• 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!

UMIK-1 or any USB MIC unsuitable for timing measurements?

Emlin

Addicted to Fun and Learning
Joined
Jul 8, 2018
Messages
845
Likes
1,172
So you think they go to the trouble of batch testing, and then scramble the SN's randomly so people don't know it.

BTW, how do you batch test mics like this? You stick ten up and do it all at once and average. You test 10 one at a time and average the results for that group. You do realize since they measured it would be easier to just use the results for each mic. The batch testing idea would be more trouble to them. If I'm guessing, I'd guess CSL has the more accurate calibration, but I don't see much reason to think these others are not individually done.
You take a few mics at random and average the results.
 

dcibel

Member
Joined
Feb 8, 2021
Messages
47
Likes
53
If one thinks the Dayton is too bumpy, using 1/6 octave smoothing will take care of it.
1/6oct smoothing isn't going to do you any favours.
1711911945326.png
 
OP
L

levimax

Major Contributor
Joined
Dec 28, 2018
Messages
2,456
Likes
3,622
Location
San Diego
So you think they go to the trouble of batch testing, and then scramble the SN's randomly so people don't know it.

BTW, how do you batch test mics like this? You stick ten up and do it all at once and average. You test 10 one at a time and average the results for that group. You do realize since they measured it would be easier to just use the results for each mic. The batch testing idea would be more trouble to them. If I'm guessing, I'd guess CSL has the more accurate calibration, but I don't see much reason to think these others are not individually done.
A lot of manufacturing is done in "batches" where the raw materials for a batch comes in and a batch of product is produced. In this case rather than test every Mic they would, using some statistical analysis based on previous batches, randomly pick a number of Mics from the beginning, middle, and end of the run and test them and then assume a "normal distribution" (assuming they weren't seeing "trends" in the run) come up with a "average" calibration file for that batch and then make that file available for download for all the serial numbers in that batch
 

DVDdoug

Major Contributor
Joined
May 27, 2021
Messages
3,165
Likes
4,178
A lot of manufacturing is done in "batches" where the raw materials for a batch comes in and a batch of product is produced. In this case rather than test every Mic they would, using some statistical analysis based on previous batches, randomly pick a number of Mics from the beginning, middle, and end of the run and test them and then assume a "normal distribution" (assuming they weren't seeing "trends" in the run) come up with a "average" calibration file for that batch and then make that file available for download for all the serial numbers in that batch.
That's perfectly valid for a "recording mic". They might not even measure the frequency response of every mic. If you buy a Shure mic it's not going to exactly-match the published frequency response.

But IMO that would be dishonest for a calibration mic with a file tied to the serial number or for a calibration mic with a specification. Anything used for measurement should be individually verified.

If you're paranoid you can send your mic to an independent calibration lab. I think that costs around $100 USD, plus shipping. If you are doing "serious work" your equipment should usually be periodically calibrated. And there are always tolerances.
 

Emlin

Addicted to Fun and Learning
Joined
Jul 8, 2018
Messages
845
Likes
1,172
Then why would each cal file differ? What you are describing is extra bother for them.
You have no proof that they all do differ, as I explained above. I honestly don't know why you are getting so aerated about this.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,217
Likes
38,582
I guess the batch testing meme is strong. Beats me why? Carry on.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,217
Likes
38,582
You have no proof that they all do differ, as I explained above. I honestly don't know why you are getting so aerated about this.
Find two files that are the same.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,217
Likes
38,582
That's perfectly valid for a "recording mic". They might not even measure the frequency response of every mic. If you buy a Shure mic it's not going to exactly-match the published frequency response.

But IMO that would be dishonest for a calibration mic with a file tied to the serial number or for a calibration mic with a specification. Anything used for measurement should be individually verified.

If you're paranoid you can send your mic to an independent calibration lab. I think that costs around $100 USD, plus shipping. If you are doing "serious work" your equipment should usually be periodically calibrated. And there are always tolerances.
I doubt Shure measures every mic. Maybe a sampling for QC. They do make them consistently. Here are three KSM 32's. Time of manufacture covers 11 years. The one slightly different in the upper frequencies is an older one with a different style mesh over it. The low end variance is noise. You can get that level of difference measuring the same mic twice due to fluctuations in below 100 hz noise.

1711916242933.png
 

Salt

Addicted to Fun and Learning
Forum Donor
Joined
Aug 7, 2023
Messages
764
Likes
452
Location
DE
May be I'm completely wrong, but for me this time delay is only a hypothetical fault.
But first: there is more than one software on this planet besides REW, as Arta or Hypex Filter Design, that all would struggle the same problem, not?
Otherwise: is it a real world problem? or only an obstacle for simulating a speaker response (is this the main question)?

For me, as I'm old fashioned, speaker design by simulation has it's place in sub/bass and the enclosure.
Any xover design was and is a dynamic and progressive process of measurement - adjustment - measurement .... in earlier days by switching C's, L's and R's, today by modeling Q's, F's .... with speaker and mic in a fixed position, and after a satisfiying solution then mic moved vertically and horizontically to see if that fits or has to be corrected.
Don't see any problem in time lag here as the whole speaker is measured all the time.
 
Last edited:

gnarly

Major Contributor
Joined
Jun 15, 2021
Messages
1,071
Likes
1,516
What measurement software do USB mics work with, besides REW?
Do any of them NOT use an acoustic timing reference, like REW's 5kHz up sweep?
 

Salt

Addicted to Fun and Learning
Forum Donor
Joined
Aug 7, 2023
Messages
764
Likes
452
Location
DE
I think it is not software limited.
I used Arta (sweep and impulse) as well as hypex (sweep) flawlessly.
 

gnarly

Major Contributor
Joined
Jun 15, 2021
Messages
1,071
Likes
1,516
I think it is not software limited.
I used Arta (sweep and impulse) as well as hypex (sweep) flawlessly.
Cool. But in single channel mode only ?

I'll try ARTA with it, if so.
 

sam_adams

Major Contributor
Joined
Dec 24, 2019
Messages
1,057
Likes
2,640
A random sample of twenty-five (25) MiniDSP Umik-1 .cal files, 615 points per file:
UMIK-1 Cal Files.png

Cross-Spectrum Labs calibrated Umik-1 .cal files are 1500 points.

A random sample of nine (9) MiniDSP Umik-2 .cal files, 615 points per file:
UMIK-2 Cal Files.png


A random sample of sixteen (16) Dayton EMM-6 .cal files, 256 points per file:
EMM-6 Cal Files.png


A random sample of twenty (20) Dayton UMM-6 .cal files, 97 points per file:
UMM-6 Cal Files.png
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,217
Likes
38,582
A random sample of twenty-five (25) MiniDSP Umik-1 .cal files, 615 points per file:
View attachment 360491
Cross-Spectrum Labs calibrated Umik-1 .cal files are 1500 points.

A random sample of nine (9) MiniDSP Umik-2 .cal files, 615 points per file:
View attachment 360492

A random sample of sixteen (16) Dayton EMM-6 .cal files, 256 points per file:
View attachment 360493

A random sample of twenty (20) Dayton UMM-6 .cal files, 97 points per file:
View attachment 360494
Now what you need is the comparison of each mic with its file. One would hope/think the result would group much closer together. A couple of those are pretty out there.
 

Salt

Addicted to Fun and Learning
Forum Donor
Joined
Aug 7, 2023
Messages
764
Likes
452
Location
DE
Here my cals, one delivered with Umik (first one), one by vendor (second one):
 

Attachments

  • 7034746.txt
    9.8 KB · Views: 15
  • UMIK1NH2_00d.txt
    2.2 KB · Views: 12

gnarly

Major Contributor
Joined
Jun 15, 2021
Messages
1,071
Likes
1,516
Single channel one speaker, what else while constructing?
Don't know what you mean by "what else while constructing" ?

Just tried a Umik-1 with ARTA. (I gave up on single channel so long ago, I'm surprised I still had the mic.)
Hey, very nice to see it works like you say! .. with windows WDM driver. Thx.:)

All that said, screw single channel ....LOL
 

Salt

Addicted to Fun and Learning
Forum Donor
Joined
Aug 7, 2023
Messages
764
Likes
452
Location
DE
Don't know what you mean by "what else while constructing" ?

Just tried a Umik-1 with ARTA. (I gave up on single channel so long ago, I'm surprised I still had the mic.)
Hey, very nice to see it works like you say! .. with windows WDM driver. Thx.:)

All that said, screw single channel ....LOL
Under construction means one mic ahead of one speaker.
 
Top Bottom