I was doing some more thinking about this PhotoLink device and how it compares to my Steinberg UR22 MKii + Arta jig loopback setup. Since all you have to do is connect the PhotoLink box to the speaker wires and point it at the microphone, the new Omnimic 40K will probably be almost as fast to set up and start taking measurements as the old OmniMic was. In addition, the PhotoLink eliminates the possiblity of overdriving and burning the inputs of the microphone soundcard. When setting up my UR22 to take measurements, I am always very careful to turn input levels way down. Then I do multiple test runs, slowly bringing up the levels until everything looks good. I need to follow this procedure during every setup to prevent accidental UR22 input overload. So the PhotoLink idea was probably, as you suggested, for quick "plug and play" setups. But it also helps to prevent loopback overdrive burnout situations.
Add a couple zener diodes to your jig to clamp down the input voltage if your prone to doing silly things like cranking out 50V with input gain set to max, they cost peanuts. I'd be more worried for your speakers honestly.
I use the signal generator to set levels, it takes seconds, and on my setup the input gain stays at the same 9 o'clock position 95% of the time.
If all you need is a timing reference, which is all the optical thing is going to provide, just use semi-dual loopback as I showed in the photo above. It's all line level signals there, low risk but input gain can still be something like +50dB with input gain knob at max ( max gain is +40dB for UR22mkii. )
@dcibel said
Is this really too complicated for people?
The physical connections is a piece of cake.
Changing processes is the harder thing to do.
If one cannot see an advantage or benefit from changing their process, they’ll be inclined to continued the same tried and trusted process. I've seen measurements made by companies using their own software or legacy systems like Linear X LMS (DOS era) and WinMLS 2000 Pro (Windows 2000 era0 and I can see why they are reluctant to change:
If it ain’t broke, Why fix it?
I enjoy using different parts as much as the next person eg. drivers, crossovers.
But beyond a 6 sided rectangular prism of a loudspeaker cabinet, I'm interested in seeing the effect of changes to dispersion and diffraction from the baffle size and shape, the placement of drivers on the baffle in 3 dimensions, listening to the effects of filters without having to build them (passive or active), and how a speaker interacts with a room in 4D.
What does (semi) dual channel measurements offer that Omnimic40K doesn't?
What does Omnimic40K offer that dual channel measurements doesn't?
What else does (semi) dual channel measurements offer that Omnimic40K doesn't?
As far as a timing reference goes, not much. But, if you compare what a USB audio Interface offers that a USB mic doesn't, it's a heck of a lot. A full dual channel setup offers a lot more as well.
What does Omnimic40K offer that dual channel measurements doesn't?
Absolute SPL without needing a sound level calibrator. But for most general use, absolute SPL isn't that important, I only got an SPL calibrator once I got deep into distortion analysis.
I heard the Omnimic 40k box also has a little light that lights up when the signal level is 2.83V +/- some percent. Maybe a nice convenience feature for some.
I still haven't seen what has been done to update the proprietary software for this new mic. Bill said it himself, that it's hard to compete with a viable for-profit product when your main competition is well executed freeware (REW). At the very least, for the timing reference to make any sense, it needs some spatial measurement tool to easily create subsequent measurements with automatic file naming. REW and ARTA have been able to do it for years. The software needs the ability to adjust windowing and smoothing settings without remeasuring, because that's just silly. Also it should allow export of the raw impulse response for processing in other software.
Honestly I'd be a lot more excited if Dayton would offer a decent measurement condenser mic with XLR connection, price at $150-$180 range. They could skip the software development and RND for some odd timing solution for all I care, just sell a good mic at a reasonable cost.
@dcibel said:
a USB audio Interface offers that a USB mic
doesn't, it's a heck of a lot. A full dual channel
setup offers a lot more as well.
What do you feel are the top 3 most useful things that a USB audio interface and full dual timing setup offers?
Honestly I'd be a lot more excited if Dayton would offer a decent measurement condenser mic with XLR connection, price at $150-$180 range. They could skip the software development and RND for some odd timing solution for all I care, just sell a good mic at a reasonable cost.
There was a proposal from Bill to the product managers at Parts Express/Dayton about a standalone microphone. It is good up to 50KHz, extremely rugged and suitable for everyday use/abuse/accidental drops.
Unfortunately it requires skills quite different to a typical electronics fab and the labor would probably drive up the price to the point it’s not commercially viable against their current line of XLR mics.
@dcibel said:
a USB audio Interface offers that a USB mic
doesn't, it's a heck of a lot. A full dual channel
setup offers a lot more as well.
What do you feel are the top 3 most useful things that a USB audio interface and full dual timing setup offers?
That will vary depending on the person and their needs. There are many reasons one may consider a USB audio Interface over a canned USB mic solution. The biggest for me when moving away from Omnimic was SNR. The old Omnimic has low end electronics internally which results in low SNR, even the cheapest USB audio Interfaces outperform it by a country mile. A USB audio Interface can be used as a PC DAC for any purpose. It allows one to record any audio input, swap mics, use for a simple PA setup, PC audio for zoom calls, podcasts, and if you're a musician of any sort who records what you play you may already have one. It allows an easy upgrade path to a full dual channel setup for only a few dollars spent, and measuring impedance as well, so a USB interface, a few resistors and some scrap wire can cover the functions of acoustic and electrical measurement, so you can now compare to the costs of interface+standard mic to that of a USB mic and DATS combined. A USB audio Interface also doesn't require proprietary software, you can choose between several software solutions, so the product longevity is a lot longer in my opinion and is simply more versatile.
A full dual channel system simply provides a feedback loop that is at the driver end of the signal chain rather than at the interface (semi-dual). Any response changes along the signal path due to the amp used, cabling, or filters is removed from the measurement allowing increased accuracy ensuring the measured result is purely the speaker with no external influence. Think of it as an auto-cal with every measurement, as the measurement is a differential between mic input and electrical feedback. It allows for insertion of a protection cap without messing up the driver response, avoiding extra post-processing steps, and opens up opportunities for more advanced signal analysis. I've used mine to provide filter transfer functions of both passive and active DSP filters for design validation.
Anyway, enough rambling on, I imagine everyone here has already made their choice on what they want to use. For anyone who hasn't, the main point to drive home is that if you have a 2 channel USB audio Interface, measuring with accurate timing to get started requires no extra diy jig whatsoever, just a patch cable.
In case anyone is interested, the new OmniMic40k software package is available for download from the Dayton Audio website. It is version 6.052 and works with both the old and new OmniMic hardware. The help file has detailed info on general setup, new feature listing, photolink setup, setting impulse response window edges, harmonic distortion percentages, a new compression testing feature, and info on how the 40k calibration was performed. Check it out:
@4thtry said:
In case anyone is interested, the new OmniMic40k software package is available for download from the Dayton Audio website. It is version 6.052 and works with both the old and new OmniMic hardware. The help file has detailed info on general setup, new feature listing, photolink setup, setting impulse response window edges, harmonic distortion percentages, a new compression testing feature, and info on how the 40k calibration was performed. Check it out:
Man, I forgot how annoying it was to constantly bounce back and forth between sweep playback and measurement recording, generally needing at least 3 sweeps to play before you get one good one. Anyway, there's no easy spatial measurement function here for this reason, playback and record are decoupled so repeated measurements for "spinorama" data will be quite tedious with the Omnimic 40k.
I wonder how well this photo link thing will work when it's not pointed at the mic, as will be the case for full 360 deg measurement datasets? It looks like any remote control IR blaster so you may have to set up a mirror lol.
Much of my complaints about Omnimic still appear to be true with the new 40k version. Solen lists this thing at CAD$792 (yikes), for comparison I could buy a Motu M2, Line Audio Omni1 and still have a bunch of money left over in the bank, just sayin..
There are a ton of peerless coming back, not just a few. Exciting times ahead as I suspect we will see plenty of new designs. The HDS 6.5 with phase plug is under $60 and you have to spend a lot more on other 6.5s with similar performance.
IIRC, the only Peerless I know with phase plugs were the Exclusives and M series. The M and Ex both had the same HDS type frames though. The M hasn't been available for quite some time now.
@4thtry said:
In case anyone is interested, the new OmniMic40k software package is available for download from the Dayton Audio website. It is version 6.052 and works with both the old and new OmniMic hardware. The help file has detailed info on general setup, new feature listing, photolink setup, setting impulse response window edges, harmonic distortion percentages, a new compression testing feature, and info on how the 40k calibration was performed. Check it out:
Man, I forgot how annoying it was to constantly bounce back and forth between sweep playback and measurement recording, generally needing at least 3 sweeps to play before you get one good one. Anyway, there's no easy spatial measurement function here for this reason, playback and record are decoupled so repeated measurements for "spinorama" data will be quite tedious with the Omnimic 40k.
I wonder how well this photo link thing will work when it's not pointed at the mic, as will be the case for full 360 deg measurement datasets? It looks like any remote control IR blaster so you may have to set up a mirror lol.
Much of my complaints about Omnimic still appear to be true with the new 40k version. Solen lists this thing at CAD$792 (yikes), for comparison I could buy a Motu M2, Line Audio Omni1 and still have a bunch of money left over in the bank, just sayin..
Thanks for your comments, dcibel. You make some very good points.
Playback/record issue: My thinking is that the playback/record issue shouldn't be too bad with the new mic. I know that when I had previously recorded polars with my old OmniMic that I would often need to cycle through multiple "whoop-whoop" cycles before hitting pause to record, because the phase curve kept drifting up and down in a repeating cycle with each whoop. But, according to the help file, the new mic's phase capture should be rock solid with each whoop. So I should be able to hit pause after a couple whoops, then hit save, then rotate the speaker, etc. OmniMic has an auto naming feature, so this should actually go a little bit faster than my current Soundeasy V20 capture and save process. The downside, however, is that the new OmniMic software does not appear to have the ability to save each measurement as a raw impulse file. So I would have to save each measurement as a blended FRD with my gating decisions permanently burned into the file. I would not be able to use VituixCAD to generate the FRDs from the raw impulse files. I don't know how much a difference this would make, because, after I do the merging using VCAD's NF+FF+Diffraction model, the resulting set of merged polar Frds should, in theory, look the same. But I would need to test this to make sure.
Photolink: I'm still on the fence about ordering one of these mics. But if I do, my plan would be to place the Photolink device off to the side of the room by my amplifier. I keep my amp, laptop, UR22, jig, etc., off to the side and well behind the DUT to avoid reflections. I then run long cables to the DUT. The OmniMic help file says that you can bounce the infrared signal off a wall, if needed, to reach the microphone pickup sensor. Probably not the most elegant solution, but it should work OK.
Upside: On the upside, the new mic can make distortion and compression measurements at SPL levels much higher than my current SoundID mic. And there is also a new feature called "Spectral Contamination" that looks very interesting. I'd like to check this out and see how it works.
@4thtry said:
...so this should actually go a little bit faster than my current Soundeasy V20 capture and save process.
Why would you be using SoundEasy for this? ARTA and REW both have way more functional systems in place for gathering spatial data, especially if the end goal is VituixCAD.
In ARTA - record spatial impulse response group. Enter file naming prefix, angle increments and destination folder. Then it's single click operation, click, rotate, click, rotate, done. Raw impulse response is saved as PIR files for easy bulk processing through VituixCAD IR-FR tool.
**In REW ** - naming with numerical increments is easy, then it's just Ctrl+m to measure, rotate speaker, repeat. IR for the bulk set can be processed in bulk directly in REW, and exported in bulk when complete.
The downside, however, is that the new OmniMic software does not appear to have the ability to save each measurement as a raw impulse file.
Well, there's these options under the file menu..
It would always be best to save the raw IR. Omnimic is lacking in windowing options, it has only one window type and no left window, so saving impulse response directly and processing it in either VituixCAD IR-FR tool or REW would be the best way to go.
Photolink: I'm still on the fence about ordering one of these mics.
I'm interest to know, if you already have a UR22, what is the appeal of the Omnimic to you? You can get a pretty decent mic for USD$500 cost of the Omnimic, and TBH if all you want is something decent for distortion analysis, Behringer B-5 is a pretty darn good option for low cost. I have one, and it does every bit as well, and should do better at higher SPL than my Line Audio mic, since it has a larger condenser diaphragm, I've only tested up to 115dB so far. However the one I have has a small 60Hz hum, so as long as you can disregard a blip at 60Hz it does really well for a low cost mic. No calibration, but you can create your own by substitution method and your existing SoundID mic.
Upside: On the upside, the new mic can make distortion and compression measurements at SPL levels much higher than my current SoundID mic. And there is also a new feature called "Spectral Contamination" that looks very interesting. I'd like to check this out and see how it works.
Looks like simply intermodulation test with the fundamental frequencies separated for easier viewing. This has been available for a long time already with REW. Run a multitone test with REW's generator and RTA, select "show distortion" on the RTA and save the measurement, then look at the distortion tab for the saved measurement. The result is the residual audio with multitone fundamental removed.
Of course, REW also has this fancy new FSAF feature as well, allowing for a similar result with any audio at all, so pick your favourite speaker destroyer clips, REW will show the spectrum of the fundamental and residual, which is a "total distortion" result including all harmonic, intermodulation and distortion products. Best of all, you can listen to the results which can be quite enlightening, the small nuances in audio now are clear as day when you can separate them from the original audio entirely. Downside of FSAF is that it requires a decent mic and a very quiet recording space for good results.
I wonder how well this photo link thing will work when it's not pointed at the mic, as will be the case for full 360 deg measurement datasets? It looks like any remote control IR blaster so you may have to set up a mirror lol.
I'll retract this statement, some brain-fart thinking the photo link would be sitting on top of the speaker being measured, but it can be anywhere in the room with line of sight to the mic, so it's functional for this purpose.
I tried these options, the save as WAV option did literally nothing. Save impulse response curve saves as a text format, an odd format that REW will not work with it, VituixCAD IR-FR tool will open it but not in a good way. Despite Omnimic V1/2 operating at 48kHz, the saved impulse response appears at a ~32kHz sample rate. VituixCAD allows for a minimum 44.1kHz sample rate for impulse response import, so the file can't be imported correctly. The data also appears to be normalized, so SPL is lost in the export making the impulse file mostly useless anyway.
Thanks for the detailed responses, dcibel, Much appreciated. Again, you make some very good points.
ARTA/REW capture: I have used ARTA and REW for capture in the past, but I was not aware of these features. Looks like I need to make a change in my workflow.
Saving IR files from OmniMic: I was aware of these options in OmniMic's drop down file menu. I get the same response that you did when clicking on "Save IR as WAV file" Nothing happens. When I click on "Save Impulse Response curve" I get a very small file. Raw Impulse files are huge, like 3 to 7 megs each. The file I got was only a few kilobytes in size.
Distortion measurements: You make some very good points. Maybe I was thinking that as long as I was spending this amount of money, I might as well get calibration out to 40k as well. Then I can measure distortion out to 20k. The downside is that then I would have to calibrate my amplifier as well. The amp that I am using is down about 1dB at 5Hz and down about 1.5dB at 40kHz. This would have to be corrected somehow and I do not see an option in OmniMic to add an amplifier calibration file. And you can only include one calibration file per microphone. So I would need to manually modify my microphone calibration file to include the effects of amplifier roll off.
@4thtry said:
Thanks for the detailed responses, dcibel, Much appreciated. Again, you make some very good points.
ARTA/REW capture: I have used ARTA and REW for capture in the past, but I was not aware of these features. Looks like I need to make a change in my workflow.
I've said 100 times before, just follow the measurement guides for VituixCAD, they're very good and provide the detailed steps of what I described above.
Saving IR files from OmniMic: I was aware of these options in OmniMic's drop down file menu. I get the same response that you did when clicking on "Save IR as WAV file" Nothing happens. When I click on "Save Impulse Response curve" I get a very small file. Raw Impulse files are huge, like 3 to 7 megs each. The file I got was only a few kilobytes in size.
It's a lot different than what you'd get from a dedicated measurement from REW or ARTA. If I run a measurement in REW, by default it contains 1 second of audio prior to the measurement, and truncates the result after 1.7 seconds regardless of the measurement length, so 2.7 seconds of audio in total. At 96kHz, 262144 samples.
ARTA impulse is defined by the sequence length you select when measuring. For exmaple, 128k sequence is 131072 samples, which is 1.365 seconds of audio at 96kHz.
Omnimic on the other hand is incapable of recording any "dead air" prior to the measurement, that whole playback and record operations are not linked thing again, and if you look at the impulse display in Omnimic, it provides only 16ms of data, 1ms before the impulse and 15ms after, and the sample rate is screwy at ~32kHz, the file contains only 438 samples. It's not unusable, but it's a lot less data than what you get from the alternatives.
Distortion measurements: You make some very good points. Maybe I was thinking that as long as I was spending this amount of money, I might as well get calibration out to 40k as well. Then I can measure distortion out to 20k. The downside is that then I would have to calibrate my amplifier as well. The amp that I am using is down about 1dB at 5Hz and down about 1.5dB at 40kHz. This would have to be corrected somehow and I do not see an option in OmniMic to add an amplifier calibration file. And you can only include one calibration file per microphone. So I would need to manually modify my microphone calibration file to include the effects of amplifier roll off.
Create custom calibration, or incorporate more post-processing steps in your process, but personally I wouldn't be too concerned about those small rolloff factors at the extremes, though they are both measurable and avoidable with that pesky dual channel thing.
My mic rolls off past 20kHz, but I still measure at 96kHz regardless, mostly because I can so why not? I'm not too concerned about knowing whether the response at 30kHz is 95dB or 98dB, so you really have to consider what you will do with the information of that calibrated "perfect" response past 20kHz.
@dcibel said:
It's a lot different than what you'd get from a dedicated measurement from REW or ARTA. If I run a measurement in REW, by default it contains 1 second of audio prior to the measurement, and truncates the result after 1.7 seconds regardless of the measurement length, so 2.7 seconds of audio in total. At 96kHz, 262144 samples.
ARTA impulse is defined by the sequence length you select when measuring. For exmaple, 128k sequence is 131072 samples, which is 1.365 seconds of audio at 96kHz.
Omnimic on the other hand is incapable of recording any "dead air" prior to the measurement, that whole playback and record operations are not linked thing again, and if you look at the impulse display in Omnimic, it provides only 16ms of data, 1ms before the impulse and 15ms after, and the sample rate is screwy at ~32kHz, the file contains only 438 samples. It's not unusable, but it's a lot less data than what you get from the alternatives.
Under the file menu, I got the "Save IR as WAV file" option to work by uninstalling OmniMic version 6.057 and re-installing OmniMic version 4.7. I then loaded the wav file into VituixCAD's IR-FR tool and exported it as an ascii text file. The text file is much larger than before and the header says that it has 29759 numpoints and a sample rate of 48kHz. The help file says that "The time length of the saved file will be about 620ms at the measurement sample rate (48ksps, or 96ksps for OmniMic 40K)". So OmniMic version 6.057 must have a bug.
My mic rolls off past 20kHz, but I still measure at 96kHz regardless, mostly because I can so why not? I'm not too >concerned about knowing whether the response at 30kHz is 95dB or 98dB, so you really have to consider what you >will do with the information of that calibrated "perfect" response past 20kHz.
Yes I'm not certain what the advantages of measuring up to 40KHz.
If I see a tweeter dome resonance, should I apply an ultrasonic notch filter, or not?
Yesterday, I found out my omnimic is failing. I have reached out to Cross Spectrum Labs to get it repaired. In the mean time I see so many others enjoying their duel channel measurement setups.
This and the $20 Beringer look like an easy entry point but the Dayton EMM-6 is probably the best value since it has a calibration file.
Please let me know your thoughts and recomendations.
My mic rolls off past 20kHz, but I still measure at 96kHz regardless, mostly because I can so why not? I'm not too >concerned about knowing whether the response at 30kHz is 95dB or 98dB, so you really have to consider what you >will do with the information of that calibrated "perfect" response past 20kHz.
Yes I'm not certain what the advantages of measuring up to 40KHz.
If I see a tweeter dome resonance, should I apply an ultrasonic notch filter, or not?
I haven't been able to produce any test that shows a difference in the audible frequency range.
@kenrhodes said:
...
This and the $20 Beringer look like an easy entry point but the Dayton EMM-6 is probably the best value since it has a calibration file.
Please let me know your thoughts and recomendations.
It depends on whether the Dayton Calibration files have improved. I compared the Dayton (but a fairly old one), Behringer and Earthworks HERE
Yesterday, I found out my omnimic is failing. I have reached out to Cross Spectrum Labs to get it repaired. In the mean time I see so many others enjoying their duel channel measurement setups.
This and the $20 Beringer look like an easy entry point but the Dayton EMM-6 is probably the best value since it has a calibration file.
Please let me know your thoughts and recommendations.
If you have interest in distortion measurements, I'd avoid both. Emm-6 calibration is contaminated with noise, honestly it does more harm than good, best to take the cal file and apply 1/3 oct or 1/1 oct smoothing before use. Good news is these mics are generally very flat without the calibration file. For the same mic and a more reliable calibration, look at Sonarworks SoundID.
For distortion evaluation, Behringer B-5 is a good option, however requires a calibration file to provide accurate frequency response in the top end. Easy to do with a known good reference, but not necessary if it's only purpose is distortion. So option here would be SoundID + B-5 for FR and distortion purpose.
I use a Line Audio Omni1, it's not a "measurement mic" in that it does not come with a calibration file and the chassis is not a long stick. But, it is very linear, "passively calibrated" to match an Earthworks reference. Not needing a calibration file has it's plug-and-play benefits. This mic performs admirably for both FR and distortion, I believe it is one of the best values on the market today.
Comments
You can bedazzle your loopback cable if you like.
I was doing some more thinking about this PhotoLink device and how it compares to my Steinberg UR22 MKii + Arta jig loopback setup. Since all you have to do is connect the PhotoLink box to the speaker wires and point it at the microphone, the new Omnimic 40K will probably be almost as fast to set up and start taking measurements as the old OmniMic was. In addition, the PhotoLink eliminates the possiblity of overdriving and burning the inputs of the microphone soundcard. When setting up my UR22 to take measurements, I am always very careful to turn input levels way down. Then I do multiple test runs, slowly bringing up the levels until everything looks good. I need to follow this procedure during every setup to prevent accidental UR22 input overload. So the PhotoLink idea was probably, as you suggested, for quick "plug and play" setups. But it also helps to prevent loopback overdrive burnout situations.
Add a couple zener diodes to your jig to clamp down the input voltage if your prone to doing silly things like cranking out 50V with input gain set to max, they cost peanuts. I'd be more worried for your speakers honestly.
I use the signal generator to set levels, it takes seconds, and on my setup the input gain stays at the same 9 o'clock position 95% of the time.
If all you need is a timing reference, which is all the optical thing is going to provide, just use semi-dual loopback as I showed in the photo above. It's all line level signals there, low risk but input gain can still be something like +50dB with input gain knob at max ( max gain is +40dB for UR22mkii. )
The physical connections is a piece of cake.
Changing processes is the harder thing to do.
If one cannot see an advantage or benefit from changing their process, they’ll be inclined to continued the same tried and trusted process. I've seen measurements made by companies using their own software or legacy systems like Linear X LMS (DOS era) and WinMLS 2000 Pro (Windows 2000 era0 and I can see why they are reluctant to change:
If it ain’t broke, Why fix it?
I enjoy using different parts as much as the next person eg. drivers, crossovers.
But beyond a 6 sided rectangular prism of a loudspeaker cabinet, I'm interested in seeing the effect of changes to dispersion and diffraction from the baffle size and shape, the placement of drivers on the baffle in 3 dimensions, listening to the effects of filters without having to build them (passive or active), and how a speaker interacts with a room in 4D.
What does (semi) dual channel measurements offer that Omnimic40K doesn't?
What does Omnimic40K offer that dual channel measurements doesn't?
As far as a timing reference goes, not much. But, if you compare what a USB audio Interface offers that a USB mic doesn't, it's a heck of a lot. A full dual channel setup offers a lot more as well.
Absolute SPL without needing a sound level calibrator. But for most general use, absolute SPL isn't that important, I only got an SPL calibrator once I got deep into distortion analysis.
I heard the Omnimic 40k box also has a little light that lights up when the signal level is 2.83V +/- some percent. Maybe a nice convenience feature for some.
I still haven't seen what has been done to update the proprietary software for this new mic. Bill said it himself, that it's hard to compete with a viable for-profit product when your main competition is well executed freeware (REW). At the very least, for the timing reference to make any sense, it needs some spatial measurement tool to easily create subsequent measurements with automatic file naming. REW and ARTA have been able to do it for years. The software needs the ability to adjust windowing and smoothing settings without remeasuring, because that's just silly. Also it should allow export of the raw impulse response for processing in other software.
Honestly I'd be a lot more excited if Dayton would offer a decent measurement condenser mic with XLR connection, price at $150-$180 range. They could skip the software development and RND for some odd timing solution for all I care, just sell a good mic at a reasonable cost.
What do you feel are the top 3 most useful things that a USB audio interface and full dual timing setup offers?
There was a proposal from Bill to the product managers at Parts Express/Dayton about a standalone microphone. It is good up to 50KHz, extremely rugged and suitable for everyday use/abuse/accidental drops.
Unfortunately it requires skills quite different to a typical electronics fab and the labor would probably drive up the price to the point it’s not commercially viable against their current line of XLR mics.
The upside?
Bill released it as a DIY project- for the skilled assemblers amongst us!
https://www.diyaudio.com/community/threads/a-diy-mems-measurement-microphone.417231/
That will vary depending on the person and their needs. There are many reasons one may consider a USB audio Interface over a canned USB mic solution. The biggest for me when moving away from Omnimic was SNR. The old Omnimic has low end electronics internally which results in low SNR, even the cheapest USB audio Interfaces outperform it by a country mile. A USB audio Interface can be used as a PC DAC for any purpose. It allows one to record any audio input, swap mics, use for a simple PA setup, PC audio for zoom calls, podcasts, and if you're a musician of any sort who records what you play you may already have one. It allows an easy upgrade path to a full dual channel setup for only a few dollars spent, and measuring impedance as well, so a USB interface, a few resistors and some scrap wire can cover the functions of acoustic and electrical measurement, so you can now compare to the costs of interface+standard mic to that of a USB mic and DATS combined. A USB audio Interface also doesn't require proprietary software, you can choose between several software solutions, so the product longevity is a lot longer in my opinion and is simply more versatile.
A full dual channel system simply provides a feedback loop that is at the driver end of the signal chain rather than at the interface (semi-dual). Any response changes along the signal path due to the amp used, cabling, or filters is removed from the measurement allowing increased accuracy ensuring the measured result is purely the speaker with no external influence. Think of it as an auto-cal with every measurement, as the measurement is a differential between mic input and electrical feedback. It allows for insertion of a protection cap without messing up the driver response, avoiding extra post-processing steps, and opens up opportunities for more advanced signal analysis. I've used mine to provide filter transfer functions of both passive and active DSP filters for design validation.
Anyway, enough rambling on, I imagine everyone here has already made their choice on what they want to use. For anyone who hasn't, the main point to drive home is that if you have a 2 channel USB audio Interface, measuring with accurate timing to get started requires no extra diy jig whatsoever, just a patch cable.
In case anyone is interested, the new OmniMic40k software package is available for download from the Dayton Audio website. It is version 6.052 and works with both the old and new OmniMic hardware. The help file has detailed info on general setup, new feature listing, photolink setup, setting impulse response window edges, harmonic distortion percentages, a new compression testing feature, and info on how the 40k calibration was performed. Check it out:
https://daytonaudio.com/product/2111/omnimic-40k-precision-measurement-microphone-system
Looks like the DA32TX has returned...
Many are now back-order-able
TC7FD, TG9- exemplary full ranges:
http://feleppa.com.au/speakermeasmid.html
830845- 12”, 8 ohm fibre cone (I use them in pairs, sealed for ~92dB/2.83V after BSC):
https://www.diyaudio.com/community/threads/15-woofer-suggestions-for-end-game-3-way.402411/post-7896830
835016- 10” 4 ohm alu cone; great in small sealed cabinets for woofers, or as pairs in dipole, as in NaO II RS:
http://musicanddesign.speakerdesign.net/products.html
Happy Days!
And the DA25TX too!
Man, I forgot how annoying it was to constantly bounce back and forth between sweep playback and measurement recording, generally needing at least 3 sweeps to play before you get one good one. Anyway, there's no easy spatial measurement function here for this reason, playback and record are decoupled so repeated measurements for "spinorama" data will be quite tedious with the Omnimic 40k.
I wonder how well this photo link thing will work when it's not pointed at the mic, as will be the case for full 360 deg measurement datasets? It looks like any remote control IR blaster so you may have to set up a mirror lol.
Much of my complaints about Omnimic still appear to be true with the new 40k version. Solen lists this thing at CAD$792 (yikes), for comparison I could buy a Motu M2, Line Audio Omni1 and still have a bunch of money left over in the bank, just sayin..
There are a ton of peerless coming back, not just a few. Exciting times ahead as I suspect we will see plenty of new designs. The HDS 6.5 with phase plug is under $60 and you have to spend a lot more on other 6.5s with similar performance.
I LOVE the HDS Peerless
I have the 6.5" and 8" HDS nomex cones. Are you referring to the silver phase plug ones?
IIRC, the only Peerless I know with phase plugs were the Exclusives and M series. The M and Ex both had the same HDS type frames though. The M hasn't been available for quite some time now.
InDIYana Event Website
Yes, HDS Exclusive:
https://www.parts-express.com/Peerless-830883-6-1-2-Nomex-Cone-HDS-Woofer-264-1094?quantity=1
Thanks for your comments, dcibel. You make some very good points.
Playback/record issue: My thinking is that the playback/record issue shouldn't be too bad with the new mic. I know that when I had previously recorded polars with my old OmniMic that I would often need to cycle through multiple "whoop-whoop" cycles before hitting pause to record, because the phase curve kept drifting up and down in a repeating cycle with each whoop. But, according to the help file, the new mic's phase capture should be rock solid with each whoop. So I should be able to hit pause after a couple whoops, then hit save, then rotate the speaker, etc. OmniMic has an auto naming feature, so this should actually go a little bit faster than my current Soundeasy V20 capture and save process. The downside, however, is that the new OmniMic software does not appear to have the ability to save each measurement as a raw impulse file. So I would have to save each measurement as a blended FRD with my gating decisions permanently burned into the file. I would not be able to use VituixCAD to generate the FRDs from the raw impulse files. I don't know how much a difference this would make, because, after I do the merging using VCAD's NF+FF+Diffraction model, the resulting set of merged polar Frds should, in theory, look the same. But I would need to test this to make sure.
Photolink: I'm still on the fence about ordering one of these mics. But if I do, my plan would be to place the Photolink device off to the side of the room by my amplifier. I keep my amp, laptop, UR22, jig, etc., off to the side and well behind the DUT to avoid reflections. I then run long cables to the DUT. The OmniMic help file says that you can bounce the infrared signal off a wall, if needed, to reach the microphone pickup sensor. Probably not the most elegant solution, but it should work OK.
Upside: On the upside, the new mic can make distortion and compression measurements at SPL levels much higher than my current SoundID mic. And there is also a new feature called "Spectral Contamination" that looks very interesting. I'd like to check this out and see how it works.
Why would you be using SoundEasy for this? ARTA and REW both have way more functional systems in place for gathering spatial data, especially if the end goal is VituixCAD.
In ARTA - record spatial impulse response group. Enter file naming prefix, angle increments and destination folder. Then it's single click operation, click, rotate, click, rotate, done. Raw impulse response is saved as PIR files for easy bulk processing through VituixCAD IR-FR tool.
**In REW ** - naming with numerical increments is easy, then it's just Ctrl+m to measure, rotate speaker, repeat. IR for the bulk set can be processed in bulk directly in REW, and exported in bulk when complete.
Well, there's these options under the file menu..
It would always be best to save the raw IR. Omnimic is lacking in windowing options, it has only one window type and no left window, so saving impulse response directly and processing it in either VituixCAD IR-FR tool or REW would be the best way to go.
I'm interest to know, if you already have a UR22, what is the appeal of the Omnimic to you? You can get a pretty decent mic for USD$500 cost of the Omnimic, and TBH if all you want is something decent for distortion analysis, Behringer B-5 is a pretty darn good option for low cost. I have one, and it does every bit as well, and should do better at higher SPL than my Line Audio mic, since it has a larger condenser diaphragm, I've only tested up to 115dB so far. However the one I have has a small 60Hz hum, so as long as you can disregard a blip at 60Hz it does really well for a low cost mic. No calibration, but you can create your own by substitution method and your existing SoundID mic.
Looks like simply intermodulation test with the fundamental frequencies separated for easier viewing. This has been available for a long time already with REW. Run a multitone test with REW's generator and RTA, select "show distortion" on the RTA and save the measurement, then look at the distortion tab for the saved measurement. The result is the residual audio with multitone fundamental removed.
Of course, REW also has this fancy new FSAF feature as well, allowing for a similar result with any audio at all, so pick your favourite speaker destroyer clips, REW will show the spectrum of the fundamental and residual, which is a "total distortion" result including all harmonic, intermodulation and distortion products. Best of all, you can listen to the results which can be quite enlightening, the small nuances in audio now are clear as day when you can separate them from the original audio entirely. Downside of FSAF is that it requires a decent mic and a very quiet recording space for good results.
I'll retract this statement, some brain-fart thinking the photo link would be sitting on top of the speaker being measured, but it can be anywhere in the room with line of sight to the mic, so it's functional for this purpose.
I tried these options, the save as WAV option did literally nothing. Save impulse response curve saves as a text format, an odd format that REW will not work with it, VituixCAD IR-FR tool will open it but not in a good way. Despite Omnimic V1/2 operating at 48kHz, the saved impulse response appears at a ~32kHz sample rate. VituixCAD allows for a minimum 44.1kHz sample rate for impulse response import, so the file can't be imported correctly. The data also appears to be normalized, so SPL is lost in the export making the impulse file mostly useless anyway.
Thanks for the detailed responses, dcibel, Much appreciated. Again, you make some very good points.
ARTA/REW capture: I have used ARTA and REW for capture in the past, but I was not aware of these features. Looks like I need to make a change in my workflow.
Saving IR files from OmniMic: I was aware of these options in OmniMic's drop down file menu. I get the same response that you did when clicking on "Save IR as WAV file" Nothing happens. When I click on "Save Impulse Response curve" I get a very small file. Raw Impulse files are huge, like 3 to 7 megs each. The file I got was only a few kilobytes in size.
Distortion measurements: You make some very good points. Maybe I was thinking that as long as I was spending this amount of money, I might as well get calibration out to 40k as well. Then I can measure distortion out to 20k. The downside is that then I would have to calibrate my amplifier as well. The amp that I am using is down about 1dB at 5Hz and down about 1.5dB at 40kHz. This would have to be corrected somehow and I do not see an option in OmniMic to add an amplifier calibration file. And you can only include one calibration file per microphone. So I would need to manually modify my microphone calibration file to include the effects of amplifier roll off.
I've said 100 times before, just follow the measurement guides for VituixCAD, they're very good and provide the detailed steps of what I described above.
It's a lot different than what you'd get from a dedicated measurement from REW or ARTA. If I run a measurement in REW, by default it contains 1 second of audio prior to the measurement, and truncates the result after 1.7 seconds regardless of the measurement length, so 2.7 seconds of audio in total. At 96kHz, 262144 samples.
ARTA impulse is defined by the sequence length you select when measuring. For exmaple, 128k sequence is 131072 samples, which is 1.365 seconds of audio at 96kHz.
Omnimic on the other hand is incapable of recording any "dead air" prior to the measurement, that whole playback and record operations are not linked thing again, and if you look at the impulse display in Omnimic, it provides only 16ms of data, 1ms before the impulse and 15ms after, and the sample rate is screwy at ~32kHz, the file contains only 438 samples. It's not unusable, but it's a lot less data than what you get from the alternatives.
Create custom calibration, or incorporate more post-processing steps in your process, but personally I wouldn't be too concerned about those small rolloff factors at the extremes, though they are both measurable and avoidable with that pesky dual channel thing.
My mic rolls off past 20kHz, but I still measure at 96kHz regardless, mostly because I can so why not? I'm not too concerned about knowing whether the response at 30kHz is 95dB or 98dB, so you really have to consider what you will do with the information of that calibrated "perfect" response past 20kHz.
Under the file menu, I got the "Save IR as WAV file" option to work by uninstalling OmniMic version 6.057 and re-installing OmniMic version 4.7. I then loaded the wav file into VituixCAD's IR-FR tool and exported it as an ascii text file. The text file is much larger than before and the header says that it has 29759 numpoints and a sample rate of 48kHz. The help file says that "The time length of the saved file will be about 620ms at the measurement sample rate (48ksps, or 96ksps for OmniMic 40K)". So OmniMic version 6.057 must have a bug.
Yes I'm not certain what the advantages of measuring up to 40KHz.
If I see a tweeter dome resonance, should I apply an ultrasonic notch filter, or not?
https://www.sweetwater.com/store/detail/ECMPro--behringer-ecm-pro-measurement-microphone?_queryID=257976a2bb1888fd08ab9c073ff53da9&_index=production_products
Yesterday, I found out my omnimic is failing. I have reached out to Cross Spectrum Labs to get it repaired. In the mean time I see so many others enjoying their duel channel measurement setups.
This and the $20 Beringer look like an easy entry point but the Dayton EMM-6 is probably the best value since it has a calibration file.
Please let me know your thoughts and recomendations.
I haven't been able to produce any test that shows a difference in the audible frequency range.
It depends on whether the Dayton Calibration files have improved. I compared the Dayton (but a fairly old one), Behringer and Earthworks HERE
I got this one
https://www.bhphotovideo.com/c/product/1668735-REG/sonarworks_12_41468_soundid_reference_measurement_mic.html
It comes with cal for 0 and 90 degrees.
If you have interest in distortion measurements, I'd avoid both. Emm-6 calibration is contaminated with noise, honestly it does more harm than good, best to take the cal file and apply 1/3 oct or 1/1 oct smoothing before use. Good news is these mics are generally very flat without the calibration file. For the same mic and a more reliable calibration, look at Sonarworks SoundID.
For distortion evaluation, Behringer B-5 is a good option, however requires a calibration file to provide accurate frequency response in the top end. Easy to do with a known good reference, but not necessary if it's only purpose is distortion. So option here would be SoundID + B-5 for FR and distortion purpose.
I use a Line Audio Omni1, it's not a "measurement mic" in that it does not come with a calibration file and the chassis is not a long stick. But, it is very linear, "passively calibrated" to match an Earthworks reference. Not needing a calibration file has it's plug-and-play benefits. This mic performs admirably for both FR and distortion, I believe it is one of the best values on the market today.
For more insight, I recommend reading through @tktran 's thread at DIYAudio here:
https://www.diyaudio.com/community/threads/in-search-of-low-distortion-omnidirectional-microphones-for-diyers.412011/