Subject:Waveforms...?

Posted by: Zoner

Date:1/3/2001 2:25:00 AM

While I'm here, I may as well mention that I noticed when I
recorded my voice, the actual waveform seems out of balance.

What I mean is...there is a center line, and the wave form
interpolates from + to - and on and on up and down....

Well, the negative values of the waveform extend further
than the positive values.

I'm just wondering if this is normal. It makes it a kind of
pain in the ass to normalize the signal when it is so out
of balance since normalizing will only increse the peak
value to the level you specify.

And when I say out of balance I mean that the negative
values on an average are about double the amplitude of the
positive values.

I'm too new to processing audio to understand if this is
normal or not. Any help would be great.

Zoner.

Subject:Re: Waveforms...?

Reply by: RickZ

Date:1/3/2001 6:54:00 AM

Try using 'tools'/'statistics' from SF menu on the file in
question, and see if there is a DC Offset listed. If there
is, you may have a problem with the audio input hardware.
Many sound cards have DC blocking capacitors on their
inputs that should prevent such an offset in the analog
domain. I'm writing this at work, and don't have SF here
to check, but I think there is a function to remove a DC
Offset.

Hope this helps,
Rick Z

Edward Kerrie wrote:
>>While I'm here, I may as well mention that I noticed when
I
>>recorded my voice, the actual waveform seems out of
balance.
>>
>>What I mean is...there is a center line, and the wave
form
>>interpolates from + to - and on and on up and down....
>>
>>Well, the negative values of the waveform extend further
>>than the positive values.
>>
>>I'm just wondering if this is normal. It makes it a kind
of
>>pain in the ass to normalize the signal when it is so out
>>of balance since normalizing will only increse the peak
>>value to the level you specify.
>>
>>And when I say out of balance I mean that the negative
>>values on an average are about double the amplitude of
the
>>positive values.
>>
>>I'm too new to processing audio to understand if this is
>>normal or not. Any help would be great.
>>
>>Zoner.
>>

Subject:Re: Waveforms...?

Reply by: Zoner

Date:1/3/2001 2:35:00 PM

I tried checking the DC offset. On both left and right
channel the offset was 0 %.

I have been recording more vocal takes and I have analysed
the waveforms. They are definately double the amplitude on
the bottom compared to the top of the waveform.

The waveform looks like a mirror in the centerline
reflecting itself top to bottom except that the bottom half
is literally stretched to double the amplitude.

Hope you can offer some more suggestions.

Zoner

Subject:Re: Waveforms...?

Reply by: Zoner

Date:1/3/2001 3:03:00 PM

Further info on this audio wavefile problem...

I have done some more recordings and this unbalanced
amplitude problem now seems to be somewhat intermittent.

On long vocal takes, there are many places where the
waveform seems to be fairly balanced as far as + and -
amplitude. In certain places however, the amplitude seems
to unbalance itself and cause the bottom half of the file
to extend to twice the amplitude as the top.

It's hard to explain. I am not just talking about a random
peak here. For instance on a whole sentence I might get
this problem whereas a new sentence right after the fact,
the waveform balances itself.

I find this very strange an I hope that someone can offer
any suggestions.

Zoner.

Subject:Re: Waveforms...?

Reply by: Rednroll

Date:1/4/2001 11:55:00 AM

Can you burn a CD? I would make yourself a test CD using
soundforge. By a test CD I mean creating Sine Waves at
different frequencies. Make sinewaves with the same
amplitude say -12dB, 30 seconds in length at
60Hz,250Hz,1kHz,4Khz,7Khz,10Khz,16Khz. These frequencies
should cover the range of your vocals. Play these test
sinewaves over a CD player and Mic the speaker and record
them back into sound forge. You can visually tell after
you record them back in, if there is a DC offset and if it
is frequency dependant. Sinewaves are much easier to look
at and analyze than a random voice wave. Also Do the same
test and this time remove the microphone and record
directly from the output of the CD player into the line in
of your sound card. See if you get the same results (ie
you just illiminated the microphone as being part of the
problem). I suspect that your microphone might actually be
the problem, probably something to do with the diaphram
element being restricted in full movement, in the positive
direction, or if it's a condenser mic it could be in the
amplification section.

There's my electrical engineering advice, good luck,
Brian Franz

Edward Kerrie wrote:
>>Further info on this audio wavefile problem...
>>
>>I have done some more recordings and this unbalanced
>>amplitude problem now seems to be somewhat intermittent.
>>
>>On long vocal takes, there are many places where the
>>waveform seems to be fairly balanced as far as + and -
>>amplitude. In certain places however, the amplitude seems
>>to unbalance itself and cause the bottom half of the file
>>to extend to twice the amplitude as the top.
>>
>>It's hard to explain. I am not just talking about a
random
>>peak here. For instance on a whole sentence I might get
>>this problem whereas a new sentence right after the fact,
>>the waveform balances itself.
>>
>>I find this very strange an I hope that someone can offer
>>any suggestions.
>>
>>Zoner.
>>
>>
>>

Subject:Re: Waveforms...?

Reply by: Zoner

Date:1/4/2001 1:35:00 PM

Yes. Other sources have also brought into suspect the
diapraghm of the microphone as well. i suspect this would
be the problem. It's a SM58 rental which I'm sure has seen
it's share of roughhousing.

I can not burn cd's. I'm pretty confident after someone
told me that the diaphraghm could be loose or damaged due
to dropping and such that that is what the problem most
likeley is.

Thanx for the help.