Rich Peet wrote:
> Actually what Rob describes, sounds like he recorded at to high of a
> gain and got the digital overload artifact. Sony Minidisc are not
> good when the pitch is high and you wanted the AGC to control the
> gain.
Yes, you can do that, but there is also a real component that's quite
common. And fits his description.
I never used the AGC on the Sony Minidisc I used to use, but always set
the gain manually. We have had a number of discussions on this subject.
Not so much for this reason but because it tends to react too slowly and
makes a recording that's very hard to process, it did respond to high
pitched sounds. On the MZ-R30 it's bar level meter also did respond to
high pitched calls and could be trusted, though it's response time was a
little slow to use it as a peak meter, so you had to pad for short
calls. I think what Rob is describing would be long sustained calls.
Which the AGC could handle if you did not mind it's messing up the start
of the call as it adjusted gain to match. Or the slow recovery of the
background afterwards.
> Actually because of the want of visualizations of "bug" calls I do
> not recommend minidisc and yes I like and use minidisc a lot. Many
> of these calls you will find to be broadranged sometimes as great as
> 4,000 through 20,000. It sounds fine to the ear when played back on
> minidisc but the compression will make the visualization look strange
> on the high freqs that are often chopped out.
You may want to note that this may be a function of a walkman MD and far
less the compression. The Portadisc does a very clean job with insect
calls. And I get to look long and hard at them by sonogram as I try to
filter them out to get the frogcall I'm after! Those broadranged calls
are the worst to try and filter.
I have strong doubts that a ideal insect recording system should stop at
20kHz. Though good mics that have higher range are quite expensive. And
going above 20kHz requires higher frequency sample rates. In fact I
don't recommend trying to do serious analysis on things above about
10kHz with a 44kHz sampling rate. Regardless of what recorder held the
samples. There are just too few samples to accurately describe the
waveforms. For this work in any detail, go to 96kHz or higher. In a way
it's amazing just how much info is available in the high end of the
44kHz sample when you look at the actual waveform of those samples, talk
about jaggies! And they play surprisingly well too!
> You also have to be careful on loud playing of these calls as you
> will not hear the the very high and loud when a medium volume low
> pitch is also present. This makes it possible on an amplified
> recording to actually play recordings that are dangerous to your
> hearing. If you want an example of this I can post a "bug" who
> sounds reasonable but when a filter is used to cut out all freqs
> below 15,000 it becomes painfully shrill.
Very good point. Though my hearing above 15,000 is so bad now I probably
would not find it painful. Certainly not loud.
> I also would recommend an omni mic or a hemispherical mic for this
> purpose. For a recorder I would recommend that you start with a
> cheap laptop (used) maybe a 233 mhz pentium. I also would just use
> the built in sound card of the laptop because most of what you will
> do will have a great signal to noise ratio anyway and you have time
> as these guys seldom fly away when calling.
To each their own.
If I remember right Klas mentioned that some use the telinga parabola as
a shield they place over the calling insect in the grass. It would
hardly have to be such a expensive shield, but might be a useful idea.
Any large, clear container might do.
I assume by hemispherical that you mean a cardioid pattern mic?
Otherwise, what is that? Gotta keep track of the toys.
Walt
________________________________________________________________________
________________________________________________________________________
|