Exploring Visual Analyser (VA)

Ah, Pano, perhaps you're running the 2011 Version (last stable version)? I'm confidently striding off into the future (probably at some personal risk!) on the 2012Beta version. Seems pretty solid.

Go to:

Visual Analyser 2011 XE

Click on "Beta available" at top right of screen.

You can "Log on as guest" to avoid needing to register. Click Latest VA version, then
click on VA2012beta.zip at bottom

Terry
 
Here is a real back to basics question.

My primary use of this software for now, is getting information on power amps. The docs deal with feeding the signal to both the program and a speaker. What is the preferred method for this setup - the use of "Y" cabling or a mixer arrangement. I'm assuming the mixer would require an additional calibration step.
 
Using two signal generators I fooled with the A cross B and B cross A. It definitely seems to be a correlation function, but after reading about such I may not be smart enough to know what to do with it. Both signals seem to be displayed in the spectrum area in a combined fashion. The scope area appears to give the correlation function, going off scale when the frequencies match. Maybe somebody can give a better description of what cross correlation is good for. It might be useful to download the manual for a commercial 2-channel vibration analyzer like SRS and read the section on correlation functions, as I know they have them.

BTW, I plugged in my Wavetek and my HP8903, which have an order of magnitude THD difference, and VA seemed to read them perfectly after using the calibration button. It read more than 10X too high before doing that. I'm not exactly sure why that calibration is even necessary.
 
Last edited:
Does seem odd, doesn't it. Calibration was always necessary when using the "null out the fundamental" approach, like my hardware N&D set does. But I can't see why it would be necessary when using the "measure all the harmonics and do the maths" approach which is what Alfredo uses. Oh well, I guess it will all become clear one day!

I was impressed when I set up some mild distortion (diode across the oscillator output), measured it on my hardware machine and got 0.6%, then to try it on VA and got 0.597%. I'll see if I can repeat it before getting too excited!

Thanks for delving into the A cross B question. Again it will probably make an awful lot of sense when it's explained to us!

Thanks MelB. I am working on the next episode, the very important one of calibration. What's confusing me a bit at the moment is that I can seem to calibrate it fine on my testbench machine running XP, but not on my office/Lab machine running Windows 7. It's possible of course that Alfredo hasn't yet dealt with this function under the newer operating system. Or I might just be being particularly obtuse. Anyway, I'll play with it a bit more and then publish the temporary findings. If anyone has managed to calibrate a Windows 7 machine, and have the calibrations stick after closing and relaunching the program, do let me know. That will mean it is possible, and I just have to work harder!

Terry
 
Yeah, all that seems OK. I've just found that VA2011 calibrates fine on the Win7 machine, so I think it must be a bug. Fortunately, it's the downstairs test-bench machine I'm more interested in having calibrated, so I'll leave the office/lab machine for now and come back to it later. I'll be interested to find out how others with a Win7 machine fare.

Terry
 
OK got it. To have the calibration stick in Windows 7 with the March 2012 Beta:
1. Download and install the 2011 with installer version.
2. Download the March 2012 Beta and unzip it into the C:\Program Files (x86)\SillanumSoft\VA2011 directory. (ie. overwrite the VA.exe file with the new version)
3. Bob's your uncle.
 
Robert may well be your mother's brother, but my VA2012 is still uncalibrated, even trying MelB's good thinking!

The odd thing is that it isn't entirely uncalibrated. When I close and relaunch after calibration, and feed it the same signal level I had just used to calibrate it, it still reads volts (and not %fs). So it's trying. But it insists the incoming voltage is 1V (plus or minus a bit), instead of what it really is. How weird is that?

I get the same result if I reload a calibration. No matter what level I had used to calibrate with, VA now claims it's 1volt.

But this might be instructive: The Level of Known Input signal defaults to 1.0. On one channel I made it the value I had calibrated to previously, and then reloaded the calibration. On that channel, it now measured properly, on the other it claimed the incoming voltage was 1. So it's responding to what's in those boxes, even though I hadn't pressed the measure buttons. I can only think it's a bug, but since we can't seem to get to Alfredo, I guess we'll have to live with it.

Terry
 
Terry, now when I restart I have the check marks in "Apply calibration Left channel and Apply calibration Right channel" and the readings are balls on accurate. The remainder of the calibration box is greyed out and yes it does put the 1.00 back in "Level of known input signal" but it is greyed out and can not be changed as the check marks are indicating it is calibrated. The entire calibration box is in fact greyed out and does not allow any input.
Are you saving your calibration settings in the "Calibration" folder?
 
Yep, I'm seeing the calibration panel is all greyed out, and I'm saving in the calibration folder. I've even tried uninstalling VA, deleting the Sillanium folder, reinstalling 2011, calibrating it, relaunching it (perfect results!), overwriting 2011 VA.exe with 2012 VA.exe, loading the 2011 calibration I'd just done, applying it, wrong results!

Are you running Win7?

Terry
 
Here is the calibration file it made (opened with notepad):
[Calibrate]
TypeUnits=1
Units=0
VoltL=1 (I assume a 1 here means a tick in the "Apply calibration" box)
VoltR=1
Rcal=40.6248062849045
Lcal=40.6612768769264
Lval=2412464.39033864
Rval=2411623.03804972
UCcalL=0
UCcalR=0
INBoard=5
LevelL=32767
LevelR=32767
Name=Line 7/8 (M-Audio Delta 1010)
dBCalL=1.00
dBCalR=1.00
[Check]
Name Check=0
[ADC]
Ottobit=0
Sedicibit=0
Ventiquattrbit=1