This thread:
http://forums.slimdevices.com/showthread.php?p=187486
Talks about difference testing. I rubbished it, although nowhere as
simply and elegantly as Opaqueice.
Adam
--
adamslim
SB3 into Derek Shek d2, Shanling CDT-100, Rotel RT-990BX, Esoteric Audio
Research 859, Living Voice
Mark Lanctot;187133 Wrote:
If the subtraction effects were overridden with time differences, the
effects would be random but 128 kbps clearly drops more material than
320 kbps in this case, as you would expect. In fact the 320 difference
test shows almost no differences in the guitar riff
opaqueice;187020 Wrote:
Actually this might be very very misleading.
You're subtracting two signals in the time domain. The resulting
signal is determined by all sorts of things which have nothing to do
with audible differences. For example, suppose the signals are
slightly mis-aligned
joatca;187132 Wrote:
I wondered if there might be a problem like that. To check it, I made up
another example. This time I've included 320Kbs, 128Kbs and VBR as well
as a 56Kbs MP3 of the original so you can hear approximately what it
sounds like (it's Wolfstone's Tinnie Run - 15 second
Mark Lanctot;187133 Wrote:
... In fact the 320 difference test shows almost no differences in the
guitar riff until the big cymbal crash. Interestingly, the cymbal
crash drops a lot of material for all tests.
When I first got my SB3, I did a lot of testing back and forth between
256k MP3,
jeffmeh;186850 Wrote:
Pretty neat. May I ask the artist and track you used to peform the
experiment?
It's Doppelkonzert F-Dur, Vivace by Telemann, track 2 of Concertos
for Recorder, Baroque Bassoon Strings, Pehrsson/McCraw/Drottningholm
Baroque Ensemble.
I was interested that the VBR
joatca;186898 Wrote:
It's Doppelkonzert F-Dur, Vivace by Telemann, track 2 of Concertos
for Recorder, Baroque Bassoon Strings, Pehrsson/McCraw/Drottningholm
Baroque Ensemble.
I was interested that the VBR differences were much greater than the
CBR. I suppose it makes sense, but until now
opaqueice;186917 Wrote:
I don't think it's as simple as that. MP3 uses a psychoacoustic model
to decide which spectral information to retain. VBR gives it the
additional freedom to concentrate on certain parts of the file, so it
might well allow the differences to be greater in parts of
I can reliably tell a difference between Fraunhofer encoded 256k and
FLAC. The highs are the giveaway, there is more resolution in the FLAC.
For a lot of music the difference doesn't matter, though.
--
pablolie
pablolie's
joatca;186844 Wrote:
More testing and I still can't tell which I prefer. However, I just
played with Audacity to subtract an encoded/decoded MP3 waveform from
the original WAV, so what's left should be the differences. I have two
30-second FLAC examples of the same piece, one the difference
Mark Lanctot;186981 Wrote:
That is VERY interesting.
[...snip...]
With only one 30-second section of classical music, this isn't
definitive, of course, but it's the first such comparison I've heard.
Thanks. I can't believe nobody did this before though. :-)
My tastes are mostly
joatca;186999 Wrote:
Note that the difference I posted is a naive one in the sense that
it's just one sample subtracted from another. If one is slightly more
rather than slightly less, you'll get a negative waveform in the
result, so the differences you hear here are probably far more
Mark Lanctot;187002 Wrote:
I'm also surprised no one's done this before. Still, I can't see why
this isn't accurate. You are comparing two waveforms, and by
subtracting one from the other, you get the difference between the two.
Actually this might be very very misleading.
You're
More testing and I still can't tell which I prefer. However, I just
played with Audacity to subtract an encoded/decoded MP3 waveform from
the original WAV, so what's left should be the differences. I have two
30-second FLAC examples of the same piece, one the difference between
320Kbps CBR and
joatca;186844 Wrote:
More testing and I still can't tell which I prefer. However, I just
played with Audacity to subtract an encoded/decoded MP3 waveform from
the original WAV, so what's left should be the differences. I have two
30-second FLAC examples of the same piece, one the difference
Mark Lanctot;185269 Wrote:
Oh yes I'm aware of HydrogenAudio. Can't post to it as I only use
webmail addresses though. :-(
But the tools they have there are for computer use. joatca's gets
SlimServer involved to some extent.
If it helps, I could generate a couple of files with my
I can't tell the difference between OGG Q6-7 and .Wav, therefore I
should't be able to tell a difference between OGG Q6-7 and .Flac.
I can spot differences up to 192kbps with MP3 quite well
--
livelock
livelock's Profile:
Eric Carroll;185070 Wrote:
Mark, are you aware of 'hydrogenaudio.org'
(http://www.hydrogenaudio.org/forums/index.php?act=idx)? There is lots
of activity and interest over there on DBT software tools to support
such.
There are freely available DBT tools for Unix and Windows currently
joatca;184813 Wrote:
I was inspired by this thread to try some blind testing of my own. For
those of you on Linux unable to test with the downloads mentioned
earlier I thought I'd post exactly what I did so you can repeat the
experiment for yourself with your own CDs.
The method is below,
Mark Lanctot;185067 Wrote:
I'm surprised there hasn't been more interest in this.
Mark, are you aware of 'hydrogenaudio.org'
(http://www.hydrogenaudio.org/forums/index.php?act=idx)? There is lots
of activity and interest over there on DBT software tools to support
such.
There are freely
I was inspired by this thread to try some blind testing of my own. For
those of you on Linux unable to test with the downloads mentioned
earlier I thought I'd post exactly what I did so you can repeat the
experiment for yourself with your own CDs.
The method is below, but my results were
SatoriGFX;179332 Wrote:
Blind testing proves little. Example, a recent study showed a group of
supposedly qualified people (classical musicians and so on) could not
tell the difference between the sound of a Strad and a far more run of
the mill violin. But, anyone who believes that they
opaqueice;179568 Wrote:
I don't follow your logic. They couldn't hear the difference, but your
faith tells you there must be one, and therefore blind tests don't
work? Wouldn't it be simpler to posit that there isn't a difference?
Do you really think a Strad sounds like any run of the mill
Pat Farrell;179027 Wrote:
agentsmith wrote:
Also, a clean track like Dire Straits Love Over Gold would be
very
nice
I'm not a great fan of using Dire Straits for critical comparison. The
tunes are good, and the recording is excellent, but their music has a
lot of distortion in
Craig;178940 Wrote:
Can you tell where it changes from mp3 to uncompressed and back?
I'll play:
starts as MP3
changes to FLAC at 0:25
changes to MP3 at 0:45
changes to FLAC at 1:14
changes to MP3 at 2:27
changes to FLAC at 4:00
changes to MP3 at 4:54
changes to FLAC at 5:59
--
cliveb
agentsmith;179051 Wrote:
Agreed in gerneral, but Love Over Gold is sort of an exception. It is
quite acoustic and used to be commonly used as a demo disc. Agreed
with classical piano. MP3 usually makes solo piano sounds grossly
distorted with sibilance. Glenn Gould's 1982 Goldberg
Pat Farrell;179027 Wrote:
agentsmith wrote:
Also, a clean track like Dire Straits Love Over Gold would be
very
nice
I'm not a great fan of using Dire Straits for critical comparison. The
tunes are good, and the recording is excellent, but their music has a
lot of distortion in
P Floding;179052 Wrote:
I think the older records, such as On Every Street sound much better.
Love Over Gold sounds very hifi IMHO. I did prefer Love Over Gold
though on my earlier systems. It's impressive for demoing.
On Every Street (1991) is much later than Love Over Gold (1982).
As it
cliveb;179063 Wrote:
On Every Street (1991) is much later than Love Over Gold (1982).
As it happens, Love Over Gold was recorded on 2 analogue tape at 30ips
without Dolby, so it ought to be ideal audiophile fodder.
OK. Didn't realize.
I do remember that the Love Over Gold CD was the first
There's a better way to do this sort of test, although I don't know a
way to use the SB rather than a sound card - download the ABXY plugin
for foobar. Then just convert a FLAC to mp3 and AB the two. The
plugin runs a nice blind test; you can see immediatly what you can hear
and what you can't.
opaqueice;179071 Wrote:
There's a better way to do this sort of test, although I don't know a
way to use the SB rather than a sound card - download the ABXY plugin
for foobar. Then just convert a FLAC to mp3 and AB the two. The
plugin runs a nice blind test; you can see immediatly what you
opaqueice;179076 Wrote:
What, exactly, is the problem with this approach? The OP just asked
if we can tell the difference between 128 kbs MP3 and FLAC. If you
can't, you can't, but there's no problem. Obviously high bitrate MP3
sounds good; sufficiently high bitrates are indistinguishable
P Floding;179077 Wrote:
So you can distinguish 128 kbps mp3 from 16/44?
Where is the bitrate limit when it sounds perfect? Does it start to
sound perfect, or does it happen suddenly at a certain bitrate? Is
there a sliding scale, or is it the same for everyone, you reckon?
I haven't
opaqueice;179085 Wrote:
but I'd be surprised if anyone can tell, say, 256 or 320 kbs MP3 from
lossless, on a normal musical selection.
I have no idea what you consider a normal musical selection but I
have compared high bitrate mp3's to FLAC and WAV and the difference is
obvious. Sure, for
No offense, but have you tried that blind? It's very easy - just
download foobar2000 and the ABXY plugin. Subjective impressions of
these things are simply not reliable - try it and you'll see for
yourself.
Anyway, it may be possible to distinguish on a good system - I
conducted my tests using
It's *no* problem at all to hear the difference on a good system - all
high frequencies are gone in a 128kbit MP3. That's if you can hear high
frequencies...
On a computer with standard computer speakers you have to listen for
artifacts introduced by the encoding, but that's as already mentioned
Craig;178940 Wrote:
I've taken a wav of Sky Blue by Peter Gabriel, made a copy and
converted the copy to a 128k mp3. Then I've compiled the two versions
back into a complete song.
What did you use to join the two files?
For some reason Audacity has been rebooting my PC when I open a
opaqueice;179085 Wrote:
I haven't experimented that much because I don't see any good reason not
to use lossless files for the SB. I did play with 192 VBR when I was
deciding which format to use with my ipod and found it hard to
distinguish from FLAC (so that's what I decided to use). My
Skunk;179109 Wrote:
What did you use to join the two files?
I use Steinberg Wavelab 6 for all my audio editing.
Craig
--
Craig
MC2Slim - Windows Shell and J River Media Center Integration for
Squeezebox.
http://www.duff-zapp.co.uk
P Floding;179126 Wrote:
It's pointless to compare one lossless bitrate to another as they might
just be equally bad. (Or, at least, if you can't hear a difference
between them, it proves nothing about higher bitrates or lossless.) I
had no problem at all hearing a difference between 224kbit
opaqueice;179157 Wrote:
Try it blind. It will take you less time than writing that post did.
And I'm not saying you won't be able to hear a difference - you might -
but it's amazing how afraid people seem to be that they might not. It's
not as if someone's going to come and take away your
opaqueice;179095 Wrote:
No offense, but have you tried that blind? It's very easy - just
download foobar2000 and the ABXY plugin. Subjective impressions of
these things are simply not reliable - try it and you'll see for
yourself.
Anyway, it may be possible to distinguish on a good
SatoriGFX;179203 Wrote:
I know all about the placebo effect and DBT etc... I have had many
experiences where I should have heard a positive difference (placebo
effect according to naysayers) but did not. I have bought newer, more
expensive products in an attempt to improve the sound only
opaqueice;179228 Wrote:
It seems to be a common misconception in these discussions that the
effects of psychology on preference, or hearing a difference versus
not, are predictable based on what you think your expectations are. In
other words people seem to think that when they buy
opaqueice;179095 Wrote:
No offense, but have you tried that blind? It's very easy - just
download foobar2000 and the ABXY plugin. Subjective impressions of
these things are simply not reliable - try it and you'll see for
yourself.
Anyway, it may be possible to distinguish on a good
P Floding;179053 Wrote:
The older Dire Straits albums sound surprisingly good on a very good
system Large, deep sound stage (real or not) and good dynamics (some
kick and crack in those drums).
I agree that there must be many much better recordings to use. I just
suggested something that
Well can you? its time to prove it :-)
I've taken a wav of Sky Blue by Peter Gabriel, made a copy and
converted the copy to a 128k mp3. Then I've compiled the two versions
back into a complete song.
Can you tell where it changes from mp3 to uncompressed and back?
A free bag of sweeties for the
The file is quite big. Could you compress it to mp3 :-)
--
Aswin
Aswin's Profile: http://forums.slimdevices.com/member.php?userid=9380
View this thread: http://forums.slimdevices.com/showthread.php?t=32576
Aswin;178942 Wrote:
The file is quite big. Could you compress it to mp3 :-)
OK, 2 bags of sweeties to the first person that can hear the joins in
the mp3 version :-)
Craig
--
Craig
MC2Slim - Windows Shell and J River Media Center Integration for
Squeezebox.
http://www.duff-zapp.co.uk
Craig;178940 Wrote:
Well can you? its time to prove it :-)
I've taken a wav of Sky Blue by Peter Gabriel, made a copy and
converted the copy to a 128k mp3. Then I've compiled the two versions
back into a complete song.
Can you tell where it changes from mp3 to uncompressed and back?
P Floding;178957 Wrote:
I'll give it a shot. Although I wouldn't go for Peter Gabriel for a test
like this. 25% downloaded...
It is a wonderfull track but if anyone has any other suggestions I can
do another one.
Craig
--
Craig
MC2Slim - Windows Shell and J River Media Center Integration
P Floding;178957 Wrote:
I'll give it a shot. Although I wouldn't go for Peter Gabriel for a test
like this. 25% downloaded...
Hmmm.. The whole thing sounds like mp3 IMO, due to Peter Gabriels taste
in mixing.
The sound did seem to collapse quite early on (I reliased I don't have
the time
mp3 to flac: 2min20
flac to mp3: 4min40
just about right or very wrong will do
--
Amauta
Amauta's Profile: http://forums.slimdevices.com/member.php?userid=2669
View this thread:
Peter Gabriel plays with phase effects on his own voice so it sound very
processed even when not mp3..
And then there are sections where just a couple of instruments play.
Put a transition in the beginning of such a section and it will be
impossible to pinpoint it.
No, sorry. He mucks about
P Floding;178963 Wrote:
The later on something happened during a shift in the music (which is a
bit like cheating, if that is where you placed the transition).
It's not cheating, it's about finding a point where I can make an edit
without a glitch
P Floding;178963 Wrote:
P.S:
Amauta;178966 Wrote:
mp3 to flac: 2min20
flac to mp3: 4min40
just about right or very wrong will do
Pretty near but you got the first one the wrong way around.
Craig
--
Craig
MC2Slim - Windows Shell and J River Media Center Integration for
Squeezebox.
http://www.duff-zapp.co.uk
Craig;178961 Wrote:
It is a wonderfull track but if anyone has any other suggestions I can
do another one.
Craig
Do you have some Dire Straits (On Every Street) or Joan Osborne
(Relish: Dracula Moon)?
Nice initiative, BTW!
--
P Floding
No, I didn't ABX it. And I won't even if you ask
P Floding;178974 Wrote:
Do you have some Dire Straits (On Every Street) or Joan Osborne
(Relish: Dracula Moon)?
Nice initiative, BTW!
Thanks
I have all Dire Straits up to Brothers in Arms and yes I have Dracula
Moon, I'll do another one over the weekend.
I'm just glad you didn't ask for
Craig;178976 Wrote:
Thanks
I have all Dire Straits up to Brothers in Arms and yes I have Dracula
Moon, I'll do another one over the weekend.
I'm just glad you didn't ask for Right Hand Man, that's bl**dy awfull
:-)
Craig
Joan Osborne would be my first choice then, when it comes to
P Floding;178977 Wrote:
Joan Osborne would be my first choice then, when it comes to sound
quality. I played that album today at live level, and it was spookily
good. Spider Web (about Ray) also sounds very good.
I'm just listening to it now, though not that loud - it's late here.
I'll do
Craig;178981 Wrote:
I'm just listening to it now, though not that loud - it's late here.
I'll do a mix of Dracula Moon over the weekend but you can provide the
sweets for the prize :-)
Craig
Interestingly Dracula Moon did not sound so good when I had the digital
out on 100%
Craig;178976 Wrote:
Thanks
I have all Dire Straits up to Brothers in Arms and yes I have Dracula
Moon, I'll do another one over the weekend.
I'm just glad you didn't ask for Right Hand Man, that's bl**dy awfull
:-)
Craig
This is an excellent idea!
But it'd be nice if you do not tell
agentsmith wrote:
Also, a clean track like Dire Straits Love Over Gold would be very
nice
I'm not a great fan of using Dire Straits for critical comparison. The
tunes are good, and the recording is excellent, but their music has a
lot of distortion in the guitars and synths in the rhythm,
63 matches
Mail list logo
