At 11:11 AM +1100 10/9/08, Paul Jacobson wrote:
>Hi Rob,
>
>The clip I posted had no post filtering, and no low cut applied on
>either mics or recorder. The only processing was conversion from
>48khz to 44.1khz. The 24bit/44.1khz wave file was converted to mp3
>using LAME 3.98 with -V 3 setting. I've replaced the mp3 with an
>version that has an additional 4dB gain to match the peak levels of
>Andrew's mp3. The rms level of my recording is still somewhat lower.
>
>Going back to the original file, the segment I posted had a peak
>level of -7.6dB, and an rms level of -34.6dB. I'd initially applied
>2dB of gain before converting to an mp3, and added a total of 6dB of
>gain to the mp3 that is currently posted. By applying 11dB
>"adjustment" to the mp3 the levels have been boosted by at least 13dB
>above the levels of the original file, and the loudest calls in the
>segment I posted would now be at +5.4dB FS.
>
>It sounds like Andrew's recording comes not just from a slightly
>different location and different day with different climatic
>conditions, but also from a different, higher intensity point in the
>dawn chorus. I've checked the time coding on the recording and
>worked out the segment I posted begins at 5.49am. This is getting
>close to the tail end of the dawn chorus, and has quite low
>intensity. I know Andrew left the camp site to record at around
>5.00am and judging by the intensity of calling in his recording it
>was made probably 20-30 minutes earlier than my recording, and far
>closer to the peak intensity of the dawn chorus. As a result I have
>concerns about just how valid the comparison is when it involves
>adjusting recordings of very different sound fields to a "matched"
>level.
>
>It is possible that Andrew and I will be able to make a "side by
>side" recording at some stage in the future, and the results would be
>a far more useful basis for comparison of these two rigs.
>
>I second Martyn's comments about Andrew being a highly skilled
>recordist. It was pretty clear from my conversations with Andrew that
>he is at a level where the technique is second nature, and that
>clearly gives him the freedom to focus on recording as an art. I had
>no intention of implying that the rig was more important than the
>recordist: as with all tools a skilled artisan will get better
>results than an apprentice. I definitely class myself as the latter.
>
>cheers
>Paul
>
>On Oct 9, 2008, at 2:51 AM, Rob Danielson wrote:
>
>> Paul, Andrew et al--
>>
>> Thanks for the interesting opportunity to compare rigs/mics. I
>> assembled a volume matched movie that might help with comparisons.
>> Paul's track is boosted 11 dB.
>>
>> Of course, the comparison is not as accurate as one could imagine.
>> The mic placements are not as close as possible, there are different
>> events and backgrounds on any two mornings and I was working from
>> mp3's that may or may not have been altered in some ways. That said,
>> the comparison is definitely interesting and worth further
>> exploration for a number of reasons.
>>
>> 6mb QuickTime movie (44.1K-16bit soundtrack)
>>
>><http://www.uwm.edu/~type/audio-reports/BoundaryMicExperiments/media/>htt=
p://www.uwm.edu/~type/audio-reports/BoundaryMicExperiments/media/
>> JacobsenPB-Vs_KnappSASS_Lrg.mov
>>
>> or
>>
>> <http://tinyurl.com/466aya>http://tinyurl.com/466aya
>>
>> Rob D.
>>
> > =3D =3D =3D =3D
Andrew and Paul--
In theory (and I'm quite sure, in reality) neither recording should
suffer any quality loss from their respective saturation levels in
the field. Sufficient saturation levels can sometimes help with mp3
encoding quality. For example, I believe that I detect some transient
distortion in both recordings that might be stemming from mp3
encoding,.. It might be better to also post shorter samples of
unaltered files if we want more accurate analysis.
The striking difference in "dynamics" between the two recordings
seems to arise partly from different mic to subject(s) distances but
I suspect that the stereo arrays used could also be playing important
roles. To my ears, when I play across the edit in the movie, there's
more lower mid-range "body" in the SASS recording. This part of the
spectrum is often very crucial in spatial imaging. There's also seems
to be more horizontal spread. Both qualities could also arise from
Andrew's rig being in a spot where the local acoustics are amplifying
this part of the spectrum, but I suspect this is not the full
explanation. Mic response and noise differences would also need to be
considered before anything definitive can be had.
I used sections from Andrews recording where the calls were softer
and I found that I had to use softer calls from your recording in
order to get the background ambience to be comparably audible. I
tried to match the backgrounds in the playback volume which are
usually more consistent than calls. But, the use of low-end roll-off
EQ can also affect this means of "matching." There could also be some
lower pitched, (250-450Hz) species calling in Andrews earlier in the
chorus that are not in yours,..?
Someday in the not so distant future, I'll have an outdoor, quiet
setting, mic "comparison environment" with accurate, repeatable
localization cues so recordists can set-up rigs for simultaneous
recordings and dive further into the mysteries. I think there could
be many factors we can better define and learn to apply towards
diffuse field stereo and surround micing techniques. Thanks for
planting some promising seeds. Rob D.
--
|