ceejay;176198 Wrote:
> Nah, can't be bothered. Its so obvious its not worth doing the test.
>
> Ceejay
>
> (Yes, I do have some high-bit rate MP3s which are for playing in
> portable devices where lossless would be pointless. But on my real gear
> the difference is pretty clear: granted, to tell the difference on
> toprated MP3s - 320kbps - you need to listen carefully to some
> non-trivial material, but the losses are very clear).
*crickets chirping*
Thought so. Do you know what it means if you can reliably distinguish
FLAC from 320kbps LAME MP3 in an A/B test but can't in an ABX test? It
means your brain is imagining differences that your ear is not actually
perceiving, and it's very common. Now I'm not suggesting you can't ABX
them--I'm just pointing out that you haven't. For what it's worth, I
can distinguish the two formats in an A/B test too, so the only
established difference so far between you and me on the perception
front is not "impairment" or a propensity to "BS", but that I don't
think A/B testing is as authoritative as you do. Incidentally, I do
recognize that there are "killer samples" that wreck even LAME MP3 at
320kbps. But what I'm talking about is representative samples
(randomly chosen from a real collection, for example). I can't ABX
them, and you haven't. So we're not really that different after
all--we're just missing a data point.
Please don't take this exchange as a challenge that you should actually
perform an ABX test yourself. It's truly a lot of time and trouble,
and, more importantly, don't ask questions if you're not prepared to
know the answer.
However, I AM taking this exchange as a cue that SlimServer could
really use an "ABX test" plugin. Not sure it's technically possible,
but wouldn't that be great--it would randomly create a sample set from
your FLAC collection, transcode it to the target format, and then
transcode it back to FLAC (so that you can't ID the format by looking
at the track info screen on the player), then create test sets with
TITLE tags like "Sample X - Trial Y - Number Z", and the test would
just be a playlist of those tracks. The user then does the whole test
from the player using the remote to go back and forth through the
tracks, and they write their responses on paper (can't imagine how
remote-control testing feedback would be possible). The paper
responses would end up being just a sequence of numbers identifying the
mismatched track throughout the playlist: 1,3,1,2,2,3,1 etc. The user
could then go back to the SlimServer, plug in those numbers, and
SlimServer could then spit out an ABX test results page.
This would certainly be worlds better than current audio format ABX
options: PC-ABX (PC-quality audio with the bonus of background fan
noise!), and homegrown options where you really need to know what
you're doing. With this plugin, SlimDevices products could actually
become the easiest way to do quality ABX testing of audio formats.
Maybe there could even be a multiformat test option.
The only downside is that this plugin would really spoil the
Transporter's budding reputation in the audiophile world ("Upon
reflection, nobody at Stereophile could even distinguish a 320kbps MP3
from a FLAC with any certainty using the Transporter's ABX plug-in, so
we've withdrawn our original review and beg the gentle reader's
forgiveness"). And I really don't want to do that to SlimDevices. So
maybe only allow the test to be run on Squeezeboxes. Or rig the test
for Transporters ;)
--
CatBus
------------------------------------------------------------------------
CatBus's Profile: http://forums.slimdevices.com/member.php?userid=7461
View this thread: http://forums.slimdevices.com/showthread.php?t=32138
_______________________________________________
ripping mailing list
[email protected]
http://lists.slimdevices.com/lists/listinfo/ripping
