Alterations in RS calibration files

[Dent]

Sonnie,

Just saw a recent post around May 12th regarding a possible corruption in the calibration files and that you have now fixed it. What was the problem and what did you do to fix it? I had used the previous calibration file that you posted (but still the recent ones based on your tests with the two analog and two digital meters) and everything looked ok to me.

Warner

[brucek]

Warner,

What was the problem and what did you do to fix it?

Long story.

To create the calibration files, we don't have the necessary equipment to do it professionally, but we decided that since Sonnie had a calibrated ECM mic/meter we might be able to use it as a 'standard' and come up with something relatively close for the RS meters that was better than the one that's been around for quite a few years.

The final test would be taking a measurement with Sonnies calibrated ECM mic/meter against a Radio Shack meter with our own calibration file to see if they measured the same. If so, we felt this was close enough for anyone to use the cal file we created. As we've always said, these are not professionally created calibrations (but they are close).

Our method of creating the calibration file was fairly simplistic. We took the same measurement with the calibrated ECM and an RS meter using RoomEQ Wizard and exported these mdat files to text. A lot of care was taken to remove as many variables as possible. Sonnie wasn't allowed to breath during the tests. The difference for each frequency point was the new raw calibration data.

There are many fast spiky anamalies in this data in areas where the response changes rapidly and for various other reasons such as small positional differences in the mic elements. So, we smooth the data points to create the final cal file. Admittedly, using comparison measurement data as we've done for a method of calibration is simply not as accurate as having it professionally done.

The file that inadvertently got on the server was an unsmoothed file. Several people downloaded it before we noticed and got the correct smoothed file back.

That's the story....

brucek

[Sonnie]

Gee... I don't even remember, unless it was the way the .cal file was saving the values. That was just a matter of changing the mime type on the .cal file extension so that it would download properly.

You have a link to that post maybe? I took a quick look but couldn't spot it right off.

If you have the old analog meter then nothing changed from the previous file and the recent test. Only the new analog and new digital meter files were different and new .cal files developed for those.

I probably need to send a message to all members to let them know about the new updated files and the new Excel worksheet corrections too.

[Sonnie]

Okay... I sat on my post for a few minutes while being interrupted watching the recorded American Idol... these Alabama folks just keep doing so well ... anyway brucek has posted in the meantime.

Yeah... I found the thread and post too and that was exactly what happened.

Had you used that old file there would not have been all that much difference.

[Dent]

Thanks for the response guys but maybe I was looking at a different set of files that differed. The first one I had download a couple of weeks ago and have been using, when opened using Notepad, causes all the numbers in appear in one long horizontal line but when I open it up in Wordpad, I can then read it more clearly because all the values are vertical. I just downloaded the calibration file again and now when I open it up in Notepad it is vertical just like when I open it in Wordpad.

Here are the two files I downloaded:
newrsdigitalold.cal
newrsdigitalnew.cal

[Sonnie]

Yep... that had to do with the mime type for the extension. I had to make it an executable file type in order for it to save properly.

[DMF]

So, are the numbers I posted on the AVS Forum from the good dataset?

[Sonnie]

Yes... what Dent is referring to is the way the file downloaded. Which wouldn't effect what you posted because it wasn't a file... you posted raw numbers. What Ken is referring to is the old analog .cal file (newrs.cal) which somehow when I was closing out all the files from the bfdguide.ws site, I transferred most of the files. I must have inadvertently got a hold of one of the .cal files that had not been smoothed. During my setting up the Downloads page I again inadvertently copied the wrong file or maybe it was during an update on that page. Anywho... as soon as brucek spotted it we uploaded the correct one, which was before we did the new test on the new analog and digital. So, you got the right ones.

[DMF]

Whew!