Listening comparison
 

Keith G wrote:
Inspired by the current visit from the Venerable Leader Of The New
World (listen to the backing vocals.... :-) I have posted a couple of
clips for preference testing by simply listening to them:

http://www.apah69.dsl.pipex.com/keit...harkey%201.mp3

http://www.apah69.dsl.pipex.com/keit...harkey%202.mp3

Anyone got a clear preference here? (Both tracks near identical @1min
/1.9 Mb)

Didn't like the first sample at all. Sounded compressed and way too bright.
Second sample, although still bright, was an easier listen and had a nicer
bass response.

--
MrBitsy

Listening comparison
 

"Keith G" wrote in message


Inspired by the current visit from the Venerable Leader Of The New
World (listen to the backing vocals.... :-) I have posted a couple of
clips for preference testing by simply listening to them:

http://www.apah69.dsl.pipex.com/keit...harkey%201.mp3

http://www.apah69.dsl.pipex.com/keit...harkey%202.mp3

Anyone got a clear preference here? (Both tracks near identical @1min
/1.9 Mb)

Two vastly different transcriptions of the same basic piece of music.

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without further
work, any preference test involving them is obviously invalid.

Looks like two different MP3 coders, or two different sets of MP3 coding
parameters were used. Again, any chance of a fair preference test is screwed
from the get-go by what looks like intentional biasing of the comparison
towards the second track.

Track 1 looks like it may have come from a CD, and was coded with a sharp
cutoff at 16 Khz.

Track 2 looks like it may have came from a LP, gentle roll-off above 16 KHz,
and with the typical up to 20 dB of extra noise below 25 Hz.

Both tracks are really tinny sounding, but Track 1 seems to be cleaner and a
little richer in the bass.

Listening comparison
 

"Keith G" wrote in message


Inspired by the current visit from the Venerable Leader Of The New
World (listen to the backing vocals.... :-) I have posted a couple of
clips for preference testing by simply listening to them:

http://www.apah69.dsl.pipex.com/keit...harkey%201.mp3

http://www.apah69.dsl.pipex.com/keit...harkey%202.mp3

Anyone got a clear preference here? (Both tracks near identical @1min
/1.9 Mb)

Two vastly different transcriptions of the same basic piece of music.

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without further
work, any preference test involving them is obviously invalid.

Looks like two different MP3 coders, or two different sets of MP3 coding
parameters were used. Again, any chance of a fair preference test is screwed
from the get-go by what looks like intentional biasing of the comparison
towards the second track.

Track 1 looks like it may have come from a CD, and was coded with a sharp
cutoff at 16 Khz.

Track 2 looks like it may have came from a LP, gentle roll-off above 16 KHz,
and with the typical up to 20 dB of extra noise below 25 Hz.

Both tracks are really tinny sounding, but Track 1 seems to be cleaner and a
little richer in the bass.

Listening comparison
 

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously invalid.

Any half-decent ABX software will do level matching automatically. I find
these samples to be not so badly matched anyway, replaygain attenuates at -
1.75dB against -1.86dB

I also think to say any preference test involving them is obviously invalid
is going a bit far, it depends what the aim of the test is.

--
Jim H jh
@333
.org

Listening comparison
 

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously invalid.

Any half-decent ABX software will do level matching automatically. I find
these samples to be not so badly matched anyway, replaygain attenuates at -
1.75dB against -1.86dB

I also think to say any preference test involving them is obviously invalid
is going a bit far, it depends what the aim of the test is.

--
Jim H jh
@333
.org

Listening comparison
 

"Jim H" wrote in message

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously
invalid.

Any half-decent ABX software will do level matching automatically.

I know of no ABX comparators that do automatic level matching. Please
provide an example.

I find these samples to be not so badly matched anyway, replaygain
attenuates at - 1.75dB against -1.86dB

There are dramatic differences in peak levels per Adobe Audition's
statistics tool:

Track one L -1.35 dB R -0.12 dB
Track two L -3.58 dB R -3.50 dB

Another tip-off that track one came from a CD, and track two came from a
LP. Track two is obviously rather highly compressed.

I also think to say any preference test involving them is obviously
invalid is going a bit far, it depends what the aim of the test is.

Listening comparison
 

"Jim H" wrote in message

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously
invalid.

Any half-decent ABX software will do level matching automatically.

I know of no ABX comparators that do automatic level matching. Please
provide an example.

I find these samples to be not so badly matched anyway, replaygain
attenuates at - 1.75dB against -1.86dB

There are dramatic differences in peak levels per Adobe Audition's
statistics tool:

Track one L -1.35 dB R -0.12 dB
Track two L -3.58 dB R -3.50 dB

Another tip-off that track one came from a CD, and track two came from a
LP. Track two is obviously rather highly compressed.

I also think to say any preference test involving them is obviously
invalid is going a bit far, it depends what the aim of the test is.

Listening comparison
 

more from the 'Arny Krueger school' of uk.rec.audio-ism:

"Jim H" wrote in message

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously
invalid.

Any half-decent ABX software will do level matching automatically.

I know of no ABX comparators that do automatic level matching. Please
provide an example.

I already mentioned higher up the thread (or lower if using outlook?), I
tend to use the FooBar2000 ABX component, a standard extension to the
player.

For the player:
http://www.foobar2000.org/

ABX components (included with 'special' version on downloads page)
http://www.foobar2000.org/components.html#foo_abx

The level matching is done via replay gain, which it can use for all
playback. see:
http://www.replaygain.org/

Hope this helps, the player runs nicely under WINE, if windoze ain't your
thing.
--
Jim H jh
@333
.org

Listening comparison
 

more from the 'Arny Krueger school' of uk.rec.audio-ism:

"Jim H" wrote in message

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously
invalid.

Any half-decent ABX software will do level matching automatically.

I know of no ABX comparators that do automatic level matching. Please
provide an example.

I already mentioned higher up the thread (or lower if using outlook?), I
tend to use the FooBar2000 ABX component, a standard extension to the
player.

For the player:
http://www.foobar2000.org/

ABX components (included with 'special' version on downloads page)
http://www.foobar2000.org/components.html#foo_abx

The level matching is done via replay gain, which it can use for all
playback. see:
http://www.replaygain.org/

Hope this helps, the player runs nicely under WINE, if windoze ain't your
thing.
--
Jim H jh
@333
.org

Listening comparison
 

"Jim H" wrote in message

more from the 'Arny Krueger school' of uk.rec.audio-ism:

"Jim H" wrote in message

more from the 'Arny Krueger school' of uk.rec.audio-ism:

The two recordings are about 50 mSec out of synch and very poorly
level-matched so they are easy to identify in an ABX test. Without
further work, any preference test involving them is obviously
invalid.

Any half-decent ABX software will do level matching automatically.

I know of no ABX comparators that do automatic level matching.
Please provide an example.

I already mentioned higher up the thread (or lower if using
outlook?), I tend to use the FooBar2000 ABX component, a standard
extension to the player.

For the player:
http://www.foobar2000.org/

ABX components (included with 'special' version on downloads page)
http://www.foobar2000.org/components.html#foo_abx

The level matching is done via replay gain, which it can use for all
playback. see:
http://www.replaygain.org/

Hope this helps, the player runs nicely under WINE, if windoze ain't
your thing.

Well, learning about new stuff is one reason why I participate in
newsgroups, and Foobar is a new one on me.

The idea of a ABX comparator that automatically changes the levels of the
files it compares is a bit aberrant, since the fact that the files have
different levels might be the point of the whole listening test. In this
case, I'd say that the exception does not break the rule, and the presence
of an ABX comparator that automatically matches levels in accordance with
some psychoacoustic theory does not eliminate or reduce the need for the
producer of files to perform their own level matching.

In this case even if Foobar's ABX and automatic level-matching features are
used, the remaining gross errors (e.g. timing) in Keith's sloppy work remain
and invalidate his proposed comparison.

There's a real problem that Keith wants to sweep under the table. He lacks
the foggiest notion of what a fair comparison involves, and lacks the
intellectual and technical tools required to set a fair comparison up. He's
been told many times, but is so arrogant and stupid that he thinks he knows
better.