TOPIC:

The posting above didn't post the way I inputted the columns.
Briefly, at 20.102Hz there is a 4.4dBs difference between my two calibration files and by the time it get to 40.191Hz that difference is 2.7dBs. Why the large difference between calibration files and why is that difference changing? Its as if I've got some other UMIK file by mistake? DevTeam, please explain. Thank you.

Okay, so I had a little bit of time today and graphed out my two files... there is a very obvious pattern/trend that becomes immediately apparent. The two graphed line are most divergent at 20Hz and notice they slowly come together until they finally meet (agree) at 990Hz after which point they are identical ...so apparently there is some mathematical error in the algorthim used to calculate the values below 990Hz that keeps getting further off the lower it goes. I wish I knew how to post a graph on this forum, it looks insane. Please correct the file... thank you

Here is a comparison of the two calibration files for my unit:



As you can see there is a 2-4dB discrepancy in the range below 100hz.

Using my ear as the tool, I believe the original calibration file (pink line) is more "correct" in that range.

Okay, figured out how to add a graph to a posting so...

These are my two UMIK calibration files plotted together... there is obviously something wrong with the lastest 10Hz calibration file. DevTeam, please see if you can get your supplier to fix the 'new' file soon. It is UMIK serial #7000275
Thank You

Here's a link to yet another graph of another UMIK calibration file set... has the same strange diverging issues below 1000Hz. www.hometheatershack.com/forums/spl-mete...e-46.html#post603340

DevTeam... you might want to just go over here and get caught up on the issues concerning the latest calibration file release with below 20Hz data.

Dear All,

We're just back from National holidays and just asked our supplier why the change of calibration file. We should hear back hopefully soon.

Before, we start answering this thread in more details please allow us to say that we're really trying to help answer all questions here in the most timely manner. Sometimes it takes longer to answer emails (as our team goes on vacation), we just come back and deal with all issues.

- As we're told by the supplier, the reasoning behind the change of calibration file is related to an incorrect reference microphone being used to generate the script. The process being as follow:
a) Use a "golden" sample of UMIK-1 as reference (i.e. needed for the shape and location of capsule)
b) Tune it against a reference microphone B&K in anechoic chamber
c) Measure each microphone on the factory floor for 2 state (low and high frequency)
d) Generate a calibration file based on the response of each microphone using a script
The golden sample is the reference for the process. We're trying to investigate with the supplier now what has changed on the golden sample. Thanks for your patience.

- Little to say that all this is upsetting although this supplier is a known company used by other companies as some of you already guessed...The whole reason why we're having this discussion and having to update the calibration file is because this large microphone supplier used to ship calibration file up until we told them that the calibration process was flawed thanks to the help of partners like CSL. Now we're working to make sure they ship products with a correct test jig and correct calibration file as per our standards. With all our miniDSP products done in-house, it's much easier for other products to deal with issue. The Umik-1 is actually the only product where we "rely" on a supplier for assembly and the calibration process. Make no mistake, we're not trying to switch the blame here as we stand behind all our products. However it's hopefully giving you a glimpse that we're not always 100% in control of the speed of answers here. We're sending emails daily to get to the bottom of the issue though so we'll get there to make sure you get the reliable data in the end! As per the process of above (4steps), you can see that the process is actually quite simple but the scripting and the way the initial referencing is done is key.

That's all for today, hoping this clarifies few questions. We will get back to you once we hear back from the microphone supplier. In the mean time, feel free to email if you have any questions you'd like to ask.

DevTeam

FINALLY!!! an answer that I can understand... a full explanation that makes sense. It's greatly appreciated and I will reward it with my patience as you work with your mic supplier to straighten it out. I'm assuming that even if the "Golden" mic wasn't quite so golden after all, hopefully a script can be generated to account for this imperfection since it is all about referencing anyway right? If not, then it will have to involve shipping back to the mic manufacturer for a re-calibration correct? Please keep us posted on further developments with the mic supplier.
Thank You...

I just tested the new 10 Hz cal file for my UMIK-1 (sn: 7000373) and the cal files are completely different below 990.251 Hz as monomer has already noticed. I attach a REW graph comparing the two frequency responses. When I bought this mic I did it because I wanted a calibrated mic for a decent price. The mic is useless to me if the calibration file is incorrect. Today I'm regretting the purchase both because of the problems with my first unit (which would have been acceptable if it had been the only problem) and now because of the obviously incorrect calibration files. Also, your customer support can be improved a lot. My trust for MiniDSP as a company has unfortunately been seriously degraded and I would hesitate before buying any other product from you in the future.

Just thought I would chime in here.

I recently got my Umik (#317), and while I love the plug and play nature, I too have some issues with cal files. My calibration data plots very similarly to the others in this thread. I do have another calibrated mic, and I will do a comparison when I have the time to do so, to see which calibration file is "correct". Personally - I am betting both the umik files are wrong.

Frequency response notwithstanding, there is another issue with my calibration data.

When I open REW and plug in my umik, then load the calibration files as prompted, the mic reads far too loud.

Old cal file (sensitivity = -21.1) reads 102.2 dB with a 94 dB calibrator.
New cal file (sensitivity = -18.3212) reads 99.5 dB with a 94 dB calibrator.
New 10 Hz cal file (sensitivity = -18.3212) also reads 99.5 dB with a 94 dB calibrator.

Mind you - this is with a Sound Level Calibrator, which generates a 94 dB or 114 dB 1 KHz calibration tone, so the frequency measured is unaffected by the differences in the LF calibration.

To "correct" things, I subtracted another 5.5 dB from the sensitivity line.

Corrected 10 Hz cal file (sensitivity = -12.8212) reads 94.0 dB with a 94 dB calibrator, and 113.9 dB with a 114 dB calibrator. At least with Beta 13, things are now spot-on as far as level goes. This might be an REW issue, as my OmniMic also reads about 6 dB high.

Similar to other users, my REW spectrum display also shows a spike at 1 KHz, then at 1 KHz intervals above that point, with levels that are about 25 dB above the noise floor. Admittedly, there is an easy 50 dB of signal above this level when I make a typical measurement, placing the "distortion" due to self-noise well below 1%. Still, seems like something that could be addressed without undue costs.

@ monomer,

Yes that's correct. It's indeed the reference file that is indeed the issue. The supplier now doing some test to understand that rise in frequency response. We'll updated everybody once we hear back from them. We are in daily contact as we mentioned before.

DevTeam

@ Megagurka,

Your comments about the innacuracy of that 10Hz file is the same as above. The 10Hz extension is not accurate. You can always use the 20Hz file for the time being. We're sorry to hear that you're unhappy with our service, we can only say that we're doing our best on our side to solve the issue.

Did you get to test the microphone at all? As mentioned by others, there are plenty of things that you can learn in the mean time on the process of doing measurements (ground plane, semi-anechoic..etc) and for example use that data when loaded into a processor. Just a thought..

DevTeam

@ Lilmike,

Thanks for the follow up and clear results of your investigations. With regards to REW, we are indeed in talks with John who is the author of the wonderful free REW software.

We have been beta testing a new version from John (without the 6dB) and it's solving that offset that you had to manually apply. We just need to make sure that the reading is consistent across all calibration files and that's what we've been doing. Once John compiles his new version and we are ready on our side, we'll ship a complete announcement to all UMIK-1 end users along with asking them to upgrade they REW install file to the latest package. It's in process.

As for the spikes, we did notice this on the batch #2 and solved it in subsequent batches. When doing measurements it shouldn't appear on the spectrum providing you're above the noise floor (required to make sure you have headroom). But with this said, for those of you who want to open the microphone, it's a simple rework of one CAP or few parts. We're happy to do it for you if you need help. George has been documenting few good things here:
www.minidsp.com/forum/18-umik-questions/...mic-?limit=6&start=6

We'll soon once again provide more info to all.

Hoping this info helps.

DevTeam

@devteam,

How can I tell whether or not my mic needs this revision/repair before it can be reliably used? My mic is number 7000634. Please let me know.

rmalak wrote:

@devteam,

How can I tell whether or not my mic needs this revision/repair before it can be reliably used? My mic is number 7000634. Please let me know.

I doubt you'll hear anything back from the DevTeam until Sunday evening if you are in the western hemisphere somewhere (which will then be Monday afternoon in Hong Kong).

Your serial number is so high, I'm sure its from the third batch of UMIKs that the DevTeam says already has the improvements. I believe its the second batch that's being targeted for these revised boards.