Roger W. Norman wrote:
And I already mentioned digging up Stephen Paul and checking out his
methodology for making a mix mp3 ready. If you don't have the opportunity
to do it from this level you're simply ripping CDs and who gives a ****?
Let's not try to kid ourselves here. You aren't contributing anything to
the situation. If you find doing an overall normalization as if it were the
god of mp3 conversions is a good thing, then by all means do so. It sucks,
but if you think it's good, then it must suck good. Again, without any
contribution to the overall creativity of the product you're simply another
link in the chain that somehow wants to continually **** well recorded music
up. There is NEVER a time when LOSSY COMPRESSION IS BETTER, whether it's
loud or not.

From your professional perspective, all of the above is true.
From my professional perspective, MP3 is a requirement for my project.
Placing voluminous amounts of CD-quality audio online is not an option.
Who gives a ****? Anyone who wants their online presence to sound as
good as possible given the obvious inherent limitations of the medium
being employed.

I am seeking facts regarding how to produce better MP3s than Joe Sixpack
can make by himself at home. It will help to give my business "a
leg-up" on the local competition. If someone can come to me to get a
"better sounding" MP3 than what the people down the street can produce,
guess who they're probably going to come to.

End of tangent.

Myke

--

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

"Lord Hasenpfeffer" wrote in message

I am seeking facts regarding how to produce better MP3s than Joe Sixpack
can make by himself at home. It will help to give my business "a
leg-up" on the local competition. If someone can come to me to get a
"better sounding" MP3 than what the people down the street can produce,
guess who they're probably going to come to.

What you are failing to understand is that you are not trying to make
'better MP3s" - you are trying to make MP3s with an average RMS level that
appeals to you. Not necessarily the same thing. More discriminating
listeners may listen to this music and say "all this band's songs sound flat
and overcompressed".

The couple of dB shift that we are talking about has *no relevance* to your
precious "absolute threshold of hearing". That threshold relates to tiny
amounts of signals at very low levels that may be discarded in your encoder.
The effect that you find so pleasing (louder, compressed) is totally
unrelated to that parameter.

geoff

Geoff Wood wrote:

What you are failing to understand is that you are not trying to make
'better MP3s" - you are trying to make MP3s with an average RMS level that
appeals to you.

That's part of it, yes. I've admitted many times that I prefer louder
*pop/rock* MP3s over quieter ones - but I believe there's also more to
it than just that.

More discriminating listeners may listen to this music and say "all this
band's songs sound flat and overcompressed".

I have posted a link to this animation several times already and you've
yet to say a word about it. It contains 3 different screenshots which
clearly reveal the effects of "normalize" on a WAV at both the default
-12dBFS and my preferred -10dBFS levels. Clearly you can see about 5
extreme peaks being slightly limited in the loudest file but I hardly
believe that my level adjustment equates to any even remotely resembling
a "weapon of mass destruction" as you say it does.

Also note that this one track is the *loudest* track on the entire CD so
if I was to "batch normalize" all the WAVs from that CD by whatever
factor is required to make their average RMS level equal to -10dBFS,
what "damage" you see occurring in this one screenshot as a result of
the limiter is *all* the *damage* that would result throughout the
entire set - and a lot fewer frequencies would also end up on the
cutting room floor as a natural result of the WAV-to-MP3 encoding
process - if my hypothesis is correct.

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

The couple of dB shift that we are talking about has *no relevance* to your
precious "absolute threshold of hearing". That threshold relates to tiny
amounts of signals at very low levels that may be discarded in your encoder.
The effect that you find so pleasing (louder, compressed) is totally
unrelated to that parameter.

Do you deny the sense in doing all that can be done in order to achieve
the best result - given the amount of inherent destruction that MP3
compression already causes? Here's another picture for you. Go ahead
and take a peek. It'll do you no harm and it better illustrates my view.

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

Myke

--

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

"Lord Hasenpfeffer" wrote in message


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


Looking again, this picture illustrates clearly just how exaggerated your
notions are. Really - not sniping or exaggerating.

The A T of H has nothing to do with the 'better sound' you like. That is
purely the higher average level that you prefer and presumably acheive with
yo command-line app. Either that or you like clipping distortion.

The 'frequencies' that you refer to are extremely low level instances in a
frequency-weighted regimen that is designed to have minimal audible impact.
Stronger levels of those 'frequencies' are not disgarded.

geoff

Geoff Wood wrote:

Because is doesn't animate for me.

It does for me just fine. What's your browser?

Anyone else having trouble with this animation?

A track with a much higher ratio of loud/soft will be affected much more
than one already near your desired average.

Can you think of any songs right off the top of your head which I might
have in my collection which I could use for testing this?

For instance, someone mentioned earlier Joe Jackson, "Steppin' Out". As
soon as I can, I'm gonna rip-and-encode that track and see how my
encoder deals with it.

The couple of dB shift that we are talking about has *no relevance* to
your precious "absolute threshold of hearing". That threshold relates
to tiny amounts of signals at very low levels that may be discarded
in your encoder.

So if the ATH has such tiny effect, which element of lossy compression
is the one most responsible for so much destruction?

And what are the "very low levels" you cite? Or what webpage is there
that has these levels spelled out for one all to clearly see?

I don't deny your word that ATH affects "very low level" frequencies.
So before you go off saying you've presented me with facts and I've
simply ignored them, please tell me what they are or where I can find them.

"The sky is blue," is true. But how blue is it? Is it royal blue? Is
it navy blue? Is it eyeball blue? No, it's *sky* blue!

Do you deny the sense in doing all that can be done in order to achieve
the best result - given the amount of inherent destruction that MP3
compression already causes?

In terms of an MP3 no. But the effect, in the scenario you variously claim
(without compression), would be made equally by turning up your volume a
tad.

You keep telling me this but your presentation is such that you're just
ballparking to derive conclusions - which is no better than what I'm
doing - except I simply ballpark more weight on the effect of the ATH
filtration than you do. We could go back and forth like this all day.

It's more effective! No, it's less effective! No, it's more effective!
No, it's less effective! That gets us totally nowhere.

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

Indicates a thorough non-understanding of relative levels,

It's just a quick-and-dirty drawing to show the basic concept of my
hypothesis - not a scientific anaysis - because obviously there are
still people out there who think I'm discussing masking and not ATH
filtration.

and that how that relates to their physical appearance
(log v. linear).

Well then tell me more about that. Or point me to a website. Or else
I'm just going to go right back to doing all the legwork on my own as my
limited time permits. I'm not as stupid as Joe Sixpack but I'm not a
professional either. I've said this all along.

Hell, I came here looking to learn some facts and to save a little time
in the process but that has not happened. I have been able to gather
some new nuggets of knowledge from some very helpful people but not
enough so far to justify all the rest of the bull**** that's been going
back and forth.

Myke

--

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

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

Because is doesn't animate for me.

It does for me just fine. What's your browser?

IE6


So if the ATH has such tiny effect, which element of lossy compression
is the one most responsible for so much destruction?

Three answers.

1. WRT CD, the whole act of lossy compression.

2. WRT MP3 files of peak normalised versus rms-normalised/compressed, the
difference is in the dymanics of the music that is different to that which
the producer intended.

3. WRT rms-normalised and not compressed (generally resulting in clipping),
what can I say ?


And what are the "very low levels" you cite? Or what webpage is there
that has these levels spelled out for one all to clearly see?

No.

I don't deny your word that ATH affects "very low level" frequencies.
So before you go off saying you've presented me with facts and I've
simply ignored them, please tell me what they are or where I can find
them.

Try your ears.
1. Normalise a file to -5dB peak. Encode it.
2. Normalise the same file to 0dB peak. Encode it.
3. Load each resultant MP3 into a competent audio editor ( I see you have
CoolEdit ?), then renormalise each to 0dB.
4. Play back and listen. Hear any difference ? I didn't on my computer
speakers, or quality bookshelf monitors.

"The sky is blue," is true. But how blue is it? Is it royal blue? Is
it navy blue? Is it eyeball blue? No, it's *sky* blue!

Try lossy-encoding it and then try comparing the colour to the original.


Do you deny the sense in doing all that can be done in order to achieve
the best result - given the amount of inherent destruction that MP3
compression already causes?

In terms of an MP3 no. But the effect, in the scenario you variously
claim
(without compression), would be made equally by turning up your volume a
tad.

You keep telling me this but your presentation is such that you're just
ballparking to derive conclusions - which is no better than what I'm
doing - except I simply ballpark more weight on the effect of the ATH
filtration than you do. We could go back and forth like this all day.

But I fully understand the size shape surface and composition of my
ballpark.

geoff

Geoff Wood wrote:
"Lord Hasenpfeffer" wrote in message

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

Looking again, this picture illustrates clearly just how exaggerated your
notions are. Really - not sniping or exaggerating.

It's just a quick-n-dirty I just threw together in a few minutes to
visually depict what I've been trying to say in words.

What about it needs to be corrected?

The A T of H has nothing to do with the 'better sound' you like.

You've told me this a hundred times now but you present it as an opinion
with not numbers or *any* kind of real proof to support it.

If you can prove this to me then you will have disproven my hypothesis!

Ta-da!!!

Either that or you like clipping distortion.

No, I don't like clipping distortion - nor wrapping distortion either.
"Normalize" has its limiter turned on by default.

The 'frequencies' that you refer to are extremely low level instances in a
frequency-weighted regimen

Define "frequency-weighted regimen". I don't want to be unclear of the
terms again.

that is designed to have minimal audible impact.

OK, then what are the biggest, baddest offenders of all lossy
compression techniques?

*) Masking filtration
*) ATH-based filtration
*) bitrate settings
*) CBR data storage limits

What else?

Stronger levels of those 'frequencies' are not disgarded.

No, no... the frequencies themselves are not allowed to pass thru if
they don't have enough "amplitudinal clout" (how's that for a term?)

It's just like at a concert - unless you have a backstage pass you
aren't gonna get backstage. (Groupies notwithstanding!)

Myke

--

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

"Lord Hasenpfeffer" wrote in message

Define "frequency-weighted regimen". I don't want to be unclear of the
terms again.

'Important' frequency ranges are disgarded less than 'unimportant' ones.

OK, then what are the biggest, baddest offenders of all lossy
compression techniques?

*) Masking filtration
*) ATH-based filtration
*) bitrate settings
*) CBR data storage limits

No idea.

What else?

Stronger levels of those 'frequencies' are not disgarded.

No, no... the frequencies themselves are not allowed to pass thru if
they don't have enough "amplitudinal clout" (how's that for a term?)

Yes, that's pretty much what I said.

geoff

Lord Hasenpfeffer wrote:

It's more effective! No, it's less effective! No, it's more effective!
No, it's less effective! That gets us totally nowhere.

No. A LISTENING TEST, that has been suggested many times, would get you somewhere.
Now go do the listening test and then come back with your results. Stop
philosophing about it and listen for yourself what the difference is. If you can't
hear a difference, or it turns out the amplified file sounds worse, you have your
answer. No matter the theory behind it.


Hell, I came here looking to learn some facts and to save a little time
in the process but that has not happened. I have been able to gather
some new nuggets of knowledge from some very helpful people but not
enough so far to justify all the rest of the bull**** that's been going
back and forth.

Then by all means, stop, leave, do a listening test. Come back if you wish. But
stop wasting your time with discussing something in a place you don't get a
straight answer. Go to the MP3 discussion group I've posted a website of, and ask
there. Maybe even the coders are there.

Get all the coding facts, draw your conclusions. But none of them will *actually*
help you further if you don't do a listening test.

Erwin Timmerman

In article writes:

I am seeking facts regarding how to produce better MP3s than Joe Sixpack
can make by himself at home. It will help to give my business "a
leg-up" on the local competition.

First, you need to define "better." We've been trying to do that for
years. On one end of the scale you have "better" being a wide dynamic
range, on the other end of the scale, it's having a very narrow
dynamic range so that it can be encoded and played as close to maximum
level as possible. And off to the sides you have the "wow, listen to
all that bass" and "better restrict the frequency response so the IM
distortion won't make the vocal warble."

CD Mastering engineers make those decisions many times a day. It seems
that you should be looking into some of the basic philosophies of
mastering, and then try to achieve those goals within the MP3 format.


--
I'm really Mike Rivers - )

OK, then what are the biggest, baddest offenders of all lossy
compression techniques?

I think it'll be hard to find such a generality. What destroy's most
depends on the encoder you use.

Some MP3 encoders cut all ferquencies above a hard coded value, the worst
I've seen was cutting everything above 15000Hz, wich of course can destroy
a lot of music. So for that decoder the cutting of frequencies was possible
the worst offender.

Different encoders also use different psycho-acoustic models. Some encoders
use rather bad models wich in those encoders can be the worst offenders.

I doubt you'll find one single mechanism wich is the worst offender in all
encoders.

Of course the bitrate is allways important, but it's not an offender. The
whole purpose of lossy compression is to make a compromize between data
size and quality. When someone encode at a low bit rate said someone does
so because the size is a lot more important than the quality. When someone
purposefully encodes to low quality the resulting low quality shouldn't
offend said someone.

/Jonas

"Mike Rivers" wrote in message
news:znr1057145008k@trad...

In article
writes:

I am seeking facts regarding how to produce better MP3s than Joe Sixpack
can make by himself at home. It will help to give my business "a
leg-up" on the local competition.

CD Mastering engineers make those decisions many times a day. It seems
that you should be looking into some of the basic philosophies of
mastering, and then try to achieve those goals within the MP3 format.

Myke, if you are pursuing this as seriously as you claim, two books that can
help you are

Bob Katz, "Mastering Audio - The Art and Science"
http://www.digido.com/

and Bobby Owsinski, "The Mastering Engineer's Handbook".
http://www.amazon.com/exec/obidos/tg.../-/0872887413/

John

If you can prove this to me then you will have disproven my
hypothesis!

Have you done any of the testing suggested here yet? I would like to see
the results.

I know you like the Lame encoder (notlame uses the Lame encoding engine),
so you should be comfortable with the command line Lame encoder wich allows
you to specify the ATH threshold.

Download an app that can create a test tone at a specified amplitude.
Create a wav file with a test tone at a very low amplitude. Encode this
with a manually specified ATH (so you know for certain what it is). If the
test tone disappears your hypothesis is correct, if it doesn't disappear
your hypothesis is incorrect. Try this with different amplitudes and
settings if you feel uncertain of exactly what the parameters in the apps
really does.

And once you've done this, post the results here so we all may know the
answer.

If this shows your hypothesis to be correct, I would suggest that
using Lame and manually setting the ATH to a lower value would be a better
way to reach your stated goals than to compress the audio before encoding.

Regards
/Jonas

"Lord Hasenpfeffer" wrote in message ...

It'll do you no harm and it better illustrates my view.

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


Good for a laugh. Perhaps you should send this to a few technical
journals and some codec developers for analysis or publication.

"Lord Hasenpfeffer" wrote in message ...

So if the ATH has such tiny effect, which element of lossy compression
is the one most responsible for so much destruction?

The entire basis for the development of the codec, that's "which element."
Booting out frequencies from the source is how the codec's compression
scheme works. You cannot stop this from occurring. I'm sorry.

And what are the "very low levels" you cite? Or what webpage is there
that has these levels spelled out for one all to clearly see?

Screw a buncha' web pages. ;-) I thought you said you had read up
on the codec principles and taken in a lot of FAQs already.

Again, compression from the codec's point of view, makes it's OWN
decision as to what is or is not below the ATH at any given instant in
time during the encoding process. It will make the same set of decisions
based on what at that particular instant, is either in front of, or behind,
other more dominant frequencies. The codec scheme determines at
that instant, which frequencies are too low to bother encoding. You can't
change that without modeling a new data compression codec that operates
in a different manner on different principles.

I don't deny your word that ATH affects "very low level" frequencies.

You just seem to consider 'low level' a product of overall volume.

"Very low level" as implied here and as applied by the codec, does NOT
mean average RMS amplitude before encoding, but is rather the level of
frequencies that are compared to the remainder of the program content
at any given instant in time during encoding. This has been explained
several times by several different contributors.

By the way, I haven't thanked you for pushing me into a little research.

--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com

"Lord Hasenpfeffer" wrote in message ...

No, no... the frequencies themselves are not allowed to pass thru if
they don't have enough "amplitudinal clout" (how's that for a term?)

When you elevate the volume of a .wav, you are elevating ALL frequencies
(it's entire content) by the exact same amount. You have changed not
a thing but the overall volume (excluding any resultant clipping if it occurs).

The *codec* will still make the decision as to which frequencies to discard
based on the amplitudinal clout of ALL frequencies present at a given point
in time.

--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com

Mike Rivers wrote:

I am seeking facts regarding how to produce better MP3s than Joe Sixpack
can make by himself at home. It will help to give my business "a
leg-up" on the local competition.

First, you need to define "better."

Oh, gosh, I forgot to remember to include the quotes around that evil
word. I usually do but this time I didn't and I definitely *know*
better than to ever leave them out in a discussion in a forum such as
this! :-) Of course, "better" cannot be defined beyond individual taste.

CD Mastering engineers make those decisions many times a day.

I understand that and I have great respect for them, their knowledge and
their craft (contrary to what some would have you believe of me).

It seems that you should be looking into some of the basic philosophies of
mastering, and then try to achieve those goals within the MP3 format.

I am indeed fascinated with *all* of these issues. As a self-styled
noisician and sound-art composer who is soon to self-produce a whole
slew of "digitally remastered" CD-R versions of his own creations
recorded all throughout the past 20 years, it is extremely important
that I know the basic philosophies of audio CD mastering - but in my
day-to-day life as a professional webmaster, etc., the MP3 requirement
must come first.

Myke

--

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

In article , MAMS\
wrote:

By the way, I haven't thanked you for pushing me into a little research.

Research all of us need to be doing!

I've been utterly shocked at the parroting of misinformation in most of
the articles on lossy-coding that have been published by our beloved
gear-pimping magazines. Most attempts to second guess lossy-coding are
doomed to sound really bad just as attempts to second guess broadcast
processing are.

--
Bob Olhsson Audio Mastery Recording Project Design and Consulting
Box 90412, Nashville TN 37209 Tracking, Mixing, Mastering, Audio for Picture
615.385.8051 FAX: 615.385.8196 Mix Evaluation and Quality Control
40 years of making people sound better than they ever imagined!

Jonas Eckerman wrote:
What destroy's most depends on the encoder you use.

Thanks. Now that you've reminded me of it, I must agree.

Myke

--

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

John LeBlanc wrote:

Myke, if you are pursuing this as seriously as you claim, two books that can
help you are

I am.

Bob Katz, "Mastering Audio - The Art and Science"
http://www.digido.com/

and Bobby Owsinski, "The Mastering Engineer's Handbook".
http://www.amazon.com/exec/obidos/tg.../-/0872887413/

Thank you.

Myke

--

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

Jonas Eckerman wrote:

Have you done any of the testing suggested here yet?

Some yes, though not as much as I would have preferred to date.
It seems that the wife and sleep are dominating all the time I've been
able to break away from all the back-and-forth in here.

I would like to see the results.

I would too.

I know you like the Lame encoder (notlame uses the Lame encoding engine),
so you should be comfortable with the command line Lame encoder wich allows
you to specify the ATH threshold.

Yes, I noticed this both a couple of years ago and a couple days ago but
had generally forgotten it in-between.

Makes me wonder if notlame can be similarly adjusted. I will review the
man page for it.

Download an app that can create a test tone at a specified amplitude.

I believe I've got a WAV of just such a test tone I made once made from
the end of an old videotape. Some videotapes used to have such long,
annoying tones at their ends - not unlike the off-air test tones of TV
past - which was, I believe, to wake you if you fell asleep while
watching the movie! Unless you have reservations about this as being
an appropriate test WAV I think I'll hunt it up and use it. I have no
clue at this point about what software I could use to create such a tone.

Myke

--

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

Jonas Eckerman wrote:

From what I could find with a quick glance at the history of Lame, the Lame
encoder now uses a contniously adjusted ATH and not a fixed one. That
doesn't mean that other encoders also works that way, only that tyhe
current Lame encoder does. Versions of the Lame encoder previous to that
modification seems to have used a fixed ATH.

It seems to me that at least in the theoretical, the fixed / not-fixed
positioning of the ATH has probably more to do with proving my
hypothesis than *most* of what else been stated here so far. (But that
is not to suggest that I'm now ignoring everyone else's input, ideas and
suggestions...)

And, if it does ... and it is true that old LAME engines v. new LAME
engines differ in this apparently most critical sense then I suspect
that (1) the effects of the ATH filtration on the quality of resultant
MP3s are greater than some might like to admit and (2) the correct
answer to whether or not my hypothesis is true is clearly "Yes and no."

Thank you Jonas, you may have won the prize. :-)

Myke

--

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

David Morgan (MAMS) wrote:

It'll do you no harm and it better illustrates my view.

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

Good for a laugh.

Well, that's one thing at least about which we can easily agree. ;-)

Perhaps you should send this to a few technical journals
and some codec developers for analysis or publication.

I believe it's already "out there" enough as it is, thankyavurrymuch.

Myke

--

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

"Lord Hasenpfeffer" wrote in message ...

Thank you Jonas, that's exactly what those three images are showing.
The 3rd and loudest one that we both agree is *slightly* compressed (not
just limited???)

Who compressed it? What was the compression source? You never
mentioned performing any function other than RMS normalization. Why
have you not mentioned before that you were using compression?

Personally, I cannot see evidence of any compression in the .gif file.

--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com

"Lord Hasenpfeffer" wrote in message ...
Jonas Eckerman wrote:
What destroy's most depends on the encoder you use.

Thanks. Now that you've reminded me of it, I must agree.


Gee !!!

My first question to you was, "What encoders have you studied?"

Your reply essentially? 'Only the one I got for free'.

And, if it does ... and it is true that old LAME engines v. new LAME
engines differ in this apparently most critical sense then I suspect
that (1) the effects of the ATH filtration on the quality of resultant
MP3s are greater than some might like to admit

Not necessarily.

The people working on Lame put in a lot of effort in every part of the
encoding process. They try hard to adress *all* issues affecting the
quality of the encoded audio as well as those affecting files size and
encoding speed. Both the big issues and the small.

From the history document we can see that modifications to the ATH handling
have been done both for raising qualilty and for minimizing file size. It
is quite possible that changes for minimizing file size actually affected
quality in a negative way in at least some cases. After all, minimizing
file size at the expense of quality is the purpose of Lame even though they
try very hard to make this expense as low as possible.

I don't have the knowledge necassary for assesing how big the impact of
Lames changes in handling of ATH has been on the quality. If knowing more
about this was important I'd go the the Lame home site, find a link to the
discussion groups/forums/lists where I'd be most likely to find people who
has that knowledge.

And I bet those same people could answer a whole lot of other questions
about the MP3 encoding process as well as tell me what parameters to Lame
they personally likes to use. :-)

Personaly I have no idea what of the changes Lame has had since it first
became a complete encoder has been the most important changes WRT sound
quality. I do know that I've never seen the ATH changes mentioned before I
looked at the history document, and that I have seen other changes
mentioned as beeing important.


Thank you Jonas, you may have won the prize. :-)

Thanks. What do I get? :-)

Actually, I feel I have won something. I've won some knew knowledge about
MP3 encoders.

But I honestly think that I'm not the only person in this group who has
come to that conclusion nor the only person who has posted it to the group.

Regards
/Jonas

David Morgan (MAMS) wrote:

Thank you Jonas, that's exactly what those three images are showing.
The 3rd and loudest one that we both agree is *slightly* compressed (not
just limited???)

Who compressed it? What was the compression source? You never
mentioned performing any function other than RMS normalization. Why
have you not mentioned before that you were using compression?

Personally, I cannot see evidence of any compression in the .gif file.

Neither can I. That's why I asked "not just limited???"

I can see very slight limiting occurring in the animation.
I can see no compression occurring.

The point of uploading the image though was to debunk the baseless
claims that I am "severely dicking with the music" by doing this to my
WAVs prior to encoding.

Myke

--

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

David Morgan (MAMS) wrote:

Gee !!!

My first question to you was, "What encoders have you studied?"

Your reply essentially? 'Only the one I got for free'.

That is correct. I have heard MP3s in the past which were obviously
poorly made. With notlame, however, I have not been able to produce a
single MP3 which has offended my ears other than those which were made
from older, quieter WAVs ripped from older, quieter CDs. "normalize"
solved that problem for me, therefore, I have no complaints with my encoder.

Myke

--

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

Jonas Eckerman wrote:

Thank you Jonas, you may have won the prize. :-)

Thanks. What do I get? :-)

How about The Lord H Platter? A set of my favourite RMS normalized
MP3s? Just kidding. They've been "severely dicked with" I tell you! :-)

Actually, I feel I have won something. I've won some knew knowledge about
MP3 encoders.

There ya go!

But I honestly think that I'm not the only person in this group who has
come to that conclusion nor the only person who has posted it to the group.

One day it was stated that ATH-based filtration was fixed in relation to
Full Scale. Another day it was stated that ATH-based filtration was
relative to max-peak level on a frame-by-frame basis. But you, sir,
were the one who first posted the URL which revealed to me that *both*
are true depending upon the age of one's LAME encoder! AFAIK, all
prior arguments - including my own - seemed to rely on the assumption
that either one or the other was true at all times without taking into
consideration the shift in the LAME developers' policy from using a
fixed-position to using relative-positions as progress ensued.

(1) If ATH-based filtration occurs with older LAME encoders from a fixed
position relative to Full Scale, my hypothesis should be true.

(2) If ATH-based filtration occurs with newer LAME encoders from a
relative position in relation to max-peak levels on a frame-by-frame
basis, my hypothesis should be false.

Therefore, the answer to all of this is clearly "yes and no".

Myke

P.S. What I want to know now is what the policy in this regard has been
throughout the history of ATRAC development for MiniDisc - but I'm not
going to start my search for that particular answer in this forum.

--

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

Personally, I cannot see evidence of any compression in the .gif
file.

Neither can I. That's why I asked "not just limited???"

My bad.

When I looked at the file I did not think about what you had actually done,
adn as I did like the way it switched pictures continuously I didn't
compare the pictures as much as I would have done if it had kep still.

To my glance it looked like you might have used a very slight compression
to get rid of the highest peaks and then turned the gain up a bit, while in
reality you had turned the gain up a bit with limiting applied to avoid
digital clipping.

The point of uploading the image though was to debunk the baseless
claims that I am "severely dicking with the music" by doing this to my

I think that some of that misunderstanding could have been prevented if you
had not just said that you were batch normalizing all the tracks from a CD.
It actually took a while for you to explain what *you* meant by "batch" in
this case.

Normally batch whatever simply means that you've told the machine to do
whatever and then left it to do that instead of interactively telling it
to do the different steps requiered for whatever.

With most applications, batch normalizing would simply mean that the
application normalizes a whole bunch of files, one file at a time, in one
batch. It does not mean that the app first analyzes the whole batch of
files and then applies the same change to the whole batch, wich is what you
seem to mean.

Regards
/Jonas

Geoff Wood wrote:
"Lord Hasenpfeffer" wrote in message

I can see very slight limiting occurring in the animation.
I can see no compression occurring.

The point of uploading the image though was to debunk the baseless
claims that I am "severely dicking with the music" by doing this to my
WAVs prior to encoding.

Music that already approaches your desired RMS level will not be altered
much. Most music will. Sugeest you look at before/after on a range of
your music. Not just sections of waqveform but whole tracks.

We're talking about an approximately +5dB boost to this one file,
however, in order to pump it up to -10dBFS.

The original levels of the MFSL DSotM CD are far lower than those of my
CD copy of U2's, "Sunday Bloody Sunday" - yet that entire album only
required a boost of +4.5dB to reack -10dBFS.

So this "normalize" application of which I am so fond does indeed appear
to consider the source WAV's soft/loud ratio appropriately. Otherwise,
the "quieter" Pink Floyd WAV would have required more pumpitude not less
than the "louder" U2 WAV in order to match their averages.

Myke

--

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

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

I can see very slight limiting occurring in the animation.
I can see no compression occurring.

The point of uploading the image though was to debunk the baseless
claims that I am "severely dicking with the music" by doing this to my
WAVs prior to encoding.

Music that already approaches your desired RMS level will not be altered
much. Most music will. Sugeest you look at before/after on a range of
your music. Not just sections of waqveform but whole tracks.

We're talking about an approximately +5dB boost to this one file,
however, in order to pump it up to -10dBFS.

The original levels of the MFSL DSotM CD are far lower than those of my
CD copy of U2's, "Sunday Bloody Sunday" - yet that entire album only
required a boost of +4.5dB to reack -10dBFS.

So this "normalize" application of which I am so fond does indeed appear
to consider the source WAV's soft/loud ratio appropriately. Otherwise,
the "quieter" Pink Floyd WAV would have required more pumpitude not less
than the "louder" U2 WAV in order to match their averages.

Myke

--

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

Jargon strikes again!

Exactly. A very common reason for misunderstanings. This time though it was
a computer term and not an audio term...

Your use of the term "batch" was correct, but not enough. In the case of
your RMS normilizing process "batch" did not explain all you did, it only
told us that you normalized a batch of files.

I think this is a danger in learning terms from the documentation for *one*
application (or appliance or whatever). You only learn what that
application does, not what the term actually means. Most software
documentation doesn't try to teach you what the terms generally mean, only
what they mean in the context of that specific application.

P.S. In the meantime, have another gander:

I can see only a very slight limiting going on here.

This one dates back a bit now so don't read too much into my personal
commentary thats included in the filename...

I guess that you got a bunch of critical replies when you showed that image
without first telling people to disregard those comments.

BTW, I think the paper looks better than the clock. :-)

Regards
/Jonas

In article writes:

I am indeed fascinated with *all* of these issues. As a self-styled
noisician and sound-art composer who is soon to self-produce a whole
slew of "digitally remastered" CD-R versions of his own creations
recorded all throughout the past 20 years, it is extremely important
that I know the basic philosophies of audio CD mastering - but in my
day-to-day life as a professional webmaster, etc., the MP3 requirement
must come first.

Bob Katz' book Mastering Audio which someone else mentioned is mostly
about listening, how to set yourself up for better listening, how to
listen better, and what to listen for. The fact that your final
delivery medium is MP3 and Bob's is mostly CD doesn't really matter
all that much - you still want to get as close as you can to the same
place. Though you might use different tools to get there, you still
need to evaluate all through the process where you are and whether
you're going in the right direction.

I'd recommend that you get a copy and study it for a while. Order it
direct from Bob's web site http://www.digido.com and he'll make a
little extra money over what he'd get if you ordered it from the
publisher or a book store.




--
I'm really Mike Rivers - )

"Lord Hasenpfeffer" wrote in message ...

So this "normalize" application of which I am so fond does indeed appear
to consider the source WAV's soft/loud ratio appropriately.

Without seeing the software, I couldn't tell you. But I suspect (very highly)
that it sees *nothing* outside of your command to use it and to what degree.
It doesn't care anything *at all* about the material... only about it's operation.
It is an algorithm designed to perform a specific function - it is not intelligent.

Otherwise,
the "quieter" Pink Floyd WAV would have required more pumpitude not less
than the "louder" U2 WAV in order to match their averages.

It would depend on which type of normalization you used (peak or avg.)

Do you understand the difference between peak and average levels?
no answer required

Do you know what a 'transient' peak is?

--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com

"Lord Hasenpfeffer" wrote in message ...
David Morgan (MAMS) wrote:

Gee !!!

My first question to you was, "What encoders have you studied?"

Your reply essentially? 'Only the one I got for free'.

That is correct. I have heard MP3s in the past which were obviously
poorly made. With notlame, however, I have not been able to produce a
single MP3 which has offended my ears other than those which were made
from older, quieter WAVs ripped from older, quieter CDs. "normalize"
solved that problem for me, therefore, I have no complaints with my encoder.

Myke


Without experimenting with codecs and much listening, you will never know.

At least we know, that to you, louder is psychoacoustically 'better'.

"Lord Hasenpfeffer" wrote in message ...

(2) If ATH-based filtration occurs with newer LAME encoders from a
relative position in relation to max-peak levels on a frame-by-frame
basis, my hypothesis should be false.


In order for the encoder to make the ATH decisions, it must sample
the data continuously while writing.

"Lord Hasenpfeffer" wrote in message ...

http://www.mykec.com/mykec/images/20...ey_Smoking.png

This one dates back a bit now so don't read too much into my personal
commentary thats included in the filename...

Really. I'm glad you included this disclaimer. I would be depressed to
believe that you were obtuse enough to believe that you were offering
some miraculous new "mastering" service solely based on your
new-found normalization tool.

But I'll answer the question you posed there... I had rather listen to the
one with all of the waveform intact, without the negative and positive
peaks lopped off. (*Even* if I had to turn up the volume control).


Please re-read this post by "WBRW" :
om...


--
David Morgan (MAMS)
http://www.m-a-m-s.com
http://www.artisan-recordingstudio.com

In article ,
Lord Hasenpfeffer wrote:
So if the ATH has such tiny effect, which element of lossy compression
is the one most responsible for so much destruction?

Call me a mad crazy person, but I think 'starting with the loudest
overtones, filling up only 128Kbs with progressively weaker overtones so
long as they're not real close to the loud overtones, and then throwing
everything else away completely, most of which is way the hell over ATH
thresholds most of the time' might have something to do with it.

All, ALL ATH gating does is it clears out a bit of sonic space to
trowel a precious bit of more data bandwidth onto the shambling hulk.
That's it. You can't crow about restoring the low level detail because
you AREN'T GETTING IT, not even at 320K. Do you know what the data
bandwidth of simple CD audio is? 1411K. _Ten_ times your 128K CBR.

All you're doing is deciding you like peak limiting and about 10 db
of crest factor. Be happy with that, you're lucky. I like 15 db and up
of crest factor, so modern CDs that are a bit too loud for you are
unlistenably squashed for me.


Chris Mad Crazy Person, er, Johnson

David Morgan (MAMS) wrote:

Otherwise,
the "quieter" Pink Floyd WAV would have required more pumpitude not less
than the "louder" U2 WAV in order to match their averages.

It would depend on which type of normalization you used (peak or avg.)

Both "Sunday Bloody Sunday" and the entire "Dark Side..." WAVs were
"limitized" with "normalize" to an average RMS level of -10dBFS for the
sake of keeping variables to a minimum.

Do you understand the difference between peak and average levels?
no answer required

Yes.

Do you know what a 'transient' peak is?

Maybe. There are many times when I understand a term from having
experienced it in practice even though I'm not familiar with the jargon
used to describe in English.

Myke

--

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