atrac3+ audio quality
CygnusX-1
it when compared to 128k mp3? Also does the sound degrades badly
when converting 128k mp3 to atrac3+? TIA
Julian Adamaitis
this list normal stereo record mode is 192kbs or something, so it is
definitely higher quality. However, I am just of the opinion it is best NOT
to cascade several compressions on top of each other. I go from wav to
ATRAC or ATRAC to wav, but converting mp3 to ATRAC or Media player or such
can start to generate artifacts worse than those normally heard by the
ending format.
Now, I'm sure there are TONS of people on this list who frequently convert
ATRAC - MP3- ATRAC. I'll let them give you the dos and don'ts and problems
that can happen from these activities. I'd be curious to hear myself.
Julian
"CygnusX-1" <_cygnu...@excite.com> wrote in message
news:c7d6db20.04061...@posting.google.com...
SCC
copy is only as good as the original,
"Julian Adamaitis" <julianp...@access4less.net> wrote in message
news:10ct63h...@corp.supernews.com...
ff123
>ATRAC 3 Plus at 48 MBPS is equal if not superior to 128 KBPS MP#.
LOL! Yes, if you really meant MEGA bits per second.
But at 48 kbps, I think you've been sipping too much of the funny
stuff. Even at 64 kbps, the best codecs, for example high-efficiency
AAC, are badly beaten by lame mp3 at 128 kbps. 48 kbps is just not
enough to be able to approach CD quality.
ff123
SCC
compression so it's KBPS ratings can be of better quality. So maybe you
aren't informed or drinking the funny stuff. If you want a very detailed
explanation let me know.
"ff123" <ff...@noneya.org> wrote in message
news:gm61d0ls9l9hj0tnj...@4ax.com...
ff123
>You are comparing apples and pineapples. ATRAC is a different type of
>compression so it's KBPS ratings can be of better quality. So maybe you
>aren't informed or drinking the funny stuff. If you want a very detailed
>explanation let me know.
I'm not interested in any bogus explanations. Prove it to me.
If you're really serious, let's get down to brass tacks and set up a
real listening test. All I need from you are about half a dozen
samples about 20 seconds long, both the original, and atrac3+ version
at 48 kbps decoded back to WAV. You can losslessly compress these so
they take up less space. Be sure to tell me what the atrac3+ file
sizes were, so that we both agree that they were 48 kbps.
If you like, you can choose some of the lossless samples I store here:
That way you don't have to bother with ripping, editing, etc. All you
need to do is unFLAC and encode/decode back to WAV.
Then contact me by email and I will arrange for anonymous FTP upload
space for you.
I, in turn, will encode the originals with lame 128 kbps, and set up
the blind test. I can also include WMA9 Standard at 48 kbps to see
how it compares.
But really, 48 kbps of the best codec can't even touch a properly
encoded mp3 at 128 kbps!
ff123
miya...@eskimo.com
SCC
Second, I don't have time for all that because I have a life.
Third, I have already been through the process at a lab with several
engineers.
Maybe if you showed a little class when you debated I would take time to go
through this with you.
"ff123" <ff...@noneya.org> wrote in message
news:p4r1d0d45q22q1k72...@4ax.com...
ff123
>First of all there is no reason to be rude.
>
>Second, I don't have time for all that because I have a life.
>
>Third, I have already been through the process at a lab with several
>engineers.
>
>Maybe if you showed a little class when you debated I would take time to go
>through this with you.
Running away from the debate noted.
But really, if you've already performed the blind listening tests,
show me the data. I can show you data from all sorts of
latest-generation codecs blind tested over multiple samples by
multiple people at 64 kbps compared with 128 kbps mp3. They all got
blown away.
So don't make me laugh about 48 kbps codecs sounding as good as, let
alone better than, mp3 at 128 kbps.
ff123
Bob Cain
SCC wrote:
> You are comparing apples and pineapples. ATRAC is a different type of
> compression so it's KBPS ratings can be of better quality. So maybe you
> aren't informed or drinking the funny stuff. If you want a very detailed
> explanation let me know.
By all means.
Bob
--
"Things should be described as simply as possible, but no
simpler."
A. Einstein
SCC
ff123
>Not running away justasking you to clean up your act.
Non-response. Do you or don't you have blind listening tests showing
atrac3+ at 48 kbps equal to or superior to mp3 at 128 kbps? If so,
please post them. If not, I'd advise against posting BS in the
future. Somebody is likely to call you on it.
ff123
ff123
>First of all there is no reason to be rude.
>
>Second, I don't have time for all that because I have a life.
>
>Third, I have already been through the process at a lab with several
>engineers.
>
>Maybe if you showed a little class when you debated I would take time to go
>through this with you.
Ok, since apparently any results won't be forthcoming from you, here
are mine. I had a friend use SonicStage 2 to encode to atrac3+ at 48
kbps, and he then used Total Recorder to capture to WAV.
I tested 7 samples in blind tests against lame 3.90.3 --alt-preset 128
--scale 1
(The --scale 1 is used so that the mp3's don't have their volumes
reduced).
I used my blind listening utility (http://ff123.net/abchr/abchr.html)
to rate on a 1 to 5 scale, where 1 is worst, and 5 is best.
Some of these samples can be found on my site at
http://ff123.net/samples.html
bartok_strings2: atrac3+ = 2.0, lame = 5.0
dafunk: atrac3+ = 3.1, lame = 5.0
hongroise: atrac3+ = 2.3, lame = 5.0
Mahler: atrac3+ = 2.3, lame = 5.0
rosemary: atrac3+ = 3.6, lame = 5.0
trust: atrac3+ = 1.4, lame = 3.9
waiting: atrac3+ = 1.8, lame = 3.6
As you can see, atrac3+ at 48 kbps loses badly to lame at 128 kbps
every time. It's just no contest.
Typical atrac3+ artifacting at 48 kbps include noise-pumping (noise
increases during louder portions of the music), a metallic sound,
lowpassing, and reduction of stereo image.
I think it sounds respectable for 48 kbps, but nowhere near the
quality of lame mp3 at 128 kbps.
ff123
Chris
format, as are MP3, Ogg Vorbis, MP4/AAC, MPC, and many others.
Was I under the wrong impression about ATRAC3?
~ Chris
"SCC" <s...@nospam.com> wrote in message news:<GK3Ac.122438$Nn4.26...@twister.nyc.rr.com>...
CygnusX-1
this in a review? I ask this because a friend has an MD recorder and
says atrac3 (not atrac3+) is not that good at its lowest bit rate. Has
Sony improved atrac3+ that much over atrac3?
"SCC" <s...@nospam.com> wrote in message news:<1iPzc.80486$mX.27...@twister.nyc.rr.com>...
Reckless
I think he merely suggested what a great many people know already. ATRAC and
MP3 compression techniques are different and not exactly comparable. I
happen to agree with the statement of what sounds good. If you're an
audiophile then you wouldn't use any compression and you'd clearly 'need' to
use DVD-A at the highest bitrate possible...
"ff123" <ff...@noneya.org> wrote in message
news:6242d0l7191natn92...@4ax.com...
ff123
<reck...@nospam.com> wrote:
>Bogus, bullshit?
>
>I think he merely suggested what a great many people know already. ATRAC and
>MP3 compression techniques are different and not exactly comparable. I
>happen to agree with the statement of what sounds good. If you're an
>audiophile then you wouldn't use any compression and you'd clearly 'need' to
>use DVD-A at the highest bitrate possible...
Even though the artifacting between different formats are not the same
(i.e., I don't think I've ever heard noise-pumping using mp3), the
sheer size of the artifacting using atrac3+ at 48 kbps makes the
comparison with 128 kbps lame mp3 extremely easy.
It's not as if the differences were marginal, so that sometimes the
listener might have a hard time saying that one sounded better than
the other. The two are not even in the same ballpark; lame is
better-sounding by far.
What I said would be bogus is some "explanation" of why 48 kbps
atrac3+ sounds as good as or better than mp3 at 128 kbps instead of
the results of an actual listening test.
And considering the results of my personal listening test, and the
results of other listening tests which have pitted some of the finest
low-bitrate codecs at 64 kbps against lame mp3 at 128 kbps (all of
which lost miserably), I maintain that SCC's claim is indeed BS.
ff123
![[ice]'s profile photo](http://lh3.googleusercontent.com/a-/ALV-UjUqmXxi87oEDTkcROU8_n7cVOwqpDrzglDKUGqy4m340GxY=s40-c)
[ice]
> format, as are MP3, Ogg Vorbis, MP4/AAC, MPC, and many others.
>
> Was I under the wrong impression about ATRAC3?
Nope, I don't think so.
Please read: http://www.minidisc.org/hi-md_faq.html#_q99
Our friend ff123 with the double-blind listening fetish :-) may be
interested in the statement:
"The first generation ATRAC was a far cry from the quality of Type-R
DSP for ATRAC introduced in 1998, and in fact it was probably similar
in quality to MP3 at 128kbps. This view is supported by scientific
listening experiments conducted on Sony's behalf by Intertek Testing
Services (UK) and TESTFactory (Germany). Both reports showed a near
equivalency of ATRAC3plus at 64kbps with MP3 at 128kbps"
There's some nice charts and graphs with wiggly lines there too.
Incidentally, ATRAC3+ is only available on Hi-MD and Sony's Network
Walkmans at present.
ff123
>interested in the statement:
>
>"The first generation ATRAC was a far cry from the quality of Type-R
>DSP for ATRAC introduced in 1998, and in fact it was probably similar
>in quality to MP3 at 128kbps. This view is supported by scientific
>listening experiments conducted on Sony's behalf by Intertek Testing
>Services (UK) and TESTFactory (Germany). Both reports showed a near
>equivalency of ATRAC3plus at 64kbps with MP3 at 128kbps"
>
>There's some nice charts and graphs with wiggly lines there too.
I saw those reports. Here is Intertek's:
http://www.sony.net/Products/ATRAC3/tech/lab/ITS_test_report.pdf
and here is TESTfactory's:
http://www.sony.net/Products/ATRAC3/tech/lab/TESTfactory_Listening_test.pdf
They say that atrac3+ at 64 kbps is comparable or slightly preferable
to mp3 CBR 128 (using MMJB 7.5).
How would it have fared against lame 128? Here's how other
high-performance low-bitrate codecs at 64 kbps compared:
http://www.rjamorim.com/test/64test/results.html
Note that not even the best of them (HE-AAC) came close to lame 128.
Would atrac3+ at 64 kbps have done better? I don't know. That would
be an interesting listening test.
However, I *do* know that atrac3+ at 48 kbps (SCC's claim) certainly
does not come remotely close to lame 128.
ff123
Bob Cain
[ice] wrote:
>
> Please read: http://www.minidisc.org/hi-md_faq.html#_q99
>
That chart also states the 256 kbps ATRAC3 is perceptually
lossless. The feasability of that casts serious doubt on
everything else.
I'd like to believe it but can't.
StArSeEd
> [ice]:
> > Please read: http://www.minidisc.org/hi-md_faq.html#_q99
> That chart also states the 256 kbps ATRAC3 is perceptually
> lossless. The feasability of that casts serious doubt on
> everything else.
> I'd like to believe it but can't.
Ahoy again there, Mr. Cain.
I've been playing around a bit with ATRAC3plus @ 256k in SonicStage 2.0, converting both from
LAME v3.93.1 "--alt preset cbr 256" MP3 and PCM WAV, and the results are pretty frickin'
impressive so far. Very transparent to these ears. Extremely suitable as a replacement for
good ol' SP.
ATRAC3plus @ 48kbps seems about on par with, if not slightly better than, ATRAC3 @ 66kbps (LP4),
at least using the live (and nearly mono) soundboard Seven Mary Three song I used for
comparison.
Whoever complained a while back about SonicStage not being able to actually play back OMG files,
you're wrong... I've successfully converted multiple tracks to ATRAC3plus, moved/deleted/renamed
the original files (MP3 & WAV), and SonicStage still knows "what's up", playing the ATRAC3plus
OMG files without a hitch (though right-clicking the track and clicking Properties does yield an
interesting "error" message). Now, if there is no database entry for the ATRAC3/ATRAC3plus
file... things might not go so smoothly.
Something else that's possibly of interest: a file can apparently only exist in one ATRAC3 and
one ATRAC3plus bitrate at once (at least within SonicStage itself); you'll have to delete the
old OMG from a track's properties sheet (whether the OMG file itself still exists or not) in
order to re-convert it using a new bitrate. And yes, self-ripped waves can be converted
directly to ATRAC3/ATRAC3plus using SonicStage 2.0 without restrictions.
Hmm.. interesting little plain text xml tidbit here at the end of the 256k OMG file itself...
OMG:FRINGE:USR:user.xml
(garbage characters)
<?xml version = "1.0" encoding = "UTF-8" ?>
<package version = "2.0">
<pid>CENSORED</pid>
<album>
<caption>
<title >Untitled</title>
</caption>
</album>
<tracklist>
<track>
<description>
<genre>Unknown</genre>
</description>
<caption>
<title >Seven Mary Three - Goldfish</title>
<artist role = "principal" >Unknown</artist>
</caption>
<renditionlist>
<rendition format = "audio/mp3">
<cid>CENSORED</cid>
</rendition>
</renditionlist>
</track>
</tracklist>
</package>
Rendition format = audio/mp3? This was converted from a PCM wave. Nowhere on my hard drive do
I have, or have I ever had, this song in MP3. Hmm...
-StArSeEd
Martin Tillman
<arc...@arcanemethods.com> wrote:
> If you want a very detailed
> > explanation let me know.
>
> By all means.
HAHAHAHA!!
I DO so love your come-ons, Bob. Nice.
John Davis
From everything I've read, seen, heard and recorded, I have to say
which will sound better depends a lot on the source. Sometimes one,
sometimes the other. I've seen Atrac actually reduce noise on a
recording making it sound better.
--
John F Davis, WA8YXM in Delightful Detroit Remove "do.not.spam." to reply
"Nothing adds excitement like something that is none of your business"
John Davis
One thing to be wary of is testing done at a company's behalf. I've
seen far too many tests (I look closely at many medical and related
tests and have the skills and knowledge to understand a good part of it,
though I do admit not all)
I have found time and again that a lab has tested, and tested, and found
EXACTLY what the purchaser of the test wished them to find (Small
surprise) Even with larger, "independent" tests... Found differently.
Examples upon request.
ff123
<john.f.davis@.do.not.spam.sbcglobal.net> wrote:
> From everything I've read, seen, heard and recorded, I have to say
>which will sound better depends a lot on the source. Sometimes one,
>sometimes the other. I've seen Atrac actually reduce noise on a
>recording making it sound better.
The reduction of noise you're talking about may be related to the
noise pumping I heard. It may well be possible that atrac3+'s
artifacting can reduce surface noise of a recording from vinyl, for
example, and that may make it sound "better than the original." But
this will always lose to another encoding scheme which preserves the
fidelity of the original.
In the type of listening test I performed, whichever encoder produces
less annoying defects *when compared against the original* is the one
that will be rated higher.
ff123
ff123
>seen far too many tests (I look closely at many medical and related
>tests and have the skills and knowledge to understand a good part of it,
>though I do admit not all)
>
>I have found time and again that a lab has tested, and tested, and found
>EXACTLY what the purchaser of the test wished them to find (Small
>surprise) Even with larger, "independent" tests... Found differently.
>
>Examples upon request.
Here are a couple of examples from the audio codec world:
Microsoft commissioned NSTL -- they found that WMT4 bested RealSystem
G2 at 20 kbps and 32 kbps, and that WMT4 at 64 kbps was comparable to
mp3 at 128 kbps.
http://www.nstl.com/downloads/Final_MSAudio_Report.pdf
RealAudio also commissioned NSTL -- they found that Real Audio 8
bested WMA7 at stereo 20 kbps and at 32 kbps.
http://www.nstl.com/Downloads/real-vs-windows-media.pdf
There are a couple ways to "stack the deck" for audio codecs:
1. Choose your competitor carefully. In Microsoft's comparison, the
mp3 candidate was most likely a Xing encoder (MusicMatch Jukebox,
1999). Not exactly the best mp3 implementation.
2. Choose your samples carefully. If you know the strengths and
weaknesses of your own codec and that of your competitors, you can
pick and choose samples to tilt the comparison.
3. Choose your listeners carefully and listening system carefully.
In the Microsoft comparison, listeners were recruited from the local
community, and they listened through computer speakers.
Untrained listeners making a comparison through 2nd rate speakers
using "easy" music to encode might find that 64 kbps WMT4 sounds
comparable to the reviled Xing mp3 encoder at 128 kbps. But is that a
fair comparison?
ff123
Ma...@uwaterloo.ca
Dear Starseed,
I'm not sure I follow that last bit, but it sounds interesting.
Is "OMG" the file format the HiMD uses? If so, does that mean you can take
an OMG file, eg., recorded by line or microphone in, and convert them to
some nonproprietary format? That is what many people want when we say
"uploading" recorders.
I'm sure we'll find out in July, but maybe you've already figured out what
is in an "OMG" file.
Richard
Bob Cain
Ma...@uwaterloo.ca wrote:
> Is "OMG" the file format the HiMD uses? If so, does that mean you can take
> an OMG file, eg., recorded by line or microphone in, and convert them to
> some nonproprietary format? That is what many people want when we say
> "uploading" recorders.
>
> I'm sure we'll find out in July, but maybe you've already figured out what
> is in an "OMG" file.
>
Yeah, I'd like to know too.
ff123
<arc...@arcanemethods.com> wrote:
>
>
>Ma...@uwaterloo.ca wrote:
>
>> Is "OMG" the file format the HiMD uses? If so, does that mean you can take
>> an OMG file, eg., recorded by line or microphone in, and convert them to
>> some nonproprietary format? That is what many people want when we say
>> "uploading" recorders.
>>
>> I'm sure we'll find out in July, but maybe you've already figured out what
>> is in an "OMG" file.
>>
>
>Yeah, I'd like to know too.
>
>
>Bob
It apparently stands for Open MagicGate, Sony's DRM scheme. "Open" is
short for Sony's Open Platform Program, which means that the
information needed to produce products compatible with MagicGate
memory sticks is publicly available on a license basis (for an annual
fee, of course). That's a funny way to use "open." It sounds more
like MG is closed, as in proprietary; but maybe that's just me.
The license includes the rights to use atrac3 audio compression
technologies.
So a .omg file is an encrypted atrac3 file. But I don't know if it
can be converted to something like mp3.
ff123
[ice]
> you're wrong... I've successfully converted multiple tracks to ATRAC3plus, moved/deleted/renamed
> the original files (MP3 & WAV), and SonicStage still knows "what's up", playing the ATRAC3plus
> OMG files without a hitch (though right-clicking the track and clicking Properties does yield an
> interesting "error" message). Now, if there is no database entry for the ATRAC3/ATRAC3plus
> file... things might not go so smoothly.
Kewl, can you explain a bit more on how you played them (no matter how
simple)? SonicStage 2.0 right?
Weird, maybe there's no 'audio/wav' entry and an mp3 with no
compression is classed as a wav? Bizarre though.
Looks like a bit of file "meta-data" (data about the data) in XML
format for display in the SS's internal trackplayer or in the HiMD
device itself? Like an ID3 tag in an mp3.
The 'encoding = "UTF-8"' I'm not suprised by as it means many, many
character alphabets can be stored in this bit of XML (Latin, japanese,
Korean, Arabic etc).
John Davis
> In the type of listening test I performed, whichever encoder produces
> less annoying defects *when compared against the original* is the one
> that will be rated higher.
Well. I'd be inclined to say the very same thing FF. However I'm not
sure you and I would agree as to what "Defects" are "most annoying"
However.... I do not consider the reduction of noise (this was an off
air recording that I was using by the way, the noise was FM-hiss) to be
a defect.
John Davis
>
> There are a couple ways to "stack the deck" for audio codecs:
>
> 1. Choose your competitor carefully. In Microsoft's comparison, the
> mp3 candidate was most likely a Xing encoder (MusicMatch Jukebox,
> 1999). Not exactly the best mp3 implementation.
>
> 2. Choose your samples carefully. If you know the strengths and
> weaknesses of your own codec and that of your competitors, you can
> pick and choose samples to tilt the comparison.
>
> 3. Choose your listeners carefully and listening system carefully.
> In the Microsoft comparison, listeners were recruited from the local
> community, and they listened through computer speakers.
>
> Untrained listeners making a comparison through 2nd rate speakers
> using "easy" music to encode might find that 64 kbps WMT4 sounds
> comparable to the reviled Xing mp3 encoder at 128 kbps. But is that a
> fair comparison?
I agree. As I said, I'm more familure with Medical studies, but yours
is an even finer example (Same lab, different results, always finding
for the payer) Kind of reminds me of a 60 minutes spot on "lie
detectors"
60 minutes went to Popular photography and asked for voluenters, they
got toether a group of four then called in the polygraph operators.
They told all concerned "There was a very expensive camera stolen from
that closet (points to the Janitor's broom/mop closet door) and we have
narrowed the list of suspects down to these four, they were the only
ones with access. (They told the polygraph operator only "We think
"<suspect name>" did it, this part they did not tell their "Suspects")
Of course the did not store expensive cameras among the mops and brooms,
and not one polygraph operator openeed the door. But all five of the
polygraph operators they hired found the prime suspect had done it.
Even though there was no missing camera that was the part that was a
lie, the suspects simply told the truth ( I did not do it ) which, since
it had not been done, was true.
There is more. but that's it for this story.
StArSeEd
> Kewl, can you explain a bit more on how you played them (no matter how
> simple)? SonicStage 2.0 right?
Yes, I used SonicStage 2.0 to both create and play the OMG files. Even after renaming the OMG
file, removing the original MP3/WAV version from the OMG database (using SS2.0, not manually
editing it in MS Access.. "invisible" remnants MAY have been left behind in the .mdb itself..),
and moving/renaming/deleting the original file(s), the OMG remained quite playable in SonicStage
2.0. A simple drag-drop re-added the file to the database (only visible file format: OpenMG
ATRAC3plus), and it played perfectly.
> Weird, maybe there's no 'audio/wav' entry and an mp3 with no
> compression is classed as a wav? Bizarre though.
Probably just some botchery of MIME types for immediate compatibility where required... While
there IS an official MIME type for WAV files (audio/x-wav), and has been for many, many years,
apparently one may not exist for OMG/ATRAC3 files (at least at the time of the creation of SS2).
There is a draft concerning an interesting new "ATRAC-X" format, however:
http://www.ietf.org/internet-drafts/draft-hatanaka-avt-rtp-atrac-family-01.txt
> Looks like a bit of file "meta-data" (data about the data) in XML
> format for display in the SS's internal trackplayer or in the HiMD
> device itself? Like an ID3 tag in an mp3.
But of course.. I just found the MP3 bit interesting. And the pid & cid.. going to have to play
around with those a bit when I get a chance.
-StArSeEd
ff123
<john.f.davis@.do.not.spam.sbcglobal.net> wrote:
>ff123 wrote:
>> In the type of listening test I performed, whichever encoder produces
>> less annoying defects *when compared against the original* is the one
>> that will be rated higher.
>
>Well. I'd be inclined to say the very same thing FF. However I'm not
>sure you and I would agree as to what "Defects" are "most annoying"
Individual preferences vary. The most reliable type of listening test
has multiple listeners and uses as many samples of different types of
music as they can.
However, I don't want to lose sight of the fact that even though
preferences vary, at 48 kbps atrac3+ vs. 128 kbps lame mp3, we're not
really talking about preferring one defect over another. We're
talking about the presence of large defects in atrac3+'s case vs. the
virtual absence of other types of defects in lame mp3's case.
>However.... I do not consider the reduction of noise (this was an off
>air recording that I was using by the way, the noise was FM-hiss) to be
>a defect.
Let's say you use two codecs on the recording -- one reduces the hiss
and the other sounds exactly like the original, with the hiss intact.
Then the second one which preserves the fidelity "sounds better," as
in "closer to the original" in the type of listening test I used. The
first one would be considered to have artifacting.
I notice on some codecs, the same thing happens with distortion.
Sometimes it is reduced compared with the original. But that is a
defect by definition.
ff123
Bob Cain
John Davis wrote:
> Even though there was no missing camera that was the part that was a
> lie, the suspects simply told the truth ( I did not do it ) which, since
> it had not been done, was true.
>
> There is more. but that's it for this story.
>
There is also the possibility that that guy reacted as if
lying while telling the truth. Especially if he knew he was
the "culprit". I've had a polygraph report a lie while
telling the truth and continue to do so on several days of
testing. Just the fact that I knew the sensitivity of the
question was enough to make my parameters scream. It was
during a job interview for NSA almost 40 years ago and I was
disqualified (thank God.)
John Davis
>
>
> John Davis wrote:
>
>
>> Even though there was no missing camera that was the part that was a
>> lie, the suspects simply told the truth ( I did not do it ) which,
>> since it had not been done, was true.
>>
>> There is more. but that's it for this story.
>>
>
> There is also the possibility that that guy reacted as if lying while
> telling the truth. Especially if he knew he was the "culprit". I've
> had a polygraph report a lie while telling the truth and continue to do
> so on several days of testing. Just the fact that I knew the
> sensitivity of the question was enough to make my parameters scream. It
> was during a job interview for NSA almost 40 years ago and I was
> disqualified (thank God.)
There were 4 "suspects" and 5 Polygraph operators. Each operator (Save
one) was given a different "Favorite suspect" all fingered their
specific "Favorite" suspect as the lier even though all 4 "Suspects"
somply told the truth, a 100% failure rate, hard to write statisticts on
that cause you can't do a percentage when it's 0/5
I think I failed to point out that they randomized the "Favorite
Suspect" to avoid just the kind of posibility you cite.
Dave Kendall
> Any mags or sites review the audio quality of atrac3+? How good is
> it when compared to 128k mp3?
There's a comparison of ATRAC3 @132kbps with a number of other formats
here:
http://www.rjamorim.com/test/multiformat128/results.html
But so far I haven't seen any similar tests that compare ATRAC3+ with
anything else. From what I've heard ATRAC3+ is a significant
improvement, it's quite possible that it's closed the quality gap with
LAME MP3, at the very least I doubt it would still get beaten by WMA
Standard.
Also does the sound degrades badly
> when converting 128k mp3 to atrac3+? TIA
I haven't tested it, but it would probably damage the sound quality
significantly, converting between different formats and bitrates
usually does.
Dave
Lewis Levin
better sound quality regardless of codecs being compared. Variable
bit rate is better than constant bit rates. All of the successors to
MP3 (Lame, OGG, MP4) use some kind of acoustic modeling to decide what
to compress the most aggressively, as does ATRAC.
Vendors should acknowledge that it is impossible to win the codec
wars. People will figure out which is best. Eventually, it will be
very difficult to make further incremental improvements and increasing
bit rate will be the only way to improve quality.
Device manufacturers should simply make multiple codec devices.
Several do already--it is just not easy for consumers to "install" new
codecs in an existing device.
It is tempting to say that compression will become passe as storage
prices come down with increasing density. This is certainly true for
spinning magnetic media like hard disks. Spinning optical media,
whether CD or MD, are already quite cheap. The rate of progress in
solid state storage has been remarkably slow, given the market demand.
It is just hard to increase density, even with circuitry etched onto
silicon.
HI-MD fits two PCM uncompressed "CD's" on on HI-MD disc. Imagine if
lossless compression were applied. Perhaps 30-40% reduction (just a
guess--I should just try it to see) would be possible. It's
reasonable to suppose that in 4-6 years extreme compression won't be
needed. Even if lossless compression still results in large files,
everyone should run 392kb. What's a little more storage? The storage
problem is really a problem on fixed media devices (flash or hard
disk), but on removeable media devices like MD, storage already
doesn't even matter.