At 6:00 PM +0100 2/23/11, Matthias Eberl wrote:
>
>
>Finally the Sennheiser Stereo Set arrived and I
>was able to build two stereo rigs. I made a
>simple test setup in a quiet room, here is a
>audio slideshow showing the results:
>
><http://rufposten.de/daten/stereo_test_feb_2011/>http://rufposten.de/daten=
/stereo_test_feb_2011/
>
>Some thoughts:
>
>- When using speakers for playback, the PBBA
>delivers a more detailed imagination of the
>outer positions as the Olson Rig. When using
>headphones, the olson rig was more accurate.
>- The Olson Rig moves everything more to the
>center, which is audible especially in the direct
>comparison.
>- The recordings with the Olson Rig have more
>reverb, wich is probably increased because
>the room walls were parallel to the boundaries.
>So the room is more present, what I definitly
>like. This effect is overdone by a offset of 6cm (2.4") to the front tough=
.
>- Both rigs outperform the simple stereo setup of the R=F8de NT4 by far.
>
>That's all my untrained ear could hear. Would love to read your comments.
>
>(A discreet rumbling on some recordings stems from the passing tramway. )
>
>Matthias
>
Hi Matthias--
Interesting idea to compare these two rigs. I've
never seen a movie created that way with stills
and coding before. Pretty clever.
The samples on the movie that I'm listening to
seem have clipping or distortion from transient
peaks. I found it harder to ignore the
distortion clicks as I listened more times.
There are some differences very much worth
understanding. I tried making some quicker pans
using just 1mm and 6 cm samples only (the samples
with the most gain and boundary effect) but the
decay of the calls is cut off. Removing the tail
of the reverberations makes it hard for me to
evaluate the imaging, spatially. The room's
"live-ness" is adding tail.
A few observations :
(1) Of course, amplitudes across the stereo field
are different with different arrays. Normalizing
all the samples might help us hear spatial
differences-- and I like that possibility-- but
it does tend to mask the way the arrays create
amplitude differences off-axis. The levels at 180
degrees is different between the two arrays --an
important, fixed feature of the two arrays. Our
ears are probably more alert to differences in
amplitude that subtleties in room reflections.
For differences in spatial imaging to be
"significant," perceptually. the amplitude
differences should be factored in, not removed.
(I can appreciate what you tried to do though).
So, we usually calibrate the comparison by match
levels and left right balance using just the 12
o'clock position. The other levels are the
correctly relative.
(2) If you have performed a similar calibration
in the first section, the wave forms suggest to
me that the Olson wing array has really
interesting symmetry in the amplitudes across the
stereo field. See yellow highlighted area
http://tinyurl.com/4hsxuot). Look at how much
difference there is between 6cm and the shorter
setbacks. If you do repackage the test samples,
using just the 1mm and 6cm samples would make it
easier for use to evaluate. The long protrusion
distances of the mic for the PBB2N are
interesting to hear, but comparing these
variables might be clearer if placed in a
separate test.
(3) We end-up running some tests inside because
its so much easier to control things. We also
know that highly reflective interior spaces are
quite different from open, large, diffuse spaces.
The the ratio of direct to indirect sound is much
higher inside. To me, this means that the very
nature of the "reflections" one is using to test
performance inside are quite different from those
I'd ask my arrays to capture outside. .
(4) Usually, the most challenging positions for
localization performance are between 12 and 1:30.
Testing with closer positions across the center
like 12:00, 12:30; 1:00: 1:30 and 2:00 can often
reveal the most telling localization performance
differences.
Thanks for sharing your arrays and test with us.
There seem to be some interesting differences to
flesh out. I look forward to isolating and
explaining differences we can perceive. Rob D.
--
|