From: Lang Elliott <>
>
> I've been doing lots of recording using my custom stereo Telinga MKH20
> setup. Go to the address below to listen to a recording I made last weekend.
> It features a Rose-breasted Grosbeak plus the sounds of some other animals,
> big critters of some sort. Maybe someone in this group can tell me what
> creature so rudely interrupted my grosbeak recording. Be sure to listen
> using headphones so that you can fully appreciate the marvelous stereo
> effect:
>
> http://www.naturesound.com/mp3/rbgroby.mp3
Where the big creatures white with black spots? Not quite sure on the
particular vocalization but that's the variant that comes to mind.
To me the grosbeak is not only slightly right, but also up above the
plane of the big beasts.
> Studying the waveforms of my various stereo parabolic recordings on the
> computer, I've discovered the following important fact. Unless the target
> bird is exactly focused, with the bird's song being identical amplitude in
> both channels, there is often a resultant phase difference between channels.
> This occurs even when the target bird is "slightly" off-axis.
Is not one of the ways we distinguish a stereo field phase differences
between channels (or at our ears if we are just listening)? Sounds like
you have been able to pick out the phase differences that belong. I
would doubt that you are picking out the reflected vs direct phase
differences as those are mixed by each mic.
> This can cause major problems when one desires a monaural recording and sums
> the two channels to obtain it. Phase cancellations then reduce the amplitude
> of the target bird's song, and may also result in other anomalies. For this
> reason, it is not a good idea to mix the two channels to monaural. It is
> better to choose the channel where the target bird is the loudest and use it
> as your monaural source.
>
> This negates my earlier conclusion that the stereo parabolic setup produces
> a monaural recording superior to that of a monaural parabola. This is true
> in theory, but only if the target bird is perfectly focused and there is no
> phase variation between the two channels. In actuality, it turns out that
> this is rarely the case, unless the target bird is in direct view and
> focusing becomes foolproof. When the bird is not seen and one focuses the
> parabola "by ear", it seems to be the norm that it will end up slightly off
> axis, with phase differences evident.
I'd not be so quick to make a blanket statement that stereo parabolics
cannot be mixed to mono well. Remember Rob's question about using M/S
with a parabolic. I'm quite sure M/S from a parabolic would mix
perfectly to mono.
However, as I noted then, two omni's become spaced omni's with a virtual
spacing larger than their real spacing. And phase difference has always
been a problem with spaced omni's when you try to mix them to mono.
A stereo parabolic setup produces a superior recording period. Mixing it
to mono makes it much worse, phase differences or not.
> It follows that if one chooses the loudest channel in a stereo parabolic
> recording to produce a monaural version, then this will produce a result
> that is somewhat inferior to that produced by a monaural parabolic setup. I
> haven't tested this in a controlled situation, but I expect that using only
> one channel will result in a loss of signal-to-noise of 3-5 db or so. Maybe
> one of you tech-heads can tell me what would be expected in theory.
I would hold that 3-5dB drop in signal up against the huge gain of the
reflector. It would be significant only at the lowest frequencies,
becoming less and less important as you go up the frequency scale.
> I expect that the Telinga Stereo DAT mike exhibits the same behavior as my
> custom setup. A quick preliminary test revealed that this is indeed the case
> (I noticed similar phase differences between channels).
The Telinga Stereo DAT mic is considerably different in design than your
setup. First off each channel has 4 mic capsules at 4 different
distances from the parabola. This means it's using a mixture of direct
and reflected distances. And it means a improvement in the mic self
noise over a single capsule. Second, those capsules are arranged PZM
style and use the boundary effect with the central barrier, not just use
it as a barrier. That will result in 4dB extra gain before the mic's
self noise. Probably also has some effect on phase differences too.
That they exhibit phase differences you find by the same methods
indicate that you are probably finding phase differences that belong.
> In spite of this potential "problem", I am totally sold on stereo parabolic
> recording. If you listen to the mp3 I've posted, using headphones, then
> you'll hear why. The listening experience is far more pleasurable that one
> has with mono. Off axis sounds come from their proper directions and the
> sense of spaciousness is superb. Listening at home becomes a real joy, and
> monitoring while recording in the field is much more exciting.
>
I've been trying to say this for years. Why I only record stereo. And
really don't worry too much about the mono.
I've one thing you might try. That's cutting off the ambient component
alltogether. And maybe getting a little boundary effect going. Try
placing a disk of your barrier material on the grids of the mics,
perpendicular to the axis of the dish. With the right size you should
cut off the ambient pickup and then get the pure reflected sound. I'm
not sure on disk size, probably not very large needed. By placing it
direct on the grids of the mic pair you will get a gap as close as you
can get to the correct PZM gap without disassembling the mics, though
it's still way too large. Worth experimenting with the distance. I
thought about trying this when you first exhibited your design, but have
not had the time to build it properly. I think you know how I build things.
At the very least such a design would get a different "view" from the
mixed ambient and reflected you have now. And different from the
barrierless spaced omni's others are using.
I've thought of making a cap for the Telinga DAT stereo element to do
the same thing. My original homemade parabolic had a barrier, and I
first experimented with and without. In mono the barrier improved the
clarity of the on axis sound. Or at least that's what it did for my old
ears. (I passed the big 60 this week) In that case, because the mic was
well outside the front plane of the dish, the barrier was a small
plastic funnel.
I've some other possibles, but they require some more elaborate
machining. Probably more than you want to tackle.
Walt
________________________________________________________________________
________________________________________________________________________
|