[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Walking down halls ...



A was written ...

>I have also noted degraded recordings from loudspeakers in enclosed
>environments.

This may not only be true of enclosed environments.


> When I record speech from a loudspeaker and then play this
>recording back to a listener on headphones I find it to be less
>intelligible than a person simply listening to the loudspeaker in the same
>environment.

Does the listener perceive that what they are listening to is the
recording of a loudspeaker? The headphones in this case become a 'second
generation' transducer (ignoring the micrphones).


> Even though I took considerable pains to calibrate and
>balance the recording and playback equipment.

Hmmm. From my experience, it is not possible to (seriously) calibrate the
frequency response of a loudspeaker. While taken with octave or one-third
octave, or even 1/12 octave equalization (meaning more than 96 filters ...
at some 60 dB/octave cut-off etc etc), the loudspeaker may appear "flat",
the examination of the frequency response characteristics of a loudspeaker
are quite different when measured at (eg) 1200 divisions to the
octave.IME, the peaks and dips can be several dB, spaced at under a few
Hertz.

This would mean that no loudspeaker is "flat", when examined with a
response curve calibrated in cents (or even 10s of cents). The
loudspeaker and the original 'cannot' sound the "same".


>Unfortunately I don't have a direct answer, but in my case I suspect it is
>related to imperfect recordings interacting with the acoustic environment.


And also possibly the frequency-depemdent directional characteristics of
the source. Which is ...

Have a person stand still, outside (free field) and speak. Walk in a circle
(at say 1 meter) around them. Listen to the _frequency response_
characteristics of the voice. Repeat with a (variety of) loudspeaker(s).

Depending upon the loudspeaker, the dispersion pattern will produce
greater and greater high frequency attenuation as one moves off-axis from
the main drivers. Also, with very rare exceptions, there is low frequency
sound coming through the walls of the loudspeaker to the sides and from
the rear. The sound heard at the back of the head is not added to by the
vibration of the skull ... in most cases.



>For example the interaction of reverberation,

Which leads to aspects of reverberation. The sound leaving the front of
the loudspeaker has different frequency response characteristics than the
sound that escapes through the walls of the enclosure. It is
(effectively) a 'complex' sound. The voice is (almost) a point-source
(in this generalized consideration). And, the speaker enclosure has added
colorations of its own, which will be added to / embedded into the
reverbertion.

(A piano is another example of a 'complex source'. Placing microphones at
different places (over the middle of the bass strings, over the middle
octave hammers, in the various holes in the frame, under the tail, under
the action, facing the (open) lid, a meter away from the lid ...)
reflects that (up close), the piano is not a 'single source', and is
anything _but_ a point source.)



> noise, and the
>directionality of the ears vs. microphones (in your case it may be related
>to the directionality of the loudspeakers vs. human-speakers).  I suspect
>that better recording equipment and perfecting the
>calibration/balancing/directionality of equipment might help and of course
>reducing the noise.  I would appreciate hearing about any answers you find
>through your research.


>On Wed, 29 Aug 2001, Michael S. Gordon wrote:

>> Dear List -

>>     I am interested in capturing the acoustic flow field of a person
walking >> down a hallway. Thus far I have made a couple of binaural
recordings of (1) >> a perceiver walking past individuals reading text
(live) and (2) a perceiver >> walking past a series of loudspeakers
projecting an assortment of recorded >> sounds (e.g., human speech,
>typewriter clicks, etc.). Those who have listened to these recordings
>have found the first set a much more compelling >> indication of the
>listener=B9s motion than the second.

I wonder ... How wide was the hall? How far away were the speakers and
the loudspeakers from the microphones? Would the results be tha same in
free-filed conditions? (out of doors?)

Is the same 'effect' present when the micropones remain still and the
sources move past the fixed mics? (These are also issues for the artistic
practice called "soundscaping".)

Best

Kevin
kaustin@vax2.concordia.ca