use REW and single channel UMIK-1 to set up my active multiway speakers, by taking measurements, then adjusting parameters "manually" without any calculations or simulation, then measuring again etc. (ad nauseam, the wrong way to do it).

I have been around diyaudio.com forums for 12 years now, and I find it the absolutely best forum for loudspeaker/audio diy hobby. There are subforums for loudspeaker desing, measurement and simulation, box construction etc.

https://www.diyaudio.com/forums/software-tools/
 
Last edited:

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
UMIK or any usb-mic is not recommended for speaker design/measurements, because of variations in delay.

Again, that is a fairly strong blanket statement. Please explain.
I'm not saying you're wrong, but I am trying to understand.
If you are indeed referring to measurements of acoustic center, I don't see how a REW + UMIK-1 system could not provide a reliable measurement with its pilot-tone-bracketed probe signal. Again, just trying to understand.
 

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_Measurement_Preparations_REW.pdf

"Note! Single channel measurement systems such as USB microphones with latency variations by default are not recommended for speaker engineering due to timing and phase variations and normalizations. REW should not be used with single channel connection or mode for far field measurements because timing is normalized by the program. Single channel connection and mode is acceptable for near field measurements only. "

You need that timing accuracy to get drivers' phases to match the way you want. Each measurement's impulse response shows amplitude variations and phase variations. A crossover slope will change both amplitude response and phase angle of signal. If simulation uses wrong values and is optimized per that, the final acoustic outcome (summation) will be wrong.

https://sound-au.com/ptd.htm
https://www.rane.com/note160.html
 
  • Like
Reactions: 617

TimW

Major Contributor
Joined
Aug 15, 2018
Messages
1,065
Likes
1,406
Location
Seattle, WA
So far I've just been using REW with a UMIK-1 and when designing a passive crossover I use that method described by Jeff Bagby in the article @617 just shared. I understand there are probably better ways to do this but I wasn't aware the method is not advisable. Looking at DIYAudio there seems to be a lot of software that can do the job. I have a usb recording interface and am willing to buy a calibrated microphone and learn a new software. Can anyone recommend a software that is cheap and easy to use? I'm looking at the ARTA website and it seems very technical (not easy).
 
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
Let me explain my understanding of the issue, including some background info for people who may be curious why we're focusing so much on this issue.

To accurately measure phase relationships between drivers, the software needs to know very precisely when the sound was sent to the speaker and when it came back. By subtracting these two times and multiplying by the speed of sound it can calculate the distance to the speaker. If the measurement signal from the woofer takes longer to come back than the tweeter, it can be assumed that acoustic 'center' of the woofer is further away, even if they are on the same flat baffle.

This difference is probably less than an inch or two. Understanding that sound travels at around 13,000 inches per second, we can see that we need extreme precision in measuring these two times (sent and return.) This is especially critical at high frequencies, where moving a driver back and forth and inch can result in the drivers being in phase or close to out of phase. Not knowing their precise spatial relationship makes accurate crossover design impossible.

If you're using two separate USB devices to send and then measure the signal, it is very easy for the latency of the usb devices to change both the sending and receiving times. A really good audio interface may have a latency of 3/1000 of a second. Let's say that's the total amount of error we're dealing with between the usb mic and the sound card output. Sound travels around 40" in that time. It's clear that two separate USB device with two different and probably variable latencies won't work, they need to be synchronized in the time domain.

Arta uses both channels (2x input and 2x output) of your recording interface to sync up the sending and returning.

I don't know if REW has some ingenious way around this but it seems like an intrinsic problem with USB mics. One way of testing would be to make the same measurement a few times and see that the impulse falls at the same point in time every time, then to move the speaker away a few inches and making sure the impulse arrives later in time repeatedly. Note that at least with ARTA, if you don't use dual channel measurement, the software will simply put the impulse at 0 seconds, which of course tells you nothing.

If REW can somehow deal with this issue that is awesome but I remain skeptical. If you can point to some part of the REW documentation that would be helpful.
 

Biblob

Addicted to Fun and Learning
Joined
Sep 13, 2018
Messages
633
Likes
600
Very helpfull, thank you. I will reconsider using the Umik-1 now.
 
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
Okay I think I understand the REW feature being discussed:
http://www.roomeqwizard.com/help/help_en-GB/html/makingmeasurements.html#acousticref

Basically it requires another speaker other than the one being tested. Very clever. My understanding is that it sends a test signal to the 'reference speaker' using one channel of your output device, and then sends another test signal (sweep) to the device being tested. Then, without moving anything, you measure the next device while using the same reference speaker. By comparing the difference in impulses between the (unmoving) reference speaker and the two devices being tested, it can calculate the time differences between the two devices under test.

That's really cool so I guess I stand corrected. I won't be using it any time soon, I can't stand REW, and my preferred speaker simulation software works with ARTA seamlessly, but yeah this seems like a viable measurement platform for speakers.
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_Measurement_Preparations_REW.pdf

"Note! Single channel measurement systems such as USB microphones with latency variations by default are not recommended for speaker engineering due to timing and phase variations and normalizations. REW should not be used with single channel connection or mode for far field measurements because timing is normalized by the program. Single channel connection and mode is acceptable for near field measurements only. "

This is the same quote that I was asking about earlier in this very thread.
I could just be dense (likely), but that does not help.

To me the key phrase there is "single channel measurement system".
We all must be talking past each other, as I am referring to using REW+UMIK-1 (or any mic) as a two-channel measurement system.
The second channel serves as the timing reference. See pic attached.

You need that timing accuracy to get drivers' phases to match the way you want. Each measurement's impulse response shows amplitude variations and phase variations. A crossover slope will change both amplitude response and phase angle of signal. If simulation uses wrong values and is optimized per that, the final acoustic outcome (summation) will be wrong.

I understand why accurate timing/phase measurements can be important.
I am asking why, or where is some evidence, that using REW with a UMIK-1 is incapable of meeting such requirements.

For instance, this post from miniDSP implies no such problem making timing measurements.

It's clear that two separate USB device with two different and probably variable latencies won't work, they need to be synchronized in the time domain.
If you're using two separate USB devices to send and then measure the signal, it is very easy for the latency of the usb devices to change both the sending and receiving times.

Offline analysis?
What about the case of making a WAV file of REW's test signal(s), using that test signal to probe some speakers by recording their responses independent from REW, then using REW offline at a later date on those recorded responses?

If one has a timing reference embedded in the probe signal and encoded on the measurement signal, I don't see that even this case is a problem with its arbitrary delay between probe and measurement.
 

Attachments

  • rew_timing_signals.png
    rew_timing_signals.png
    137.8 KB · Views: 149
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
This is the same quote that I was asking about earlier in this very thread.
I could just be dense (likely), but that does not help.

To me the key phrase there is "single channel measurement system".
We all must be talking past each other, as I am referring to using REW+UMIK-1 (or any mic) as a two-channel measurement system.
The second channel serves as the timing reference. See pic attached.



I understand why accurate timing/phase measurements can be important.
I am asking why, or where is some evidence, that using REW with a UMIK-1 is incapable of meeting such requirements.

For instance, this post from miniDSP implies no such problem making timing measurements.




Offline analysis?
What about the case of making a WAV file of REW's test signal(s), using that test signal to probe some speakers by recording their responses independent from REW, then using REW offline at a later date on those recorded responses?

If one has a timing reference embedded in the probe signal and encoded on the measurement signal, I don't see that even this case is a problem with its arbitrary delay between probe and measurement.

Please see my next post. I understand the REW timing reference now. I thought it was sending the reference signal to the DUT rather than another speaker entirely.
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
Please see my next post. I understand the REW timing reference now. I thought it was sending the reference signal to the DUT rather than another speaker entirely.

Ah, we were talking past each other! Sorry for that @617.
 
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
Ah, we were talking past each other! Sorry for that @617.

No problem. I personally see some issues with the acoustic reference - you'd definitely want the reference roughly behind or in front of the speaker you're measuring, in other words aligned with the mic-device axis. You've used this feature with good results I take it? I need to do some measurements soon, maybe I'll try it.
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
You've used this feature with good results I take it?

In fact, I have not. But, I am about to venture down that road, thus my interest.
On paper, the method seems fine to me. But, I'm a noob at acoustics and looking to learn.

I personally see some issues with the acoustic reference - you'd definitely want the reference roughly behind or in front of the speaker you're measuring, in other words aligned with the mic-device axis.

I'm afraid I don't understand: if the reference is relative, why be concerned with the plane of position?

EDIT: my thinking is that as long as the two drivers under test (say, tweeter & woofer) are in the same plane, then the reference position can be arbitrary (within reason!).
 
Last edited:

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
https://www.diyaudio.com/forums/multi-way/76977-arta-60.html#post5492173
" I've tested latency variations with Umik-1 and external sound card (but not with ARTA). Conclusion was that severe timing/phase differences exist between different measurements. Timing may stay quite stable few measurements but then delay might jump ~0.5...>1 ms back and forth. Rebooting or starting another program between measurements was not needed to generate delay jumps, and latency adjustment of sound card did not help.
So, it is not valid gear for advanced speaker measurement needed to simulate multi-way to off-axis, power, DI responses etc. because that requires very stable timing i.e. semi-dual (or dual) channel connection and measurement mode selected in ARTA (or REW). "


https://www.diyaudio.com/forums/multi-way/76977-arta-60.html#post5492272
"Single channel gear such as simple USB mics are very common today. They have spread from room acoustics measurements to speaker design which is unfortunate because more and more new users end up to bad, inaccurate/random and very slow design methods. This is not my fight, so I simply recommend to ignore messages and authors supporting or advertising such gear from now on. "

REW has given possibility for 2-ch measurements only for some time, after kimmosto got John to understand this. REW is primarily a room measurement/analysis program, impedance meas and 2-ch are extra add-on features .
 
OP
617

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
In fact, I have not. But, I am about to venture down that road, thus my interest.
On paper, the method seems fine to me. But, I'm a noob at acoustics and looking to learn.

I'm afraid I don't understand: if the reference is relative, why be concerned with the plane of position?

EDIT: my thinking is that as long as the two drivers under test (say, tweeter & woofer) are in the same plane, then the reference position can be arbitrary (within reason!).

See if this diagram makes sense:

error trig.png

Not sure if I have it right in my mind. Again this stuff is easy to test. If I'm wrong then moving the reference speaker shouldn't impact the relative delays of the two drivers.
 
Top Bottom