Pete Carney wrote:

The first image shows the original at the location of the max peak in the
file. Then the second image is proper normalization in CoolEdit 2000 of the
exact same time frame. Then the third is what the command line program
"Normalize" does to the file.

With what parameters?
Normalize doesn't just do one thing and nothing else.

Definitely a bit of limiting going on there

Yes, "normalize" will do that if it must in order to obey the parameters
specified by the user. You can even use "normalize" to reduce the
amplitude of your file rather than boost it if you want.

but only at a couple of places in the whole file. The perceived volume
level of the properly normalized file and the After "Normalize"
file is significant in my opinion.

Significant in what way?

If better is having all your files RMS at about the same level, then this
file is definitely better.

Which file? You're working with three files there.

If anyone were to suggest that this "Normalize" file sounds "better"
I would have to call them "foolish" :-)

My main concern is to determine whether or not using "normalize" to
boost the amplitudes of low-RMS-level WAVs will cause fewer frequencies
in the recording to be axed on MP3's ATH chopping block as I believe
they would be in the WAV's default state for being "too low to be heard".

"Normalize" is not just a Unix command line program. It is available for
most platforms including Windows.

I use "normalize" under Linux. If "normalize" is also available for
Windows, great. So is "Mozilla". So is "The Gimp". So is StarOffice
and OpenOffice. Right on down the line. Makes no difference to me.

It is definitely doing some limiting where the peaks above 0 dB would clip.

Well, if it *has* to limit the peaks in order to meet a user-specified
demand, great. I'd rather it do that than clip 'em off. Meanwhile, in
my "normalized" WAV of MFSL's "Dark Side Of The Moon", there is no
limiting present at all even at the greatest peak. There is only
textbook normalization there. And for causing that to happen, I'm a
*heathen*. But Capitol's version's OK 'cuz everybody's heard of them.

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Bob Cain wrote:

Now, that's cool and certainly makes the point. How do you
make a GIF animation?

Well, y'know, I've already uploaded a screenshot of that maximum peak in
my Pink Floyd MFSL CD at

(1) it's original level
(2) at my +4.5dB personally preferred level and
(3) at a rudely pumped up +10dB level where the limiting actually
becomes visible.

Maybe I should animate those as well.

I just assumed that everybody could look up, look down and then look
straight ahead and see the differences - 'specially since they're lined
up perfectly straight one on top of the other.

Doh!

Myke


--

-================================-
Windows...It's rebootylicious!!!
-================================-

Geoff Wood wrote:

Foolish not to normalise maybe, but defintiely not to "normalise" - your
version which includes applying the compression that you seem to acknowledge
as a Bad Thing while similtaneiously explaining that you prefer it .

Duhhh..........................

Doh!

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Geoff Wood wrote:
"Martin Tillman" wrote in message

What? You are a troll.

However, if normalizing the WAV causes the amplitude of that
frequency to be boosted enough that it then becomes audible, it will be
retained.

Troll, troll, troll. I've spotted you, hehehehe!

The real sad thing is that I don't think he is .....

Truth really hurts, don't it?

Myke


--

-================================-
Windows...It's rebootylicious!!!
-================================-

StArSeEd wrote:

Alchemy Mindworks' GIF Construction Set is probably the most popular oldschool GIF animator.. at
least it used to be. Microsoft made one too at some point, though as I recall, it wasn't quite
as full-featured as the former.

No, no, no... All you need is The Gimp. It's like having Alchemy and
PhotoShop in one package - and it's free 'cuz it's open source. And
it's available for both Linux and Windows.

http://www.gimp.org

Get it today and never screw with another stupid commercial graphics app
unless your job requires you to have to do so. And you don't even have
to pay a dime to prevent those stupid annoying Shareware windows from
popping up every time you open it up - 'cuz they're never there to start
with... 'cuz it's FREE!

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Geoff Wood wrote:
"Yosah Akbah Muhammed" wrote in message

MP3 to 'archiving', is like chopping 10% off all your $ notes and putting
them in a draw, while throwing the rest away and calling it 'saving'.


Is that a Geoffact or just an opinion?

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

On Mon, 30 Jun 2003 03:22:54 -0500, Lord Hasenpfeffer
wrote:

Pete Carney wrote:

The first image shows the original at the location of the max peak in the
file. Then the second image is proper normalization in CoolEdit 2000 of the
exact same time frame. Then the third is what the command line program
"Normalize" does to the file.

With what parameters?
Normalize doesn't just do one thing and nothing else.

Weren't you claiming that this was *exactly* what it did with the MFSL
disc? That it just increased *every* sample until the peaks were just
below 0dB FS, with no other effect *whatever*?
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Mon, 30 Jun 2003 03:22:54 -0500, Lord Hasenpfeffer wrote:

Meanwhile, in my "normalized" WAV of MFSL's "Dark Side Of The Moon", there
is no limiting present at all even at the greatest peak. There is only
textbook normalization there.

Wrong. Explain your levels at 11 min and 24.5 min with respect to the
original.

And for causing that to happen, I'm a *heathen*.

Yep. Even moreso because you claimed adamantly that you were not doing
it (distorting the dynamic range).

But Capitol's version's OK 'cuz everybody's heard of them.

No one has said as far as I recall.

On Sun, 29 Jun 2003 18:37:05 -0500, Lord Hasenpfeffer wrote:


This is why I normalize before I encode. If you're just talking about
boosting amplitude for subsequent playback of the WAV from CD, fine.
There is no advantage to that

Hey, you've finally seen the light?

It's about ****ing time that you've finally realized that I understand
this

You certainly demonstrated that you absolutely did not believe this
earlier. You claimed your normalised DSOTM was better than the
original. Of course, what you did wasn't just normalisation...

and that CD-audio quality after normalization has nothing at all to
do with my hypothesis.

Previously...

Message-ID:

It amazes me that I can rip a WAV directly from an older commercial CD,
"normalize" it to 2dB beyond zero (i.e. -10dBFS) and then encode from it
an MP3 that sounds dramatically better than its own CD source.

Which means, if this is true:

If you're just talking about
boosting amplitude for subsequent playback of the WAV from CD, fine.
There is no advantage to that

meaning that this is the equivalent up twiddling the knob, then you're
saying, given the exact same source, but at different levels, you can
make two dramatically different mp3s!! This is just pure nonsense.

I suggest you restate exactly what you think you are doing and why,
because your story has changed several times over the last few days.

On Sun, 29 Jun 2003 02:02:09 -0500, Lord Hasenpfeffer wrote:

What about the Christopher Cross, "Another Page" CD I mentioned early on
in the other thread?

That "older, quieter" disc required a good 13.33dBs of pumpitude before
I was able to do anything reasonable with it.

And here are two MP3 samples in a zip archive:

http://www.mykec.com/mykec/audio/All_Right.zip

Eat that!

Pity we can't see the original, or that you didn't encode to a higher
bitrate. However, despite that, there is very little doubt in my mind
that your 'after' has limiting applied to it in order to make it sound
louder.

Yep, the only difference between them is that one is louder. Muppet.

All I had to do to make them sound the same was to add 2.5dB to the
quieter one (and let it clip in a few places). I could post a
compilation of the two and challenge you to spot the joins...

"Stewart Pinkerton" wrote in message


Which part of 'no effect on masking' did you fail to understand? The
threshold of hearing is dependent on sound levels, not on closeness to
0dB FS on the CD, IOW just crank up the amplifier a little.

Point of order...

The threshold of hearing as usually given *is* a sound level, and therefore
can't possibly be dependent on a sound level.

(Stewart Pinkerton) wrote in message ...

ATRAC is widely regarded as superior, but they both work in the same
basic way.

Despite that, good quality encoders such as Fraunhoffer and LAME are
clearly superior to latest versions of "standard" ATRAC. It's easy to
verify throwing at them some medium and hard-to-encode samples such as
the ones at PCABX pages.

On 1 Jul 2003 08:32:18 -0700, (KikeG) wrote:

(Stewart Pinkerton) wrote in message ...

ATRAC is widely regarded as superior, but they both work in the same
basic way.

Despite that, good quality encoders such as Fraunhoffer and LAME are
clearly superior to latest versions of "standard" ATRAC. It's easy to
verify throwing at them some medium and hard-to-encode samples such as
the ones at PCABX pages.

AAC and MP3 *are* Fraunhoffer algorithms................

Agreed that AAC is likely the best of the bunch.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Martin Tillman wrote:

Previously...

Message-ID:

It amazes me that I can rip a WAV directly from an older commercial CD,
"normalize" it to 2dB beyond zero (i.e. -10dBFS) and then encode from it
an MP3 that sounds dramatically better than its own CD source.

This is true. It does amaze me. Only now I know more about why this is
the case. The original CD source is "unnormalized" in my sense of the
term. The MP3 is made from a normalized WAV. In in general, most
people (including me) tend to believe that louder is better ... because
with loudness comes clarity.

This remains... an MP3 made from an older, quieter, unnormalized WAV
sounds poor to me compared to an MP3 made from a normalized one at the
same level of volume. I frequently listen to my MP3s in random shuffle
mode. Without "normalization", "remastered MP3s" sound are louder and
clearer sounding that "unremastered MP3s". Therefore, it is useful for
me to normalize older WAVs so that all of my MP3s have a nice, even
loudness. If I don't the older MP3s sound like crap in comparison the
newer "remastered" ones.

This has become a very boring conversation.

meaning that this is the equivalent up twiddling the knob, then you're
saying, given the exact same source, but at different levels, you can
make two dramatically different mp3s!! This is just pure nonsense.

Do as you like.

I suggest you restate exactly what you think you are doing and why,
because your story has changed several times over the last few days.

My mistake: Providing no well-defined hypothesis in this thread.
My correction: Providing a well-defined hypothesis in my other thread.

So it's been done.

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Arny Krueger wrote:
"Stewart Pinkerton" wrote in message



Which part of 'no effect on masking' did you fail to understand? The
threshold of hearing is dependent on sound levels, not on closeness to
0dB FS on the CD, IOW just crank up the amplifier a little.

Point of order...

The threshold of hearing as usually given *is* a sound level, and therefore
can't possibly be dependent on a sound level.

To clarify:

An ATH chart shows that frequency Y is "inaudible" below amplitude X.

In a given WAV, frequency Y's amplitude is X-2dB, therefore, frequency Y
is discarded and does not become a part of the final MP3.

However, after "normalizing" the WAV by a factor of +3dB, frequency Y's
amplitude now becomes X+1dB, therefore, frequency Y is retained and
becomes a part of the final MP3.

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

In article ,
Lord Hasenpfeffer wrote:


This is true. It does amaze me. Only now I know more about why this is
the case. The original CD source is "unnormalized" in my sense of the
term. The MP3 is made from a normalized WAV. In in general, most
people (including me) tend to believe that louder is better ... because
with loudness comes clarity.

This remains... an MP3 made from an older, quieter, unnormalized WAV
sounds poor to me compared to an MP3 made from a normalized one at the
same level of volume. I frequently listen to my MP3s in random shuffle
mode. Without "normalization", "remastered MP3s" sound are louder and
clearer sounding that "unremastered MP3s". Therefore, it is useful for
me to normalize older WAVs so that all of my MP3s have a nice, even
loudness. If I don't the older MP3s sound like crap in comparison the
newer "remastered" ones.

The real question, then, is "Where does the enhanced sense of clarity
come from?" Or, conversely, "Why do the MP3 encodings made from
lower-level CDs sound poorer to you, and why?"

I can see at least two possible hypotheses:

Hypothesis 1: the _relative_ amount of frequency content in the two
MP3 versions is identical, to within the inevitable low-level errors
implied by any coding system. Or, in other words, the lower-level
(duller-sounding-to-you) MP3 would sound identical to the higher-level
(clearer-sounding) MP3, if you simply turned up the volume a bit. The
differences in how you perceive the two aren't due to the actual
content of the MP3, only to the playback amplitude.

You might be perceiving the two as substantially different either due
to the well-known Fletcher-Munson effect: your ears' relative
sensitivities to bass and treble fall off faster than their
sensitivity to midrange, as the volume is reduced, and thus quieter
signals tend to sound as if they lack both bass and treble.

Or, it might be a masking and ambient-noise effect. If you typically
listen to MP3s in conditions of high ambient noise (office, car,
outdoors, etc.) then more of the lower-level-MP3 playback would be
drowned out by ambient noise, making it sound less clear.

The cure for this would be simply "Turn up the volume during playback."

Hypothesis 2: the MP3s encoded from lower-amplitude (non-normalized)
WAV inputs are actually, and significantly different than their
higher-amplitude cousins in ways other than just amplitude. Possibly
some frequencies are missing (excessive masking, or sounds below
threshold), possibly some frequency bands are less accurately encoded
and have a higher noise level.

This would suggest that the MP3 encoder you are using is less than
optimal. Possibly it has a poorly-set threshold detector, which is
sensitive only to the absolute signal level and not to the relative
levels. Possibly one of its other encoder or bit-allocation
algorithms is misbehaving, and is dedicating larger portions of the
available bit bandwidth to certain frequencies and is being forced to
discard other bands for lack of sufficient bit-reservoir in the
encoding.

The cure for this would be "Use a better encoder, or better settings
for the one that you have."

--
Dave Platt AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

An ATH chart shows that frequency Y is "inaudible" below amplitude X.

In a given WAV, frequency Y's amplitude is X-2dB, therefore, frequency Y
is discarded and does not become a part of the final MP3.

However, after "normalizing" the WAV by a factor of +3dB, frequency Y's
amplitude now becomes X+1dB, therefore, frequency Y is retained and
becomes a part of the final MP3.

AH.... I think I see the problem with your understanding of frequency
masking (as it was originally called).


The threshold of audibility is not a defacto amplitude. It is the amplitude
of a one frequency as compared to another frequency.

The codec works based on the concept that one loud tone will mask our
perception of another quieter tone which very close in frequency to the
louder tone.

If you have a tone of kHz at -5dB and another tone of 1.001KHz at -20dB,
that second tone may be removed from the signal since it is unlikely that
anyone could distinguish the presence of the quieter tone.

If the kHz tone goes away and the 1.001KHz tone is still at -20dB, the codec
will restore the 1.001KHz tone to the signal. as it will now become audible
since it is no longer masked by a louder, nearby tone.

If, however the second tone is several octaves away from the first, say
200Hz at -20dB, it will still be audible even though it is much quieter.
Since there is not another, louder tone nearby to mask it, will be audible.


- FLINT

Lord Hasenpfeffer wrote:

Arny Krueger wrote:
"Stewart Pinkerton" wrote in message



Which part of 'no effect on masking' did you fail to understand? The
threshold of hearing is dependent on sound levels, not on closeness to
0dB FS on the CD, IOW just crank up the amplifier a little.

Point of order...

The threshold of hearing as usually given *is* a sound level, and therefore
can't possibly be dependent on a sound level.

To clarify:

An ATH chart shows that frequency Y is "inaudible" below amplitude X.

In a given WAV, frequency Y's amplitude is X-2dB, therefore, frequency Y
is discarded and does not become a part of the final MP3.

However, after "normalizing" the WAV by a factor of +3dB, frequency Y's
amplitude now becomes X+1dB, therefore, frequency Y is retained and
becomes a part of the final MP3.

Have you checked with the Lame developers to see if this is
how the ATH is applied or if it is applied after the block
is normalized internally?


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

On Tue, 01 Jul 2003 20:48:05 -0500, Lord Hasenpfeffer
wrote:

Arny Krueger wrote:
"Stewart Pinkerton" wrote in message



Which part of 'no effect on masking' did you fail to understand? The
threshold of hearing is dependent on sound levels, not on closeness to
0dB FS on the CD, IOW just crank up the amplifier a little.

Point of order...

The threshold of hearing as usually given *is* a sound level, and therefore
can't possibly be dependent on a sound level.

To clarify:

An ATH chart shows that frequency Y is "inaudible" below amplitude X.

In a given WAV, frequency Y's amplitude is X-2dB, therefore, frequency Y
is discarded and does not become a part of the final MP3.

However, after "normalizing" the WAV by a factor of +3dB, frequency Y's
amplitude now becomes X+1dB, therefore, frequency Y is retained and
becomes a part of the final MP3.

We are well aware of what you're talking about, but there are two
problems with this:

1) The reference is to *sound level*, not amplitude on the CD. As
previously noted, just turn up the volume.

2) Have you checked that this is how your MP3 encoder actually works,
rather than on time and frequency-proximate masking (which is a
completely different matter)?

--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Tue, 01 Jul 2003 20:38:11 -0500, Lord Hasenpfeffer
wrote:

Martin Tillman wrote:

Previously...

Message-ID:

It amazes me that I can rip a WAV directly from an older commercial CD,
"normalize" it to 2dB beyond zero (i.e. -10dBFS) and then encode from it
an MP3 that sounds dramatically better than its own CD source.

This is true. It does amaze me. Only now I know more about why this is
the case. The original CD source is "unnormalized" in my sense of the
term. The MP3 is made from a normalized WAV. In in general, most
people (including me) tend to believe that louder is better ... because
with loudness comes clarity.

No, with loudness comes compression, and a *false* impression of
clarity given by the restricted dynamics of the reproduction. Radio
stations have been doing this for decades.


This remains... an MP3 made from an older, quieter, unnormalized WAV
sounds poor to me compared to an MP3 made from a normalized one at the
same level of volume. I frequently listen to my MP3s in random shuffle
mode. Without "normalization", "remastered MP3s" sound are louder and
clearer sounding that "unremastered MP3s". Therefore, it is useful for
me to normalize older WAVs so that all of my MP3s have a nice, even
loudness. If I don't the older MP3s sound like crap in comparison the
newer "remastered" ones.

This has become a very boring conversation.

This is true................

Basically, mix 'em up any way *you* like, but please refrain from
coming on here claiming that you've discovered some wonder product,
and 'whopped the ass of MFSL', 'cos both statements are *wayyy* wide
of the mark. What you're talking about has been known for decades, and
is just the sort of barbarity that MFSL deliberately tried to avoid.

--

Stewart Pinkerton | Music is Art - Audio is Engineering

(Stewart Pinkerton) wrote in message ...

AAC and MP3 *are* Fraunhoffer algorithms................

First, whilst one could agree that Fraunhoffer did develop the MP3
format and its first implementation on their own , I think that's not
the case of AAC format, where other companies participated too in the
development.

Second, I was talking about encoders, or compressors. Both a encoder
and a decoder and its algorithms do not need to share any code from
Fraunhoffer, they just need to conform to the standard. LAME encoder
is said not to have a single line of code written from Fraunhoffer
developers, and is considered to be superior to Fraunhoffer encoders
at bitrates over 128 kbps.

Agreed that AAC is likely the best of the bunch.

True. And ATRAC would be at the bottom.

Dave Platt wrote:

I can see at least two possible hypotheses:

Dave, I have drawn a picture. Please click the on the link provided
below and tell me which of your two hypotheses are best described by
what you see.

http://www.mykec.com/mykec/images/Good_MP3_Bad_MP3.gif

Myke

Hypothesis 1: the _relative_ amount of frequency content in the two
MP3 versions is identical, to within the inevitable low-level errors
implied by any coding system. Or, in other words, the lower-level
(duller-sounding-to-you) MP3 would sound identical to the higher-level
(clearer-sounding) MP3, if you simply turned up the volume a bit. The
differences in how you perceive the two aren't due to the actual
content of the MP3, only to the playback amplitude.

You might be perceiving the two as substantially different either due
to the well-known Fletcher-Munson effect: your ears' relative
sensitivities to bass and treble fall off faster than their
sensitivity to midrange, as the volume is reduced, and thus quieter
signals tend to sound as if they lack both bass and treble.

Or, it might be a masking and ambient-noise effect. If you typically
listen to MP3s in conditions of high ambient noise (office, car,
outdoors, etc.) then more of the lower-level-MP3 playback would be
drowned out by ambient noise, making it sound less clear.

The cure for this would be simply "Turn up the volume during playback."

Hypothesis 2: the MP3s encoded from lower-amplitude (non-normalized)
WAV inputs are actually, and significantly different than their
higher-amplitude cousins in ways other than just amplitude. Possibly
some frequencies are missing (excessive masking, or sounds below
threshold), possibly some frequency bands are less accurately encoded
and have a higher noise level.

This would suggest that the MP3 encoder you are using is less than
optimal. Possibly it has a poorly-set threshold detector, which is
sensitive only to the absolute signal level and not to the relative
levels. Possibly one of its other encoder or bit-allocation
algorithms is misbehaving, and is dedicating larger portions of the
available bit bandwidth to certain frequencies and is being forced to
discard other bands for lack of sufficient bit-reservoir in the
encoding.

The cure for this would be "Use a better encoder, or better settings
for the one that you have."



--

-================================-
Windows...It's rebootylicious!!!
-================================-

Stewart Pinkerton wrote:

To clarify:

An ATH chart shows that frequency Y is "inaudible" below amplitude X.

In a given WAV, frequency Y's amplitude is X-2dB, therefore, frequency Y
is discarded and does not become a part of the final MP3.

However, after "normalizing" the WAV by a factor of +3dB, frequency Y's
amplitude now becomes X+1dB, therefore, frequency Y is retained and
becomes a part of the final MP3.

We are well aware of what you're talking about, but there are two
problems with this:

1) The reference is to *sound level*, not amplitude on the CD. As
previously noted, just turn up the volume.

Where is the volume knob on my MP3 encoder?

2) Have you checked that this is how your MP3 encoder actually works,
rather than on time and frequency-proximate masking (which is a
completely different matter)?

I am absolutely certain that MP3 and ATRAC employ (1) masked frequency
and (2) Absolute Threshold of Hearing filtration methods.

Frequencies which are masked are removed AND frequencies which are too
quiet to be heard are removed.

My concerns lie NOT with masked frequencies. My purpose in normalizing
WAVs prior to encoding them is to assist them in successfully passing
through the A.T. of H. filter so that they end up as part of the final
MP3 rather than on the cutting room floor.

If the A.T. of H. filter prevents the under-amplified frequencies from
getting across and into the resultant MP3, you can playback that MP3 all
you want at whatever volume you want and those frequencies will still be
absent from the file.

Here's a pictu http://www.mykec.com/mykec/images/Good_MP3_Bad_MP3.gif

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Stewart Pinkerton wrote:

No, with loudness comes compression, and a *false* impression of
clarity given by the restricted dynamics of the reproduction. Radio
stations have been doing this for decades.

Radio stations have not been encoding MP3s for decades!

Basically, mix 'em up any way *you* like, but please refrain from
coming on here claiming that you've discovered some wonder product,
and 'whopped the ass of MFSL', 'cos both statements are *wayyy* wide
of the mark. What you're talking about has been known for decades, and
is just the sort of barbarity that MFSL deliberately tried to avoid.

http://www.mykec.com/mykec/images/Good_MP3_Bad_MP3.gif

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

"Lord Hasenpfeffer" wrote in message

My concern with creating better MP3s from louder WAVs lies not with
frequency masking but with filtration based upon what it known as the
Absolute Threshold of Hearing.

For the benefit of those not following the r.a.p thread. The (now) famous
"Absolute Threshold Of Hearing" relates to signals at extremely low levels
that are disgarded in MP3 and ATRAC coding. Not levels like -1dB, or -10dB,
or -20dB, but in the order of -60dB and lower.

geoff

Geoff Wood wrote:

Radio stations have been *restricting dynamics* for decades.

http://www.mykec.com/mykec/images/Su...Sunday_012.gif

0) Original WAV
1) "Normalize" (default -12dBFS)
2) "Normalize" (default -10dBFS)

Where is the evidence of this gawd-awful *restriction of dynamics* of
which you accuse me of wreaking against my WAVs?

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Geoff Wood wrote:

For the benefit of those not following the r.a.p thread. The (now) famous
"Absolute Threshold Of Hearing" relates to signals at extremely low levels
that are disgarded in MP3 and ATRAC coding. Not levels like -1dB, or -10dB,
or -20dB, but in the order of -60dB and lower.

Well, thank God, somebody finally cited some hard numbers as opposed to
hard opinions regarding the Absolute Threshold of Hearing! Sheesh!!!
What was so hard about that? And what is your source for this
information? Or are you just picking them out of the blue?

-60dB in relation to what? Full Scale?

Of course this value fluctuates higher and lower depending on frequency
value. It's definitely *not* a straight line even though I drew it to
look like one in my "Good_MP3_Bad_MP3.gif".

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

Geoff Wood wrote:

For the benefit of those not following the r.a.p thread. The (now) famous
"Absolute Threshold Of Hearing" relates to signals at extremely low levels
that are disgarded in MP3 and ATRAC coding. Not levels like -1dB, or -10dB,
or -20dB, but in the order of -60dB and lower.

It's nice to see you talking more on the same level with me here.
Maybe we're finally beginning to see eye-to-eye just a little bit
better. Reading your quote above sure beats having to read about
how no frequecies magically or disappear as a result of the
"normalization" - which I've never said nor even implied.

http://www.mykec.com/mykec/images/Good_MP3_Bad_MP3.gif

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

how no frequecies magically or disappear as a result of the

should have been:

"how no frequecies magically appear or disappear as a result of the"

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

"Lord Hasenpfeffer" wrote in message

Geoff Wood wrote:

For the benefit of those not following the r.a.p thread. The (now)
famous "Absolute Threshold Of Hearing" relates to signals at
extremely low levels that are disgarded in MP3 and ATRAC coding.
Not levels like -1dB, or -10dB, or -20dB, but in the order of -60dB
and lower.

Well not really but kinda-sorta.

The usually-given threshold of hearing is the "0 dB" line on charts like
this one:

http://www.webervst.com/fm.htm

However, due to masking the just noticeable levels for reliable perception
in normal day-to-day circumstances are way far higher.

Well, thank God, somebody finally cited some hard numbers as opposed
to hard opinions regarding the Absolute Threshold of Hearing!
Sheesh!!! What was so hard about that? And what is your source for
this information? Or are you just picking them out of the blue?

For a real thrill, search google for "Fletcher Munson Curve".

-60dB in relation to what? Full Scale?

Actually, the threshold of hearing is closer to 0 dB SPL @ 1000 Hz, and 0 dB
SPL is defined in terms of physical units relevant to acoustics (pressure
over area).

Of course this value fluctuates higher and lower depending on
frequency value.

Well doooh!

It's definitely *not* a straight line even though I
drew it to look like one in my "Good_MP3_Bad_MP3.gif".

I can't believe how much bandwidth and energy has been used up with this
tiring discussion when a few minutes of research with google would provide
most of the relevant facts.

"Lord Hasenpfeffer" wrote in message

KikeG wrote:

Agreed that AAC is likely the best of the bunch.

True. And ATRAC would be at the bottom.

You're implying that ATRAC is worse than MP3?

High bitrate MP3 can be really pretty good, expecially compared to low
bitrate ATRAC.

If you want to listen for yourself, please see:

http://www.pcabx.com/product/coder_decoder/index.htm

and

http://www.pcabx.com/product/mds-jb920/index.htm


You might want to start looking at this site starting with its home page at
www.pcabx.com .

"Lord Hasenpfeffer" wrote in message
...
Geoff Wood wrote:

For the benefit of those not following the r.a.p thread. The (now)
famous
"Absolute Threshold Of Hearing" relates to signals at extremely low
levels
that are disgarded in MP3 and ATRAC coding. Not levels like -1dB,
or -10dB,
or -20dB, but in the order of -60dB and lower.

Well, thank God, somebody finally cited some hard numbers as opposed to
hard opinions regarding the Absolute Threshold of Hearing! Sheesh!!!
What was so hard about that? And what is your source for this
information? Or are you just picking them out of the blue?

Out of the blue.

-60dB in relation to what? Full Scale?

Either FS or the wavfile peak level, which in most circumstance will be not
too dissimilar. As long as the peak isn't bizarrely low.

geoff

Arny Krueger wrote:

You might want to start looking at this site starting with its home page at
www.pcabx.com .

Thanks Arny. I've bookmarked all three pages. I'll look at 'em closer
ASAP.

It's beginning to look a lot like Usenet... the one I used to know...

Myke


--

-================================-
Windows...It's rebootylicious!!!
-================================-

On Wed, 02 Jul 2003 06:15:29 -0500, Lord Hasenpfeffer
wrote:

Arny Krueger wrote:

The usually-given threshold of hearing is the "0 dB" line on charts like
this one:

http://www.webervst.com/fm.htm

Ba-da-bing!! That's definitely right on track with what I've been
trying to find.

God heavens! Where did you look? Really?

This is so basic for most of us around here, that it is hard to
envision somebody not knowing it. You cannot open any book, however
basic, on hearing that does not start of with these facts!

Do you hate books that much, Myke?
If not, your public library could probably supply you with a book that
takes you thru this a lot faster than it takes to ask all these
questions on Usenet.

Per.

Lord Hasenpfeffer wrote in message ...

You're implying that ATRAC is worse than MP3?

Yes. And not only ATRAC3 long play modes at 132 kbps and below. Best
quality, standard 292 kbps ATRAC, is quite inferior to a LAME 3.90.2
200 kbps variable bitrate MP3, according to my blind tests: see first
and last post at http://www.hydrogenaudio.org/index.p...bb10a72975fb36

One of the ATRAC compression artifacts, the pre-echo of the castanets
sample, is blantantly evident. LAME 3.90.2, 3.90.3 or 3.91 with the
--alt-preset standard commandline are far better in this particular
sample, being transparent for many untrained people. Artifacts in the
other samples are more subtle, but LAME still does better in all of
them.

I'd dare to say that even a 192 kbps constant bitrate, joint-stereo
LAME mp3 is probably better than 292 kbps ATRAC.

Per Stromgren wrote:

Do you hate books that much, Myke?

I own a huge personal library of programming manuals persuant to my
purpose as a web-designer and Linux system/database administrator, etc..

Too many demands on my time for programming related needs throughout the
past 4 years have not left time for any other kinds of book-related
studies. My needs along these lines are very few and far-between.

If not, your public library could probably supply you with a book that
takes you thru this a lot faster than it takes to ask all these
questions on Usenet.

Well, the chart isn't the answer to everything I'm trying to find. It's
just a good, solid clue towards the solution - much closer and a *lot*
more helpful than all that other sometimes obvious stuff that's been
being discussed which while possible true is completely beside the point
and tangential at best.

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-

On Tue, 01 Jul 2003 20:38:11 -0500, Lord Hasenpfeffer wrote:

Martin Tillman wrote:

Previously...

Message-ID:

It amazes me that I can rip a WAV directly from an older commercial CD,
"normalize" it to 2dB beyond zero (i.e. -10dBFS) and then encode from it
an MP3 that sounds dramatically better than its own CD source.

This is true. It does amaze me. Only now I know more about why this is
the case. The original CD source is "unnormalized" in my sense of the
term. The MP3 is made from a normalized WAV.

Your definition of normalise is flawed, as is your practice of it.


In in general, most
people (including me) tend to believe that louder is better ... because
with loudness comes clarity.

Louder is just louder. Depending on the circumstances louder is either
better or worse than quieter. Therefore, 'louder is better' is actually
totally meaningless.


This remains... an MP3 made from an older, quieter, unnormalized WAV
sounds poor to me compared to an MP3 made from a normalized one at the
same level of volume.

I've proved you completely wrong with your Chris Cross example. Care to
listen to my proof?

I frequently listen to my MP3s in random shuffle
mode. Without "normalization", "remastered MP3s" sound are louder and
clearer sounding that "unremastered MP3s". Therefore, it is useful for
me to normalize older WAVs so that all of my MP3s have a nice, even
loudness. If I don't the older MP3s sound like crap in comparison the
newer "remastered" ones.

Faulty terminology, faulty logic, faulty use of the software.

Hint: There is nothing fundamentally wrong with true normalistion, nor
the desire to to have certain tracks sound more or less as loud as other
certain tracks, under certain circumstances. However, buggering the
dynamic range is totally wrong if you care about music, and even more
wrong when you don't realise you're doing it.

Hint 2: Do a search for 'Replay Gain'. If you can understand Replay
Gain you will be a better person. (Shock Horror! Replay Gain will
require that you normalise many of your tracks to around 6-10dB below
maximum!!!!!)

Lord Hasenpfeffer wrote:


Actually, flint, what you've just described *is* frequency masking which
is only one of 2 different filtration methods employed by lossy
compression methods such as MP3, AAC, MiniDisc, etc..

Just a minor terminology quibble, Myke. Filtration is what
is done to sewage. I'm an old E.E. with a DSP upgrade and
have never heard the word applied to signals. "Filtering"
is what you want.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

On Wed, 02 Jul 2003 04:46:35 -0500, Lord Hasenpfeffer wrote:

Geoff Wood wrote:

Radio stations have been *restricting dynamics* for decades.

http://www.mykec.com/mykec/images/Su...Sunday_012.gif

0) Original WAV
1) "Normalize" (default -12dBFS)
2) "Normalize" (default -10dBFS)

Where is the evidence of this gawd-awful *restriction of dynamics* of
which you accuse me of wreaking against my WAVs?

Right in front of your eyes! It shows absolutely that you are
compressing the original. If you can't see it you really don't know
what you are looking for.


Clues:

Left channel peaks at .55 and .58 mins retain relationship to lower
peaks. No compression.

Left channel peak just before 2.5. Look at the slightly lower peak just
before that one on top of waveform. See how it gets closer to the value
of the higher peak on the two higher level waveforms. What do you think
we could call this phenomenon?

See also grosser example on left channel lower waveform at 4.37 and
4.55.

Thanks for providing such a classic example of how wrong you are.

While there is no scale, I can guess what it is, and this also shows
that the highest peaks in the original are pretty close to full scale,
perhaps a couple of dBs down, indicating that there is precious little
room for REAL normalisation to maximum level anyway.

Bob Cain wrote:

Just a minor terminology quibble, Myke. Filtration is what
is done to sewage. I'm an old E.E. with a DSP upgrade and
have never heard the word applied to signals. "Filtering"
is what you want.

Hahahaha!!! As they say, "my bad!" :-D

Well here's what my gDict for Linux has to say about this issue:

Filtration Fil*tra"tion, n. Cf. F. filtration.
The act or process of filtering; the mechanical separation of
a liquid from the undissolved particles floating in it.

So, I guess we're both right!

The answer is clearly "yes and no"! ;-)

Myke

--

-================================-
Windows...It's rebootylicious!!!
-================================-