[sdiy] Analyzing a Waveform

[Glen]

I recently recorded the sound of thunder. I used a dynamic mic, my computer
soundcard, and a pre-amp circuit that I was breadboarding and testing at
the time. My low-impedance microphone is on long cord, and placed outside
my room, on the back porch of the house.

My recording of ambient outdoor sounds turned out fine, except for a loud
thunderclap on the recording. I looked at the loud thunder waveform in an
audio editing program, and saw the following waveform:

http://webpages.charter.net/glenweb/diy/Transient_01.gif

You will notice that what should have been the largest peaks of the audio
look "clipped", but not in a way that I totally understand. Notice that the
"clipped" peaks of the audio are not "level" but ramp downward. What is
causing this "ramping" effect? 

I've drawn red marker lines on the waveform, and you will notice that not
all the clipped waves are of the same amplitude. This doesn't seem like the
typical clipping I've seen when an op-amp stage is driven too hard,  and
its output reaches its power rails. I'm used to seeing all the clipped
peaks having the same exact amplitude, and I'm not sure what it means when
some clipped waves are larger than others. What might cause this? 

I should mention that the WAV recording itself is not clipped. The height
of the white background area of the image represents the total dynamic
range of the WAV file itself. As you can see, the height of the waveform is
noticeably smaller than the white background, so the limits of the WAV file
itself haven't been reached.

It seems like I used to know the answers to these questions, but it's
apparently been a long time since I really had to use that particular
knowledge. I guess I haven't done that much design work over the years.
Please help me refresh my memory.


thanks,
Glen