>>> Here's a blind comparison movie created from the results of a moving
>>> pink noise test that is hopefully very similar to what Dan suggested.
>>> The (3mb .zip) download link is at the bottom of the blog post here:
>>> <http://tinyurl.com/2egg33s>http://tinyurl.com/2egg33s Play it
>> with QuickTime or a
>>> QuickTime-compatible player on your computer.
>>
>> Very nice, thank you, Rob.
>>
>> I've posted a Spectrafoo sonogram of the MP3 at:
>>
>> <http://www.dandugan.com/downloads/Miscellany/20100726_BoundaryCombFilte=
ring.png>http://www.dandugan.com/downloads/Miscellany/20100726_BoundaryComb=
Filtering.png
>>
>> and another with higher gain trying to show the highest frequencies bett=
er
>>
>> <http://www.dandugan.com/downloads/Miscellany/20100726_BoundaryCombFilte=
ring_2.png>http://www.dandugan.com/downloads/Miscellany/20100726_BoundaryCo=
mbFiltering_2.png
>>
>> The most audible effects I hear are a descending tonal effect in the
>> mid-range and the directivity of the mics at high frequencies (HF
>> comes up as pan comes on axis). Both happen on all three mountings.
>> I don't understand the source of the mid-range effect.
>>
>> Comb filtering effects at the top end that were my concern when I
>> proposed the test are more subtle than I expected. I found it useful
>> to high-pass the playback way up like 5k-8K to eliminate the
>> distractions of the lower effects.
>>
>> -Dan
>>
>
> Hi Dan--
>
> Both files I posted are full res (movie's soundtrack 16/48K and .wav
> 24/48K). So no mp3 artifacts should be involved.'
I was mistaken in saying MP3.
> I'm noticing a dropping chord with a fundamental around 407 Hz that
> changes to about 347 Hz in all samples.
Exactly.
> It sounds a bit like a
> Doppler effect but that doesn't seem possible as the speaker isn't
> getting closer and the velocities are so low. Could it be something
> like a difference tone created by sliding pitch changes high in the
> spectrum? Yet, that would be psychoacoustic and there may be a faint
> trace of the 407- 347 Hz drop in the sonograms.
I don't think so. Any reflective surface other than the ground nearby? Effe=
ct of the ground should be constant unless you actually changed elevation a=
s you moved...comb filter effects are very sensitive to small differences.
> Can one tell if the steep, downward sloped bands in the last 2/3rds
> of the A samples and the less steep upwards sloped bands in the last
> half of C samples is comb filtering?
That's how I interpret them.
> My older ears are pressed to
> detect this HF content . Rob D.
Me, too, after high-passing 12dB/oct at 8K I hear a little something just a=
t the end of #1, and swishing in no. 3.
-Dan
|