| I rendered the same piece that I had used in my earlier WinABX test, only
with reverb added. Both renderings used Csound 5, 64 bit sample version, at
ksmps=1. Orchestra and score were completely identical. The only difference
was that the first rendering was at 44,100 cycles per second to a 16 bit
stereo soundfile, and the second rendering was at 96,000 cycles per second
to a float stereo soundfile.
I was not confident that I could tell the difference at every point in the
soundfile, but there many sections were I could zero in on obvious
differences. As the test results below show, in those cases I could hear a
clear difference between the renderings _in every single guess_.
Both renderings were subjectively of crystal clarity, even in the highest
highs. Subjectively, perhaps the 32/96 rendering was clearer, i.e. less
muddy, in the midrange where most of the notes were happening.
The main difference is that the actual Csound instrument sounds were
different in the two pieces. In a number of cases there were notes that were
audibly louder in one rendering than in the other, or at a slightly
different pitch. I think the most reasonable explanation for this is the
difference in sampling rate.
Note that this test is not a pure test of difference in soundfile bandwidth.
Csound instruments evidently behave differently at different sampling rates.
But I think that this test is still a pretty convincing argument for
rendering at high bandwidth.
I will try this test again by making a 32/88200 soundfile, and then
converting it to a 16/44100 soundfile. I suspect that that difference will
be quite a bit harder to hear.
Regards,
Mike
WinABX v0.42 test report
09/28/2006 20:19:23
A file:
D:\utah\home\mkg\projects\csoundabx\f--2002-01-28--17-37-42.042.mml.1644.csd.wav
B file:
D:\utah\home\mkg\projects\csoundabx\f--2002-01-28--17-37-42.042.mml.3296.csd.wav
Start position 00:00.0, end position 03:59.3
20:30:44 1/1 p=50.0%
20:30:54 2/2 p=25.0%
20:31:08 3/3 p=12.5%
20:31:27 4/4 p=6.2%
20:31:39 5/5 p=3.1%
20:31:49 6/6 p=1.6%
20:32:08 reset
Start position 00:19.5, end position 00:26.3
20:34:08 1/1 p=50.0%
20:34:31 2/2 p=25.0%
20:34:47 3/3 p=12.5%
20:34:57 4/4 p=6.2%
20:35:08 5/5 p=3.1%
20:35:21 6/6 p=1.6%
20:35:33 reset
Start position 02:34.3, end position 02:39.1
20:41:41 1/1 p=50.0%
20:41:52 2/2 p=25.0%
20:42:13 3/3 p=12.5%
20:42:29 4/4 p=6.2%
20:42:37 5/5 p=3.1%
20:42:46 6/6 p=1.6%
20:42:53 7/7 p=0.8%
20:43:01 8/8 p=0.4%
Start position 02:52.9, end position 02:57.1
20:44:06 9/9 p=0.2%
20:44:18 10/10 p< 0.1%
20:44:23 11/11 p< 0.1%
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel |