Problems using Speaker Workshop Softwa
http://www.speakerworkshop.com/

This is the situation:
Speaker box is already built, I just wanted to do frequency response
measurement, using MLS. I've read through the tutorial twice and used the
help file extensively; but I'm still in limbo.

First problem is with line-in level, where I'm supposed to be getting 16K
but in the wave graph I only get 8 peak (NOT 8K; just 8 ). I'm on win98 se,
my sound card is an obscure Audigy OEM model similar to the Audigy LS, but
without the 96KHz adc, only the dac is 96/24. Mike is Earthworks M30BX and
have preamp with lots of gain. I know that the connections are okay because
if I raise the gain slider for line-in to output just above zero it
feedbacks. But I still don't get anywhere near 16K in record, neither using
1KHz nor MLS.

Another problem, perhaps a consequence of the first, is that I can't make
sense of the waves I see. With a 1 KHz test signal, for example, the output
is what I would expect it to be, but the input wave looks like low amplitude
for the first 0.15 seconds, then the amplitude doubles for the rest of the
second.
I tried to do the amplifier gain calibration but the wave that got recorded
looked like low level noise and then one big spike at 17 milliseconds, and
another big spike at about 21 milliseconds. How could my amp be doing that?

Somewhat similar problem, when I try MLS on my speaker sometimes I hear a
"double signal", meaning that I hear the amplitude fall, then return in the
span of possibly a few hundred milliseconds.

Last problem for now (since I have problems articulating the rest ) is that
I did close-miking tests using MLS and frequency sweep, and got totally
different frequency response curves. But then again, I probably don't even
have the levels right to begin with...

The calibrations I saw were for impedance measurement, except for mike
level; but somehow it doesn't seem to vary much when I change the record
level. Are there any other calibrations I need to do? The amplifier response
calibration I couldn't get through. The mike didn't come with a disk; only a
printed curve; and the help file had no example for how to write a file by
hand.

My play levels (in Creative mixer) a
Master volume: 100 (maxed out)
Bass/Treb: 50 (0 dB)
Wave/MP3: 100
MIDI: 100
CDaudio: 77
Line in: 0 (and muted)
CDdigital: 0 (and muted)

My record setting is at 100, and is set to mixing
Line/CD/Aux/TAD/PC (not sure what the last 3 are)

description:
I don't have a website to post pictures; so a few thousand words will have
to do... I'm using no jig, since I wasn't planning to do impedance tests;
and so I'm using a lot of alligator clips; except the speaker cables, of
course.

From the output of the SC (green jack) I mix the two channels through
resistors and go to my amplifier (very high input impedance). During tests I
hear the signals loud and clear.

From the mike I go to the input of the preamp (built it myself, using an
OP27). Output goes to Line-In (blue jack) between ground and the tip of the
plug (I hope this is the right channel, but I'm not sure; I don't have
stereo speakers; just one big monster). Should be; I know the signal is good
and that the SC is working because if I un-mute and raise the Line-In-Out
gain from the volume controls I get feedback oscillation.
Uhmm, maybe not; feedback would happen from the left channel, now that I
think about it; I guess I'll try the middle ring tonight.
When I select New-Signal- SineWave - 1KHz - 1 second, 0.8sec record, the
output graph shows a wave spanning -33K to +33K or so in amplitude; but the
input graph shows an amplitude of about +/- 8; not +/-16K.

I've lost all my settings since I lost power and had never saved anything;
but I remember a few things:
My SR I set to the maximum: 48 KHz.
Number of samples I tried 1K for 1m away measurements; 8K and 32K when
close-miking. The markers I placed at 3.33 mS and 24.66 ms for the 1 m away
test, since I figured it would take 3.33mS to reach the mike, plus 8K/48000
= 21.33 ms for the sample length; then in MLS I specified convert to
frequency response, using markers.

Thanks in advance

mark wrote:
Problems using Speaker Workshop Softwa
http://www.speakerworkshop.com/

snip
I've lost all my settings since I lost power and had never saved
anything; but I remember a few things:
My SR I set to the maximum: 48 KHz.
Number of samples I tried 1K for 1m away measurements; 8K and 32K when
close-miking. The markers I placed at 3.33 mS and 24.66 ms for the 1
m away test, since I figured it would take 3.33mS to reach the mike,
plus 8K/48000 = 21.33 ms for the sample length; then in MLS I
specified convert to frequency response, using markers.

Thanks in advance

I do not know that particular software, seems still in a beta-state.
But I can advise you one thing. To really get meaningful results you should
gain some understanding about what you are doing.
For example to suppress room reflections the window has to be set to
different values. The speed of sound is 344m/s, so the window must start
before 2.9ms. The first reflection( from the floor) arrives after
sqrt((2*h)^2 + d^2). h= distance to floor, d= distance to mike. So if you
want only direct sound you have to finish the window after 2.236m, given
both h and d are 1m. This would result in 6.0ms.

I found this book to be informative and easy to understand:
Joseph D'Appolito; Testing Loudspeakers; Audio Amateur Press; 1998

So to troubleshoot go logic:
1. Is there a noise signal heard when you are measuring? The MLS signal
should be quite loud and it sounds like the white noise when the television
channel shuts off.
2. Do your mike+preamp really record the signal? did you solder the adapters
right? On XLR Pin1 is GND. If you talk and record this with the windows
recorder, is it clear and undistorted? Or maybe the preamp is oscillating at
high frequencies, just try a TL071 instead of that OP27 which easily
oscillates.
3. Does the software work flawlessly together with your hardware and OS?
Maybe download another software. Something like Hobbybox, which I have been
using already 7yrs. ago. They have a 30days functional free trial version.
4. Do you operate the program allright? Follow the Help files accurately.
Read the Forum questions and answers, then only ask in the NG.

Mark, I really would like to help you to get meaningful results, but is a
little difficult from the distance without hands.on experience.

--
ciao Ban
Bordighera, Italy