The Logic Off Topic list

Gilles Ruppert wrote:
> Sorry to tell you Sascha: wrong, see my other post. if you don't believe
me
> check out Ken C Pohlmann, Principles of Digital Audio. You might not hear
> the mistake, but it comes up & you will see it when you analyse it.

Sorry, but I DID anylyse it.
Or do you think that digital silence is something which just happens by
accident?
It only (!) happens when inverting two signals if the two sources are 100%
identical.

> More
> steps & you will start to hear the mistake (I speak about A LOT MORE
> PROCESSES).  It becomes eventually audible in Fade outs or low volume
> passages.

I repeated my steps 5 times, still there was digital silence.

> OK. Imagine this: you have a 3 bit soundfile, which gives you 8 steps. The
> volume of your sine tone is "7". You reduce the volume by -6 dB
(normalizing
> does nothing else than changing volume), though half as loud.

Ok, there's your mistake.
Normalizing NEVER (!) lowers any level. All it does is to bring up the level
to 0dB and that's it.

> 7/2 = 3.5, but
> how will it be showed, as 3.5 does not exist? It is either 3 or 4 in the
> digital domain & voila our rounding error & quantization noise.

Sure, but we are NOT reducing (= dividing) any levels when normalizing.
You have to do your calculations the other way round:
7x2 = 14
14/2 = 7
=> no rounding errors possible.


> Well with 16
> bit & 24 bit, the resulting noise is of course of much lower volume,
because
> the steps are far smaller. Of course you are also working with a complex
> wavefile & the result will be different for every cycle & at very low
> volume, but believe me, it has a reason why Mastering Engineers do not
> normalize!

There *might* be some reasons mastering engineers don't normalize, I don't
know anything about that, all I know is that normalizing doesn't change the
sound at all, this is perfectly proveable, see my last posts example, it
works under any circumstance.

Regards,
Sascha

I try to have the biggest sound possible before clipping when I record from my VMC7200 roland
into logic, and I wonder if it's better to normalize the mix before mastering.
Thanks...phil...

don't normalize before mastering: every process which ios done is
"worsening" the audio quality (even if you do not hear it, but you
eventually will having enough processes in the chain). You better put the
treshold of your compressor a little bit lower, or increase the input gain
or what ever of your mixing desk. May be you could normalize the file if the
preamps of your mixing desk are noisuy, but I was told by several mastering
engineers that you really do not need normalizing (and should not use it),
because it is 1 more calculation process which detoriates the audio quality.
Hope this helps,
Gilles

> -----Original Message-----
> From: salziger@... [mailto:salziger@...]
> Sent: 15 November 2001 18:25
> To: logic-ot@yahoogroups.com
> Subject: [L-OT] normalize before mastering or not
>
>
> I try to have the biggest sound possible before clipping when I
> record from my VMC7200 roland
> into logic, and I wonder if it's better to normalize the mix
> before mastering.
> Thanks...phil...
>
>
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

No.

> From: salziger@...
> Reply-To: logic-ot@yahoogroups.com
> Date: Thu, 15 Nov 2001 18:24:41 -0000
> To: logic-ot@yahoogroups.com
> Subject: [L-OT] normalize before mastering or not
> 
> I try to have the biggest sound possible before clipping when I record from my
> VMC7200 roland
> into logic, and I wonder if it's better to normalize the mix before mastering.
> Thanks...phil...
> 
> 
> 
> 
> 
> 
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>

marc@... writes:
>No.
why not, if the normalising process is of a high quality, and the peak is 
maintained below approx. -0.5 dB?
best,
dt / splattercell

>> From: salziger@...
>> Reply-To: logic-ot@yahoogroups.com
>> Date: Thu, 15 Nov 2001 18:24:41 -0000
>> To: logic-ot@yahoogroups.com
>> Subject: [L-OT] normalize before mastering or not
>> 
>> I try to have the biggest sound possible before clipping when I record
>from my
>> VMC7200 roland
>> into logic, and I wonder if it's better to normalize the mix before 
mastering.
>> Thanks...phil...

salziger@... on 16/11/01 5:24 AM, salziger@... at salziger@...
wrote:

> I try to have the biggest sound possible before clipping when I record from my
> VMC7200 roland
> into logic, and I wonder if it's better to normalize the mix before mastering.
> Thanks...phil...

Don't do it. If you're music is of a percussive nature and you're mixing it
well, the peaks of your mix should be above 80-85% of the maximum. For the
little you gain by normalising you'll do more harm than good.

Also, I'm not up on the theory of this myself but it's been said by many a
knowledgable person that if you're going to be mastering in the digital
domain and apply digital calculations to your mix, then you do need to allow
a bit of headroom in your mix for these digital calculations to take place
without making the maths involved potentially cause problems.


-- 
/\/\/\/\/\-/\/\/\-/\/\/\/\/\
Paul Najar
Jaminajar Music Production
Studios 301 - "The Lounge Room"
http://www.jaminajar.com.au/the_Lounge_Room.html
http://www.301.com.au

See my last mail:
more processes, worse quality.

>
> marc@... writes:
> >No.
> why not, if the normalising process is of a high quality, and the peak is
> maintained below approx. -0.5 dB?
> best,
> dt / splattercell

gilles,
>See my last mail:
>more processes, worse quality.
maybe i'm wrong, but i thought that was only true in cases wherein dithering 
and/or noise-shaping processes were repeatedly applied.
?
best,
dt / s-c

<texture444@...> wrote:
> maybe i'm wrong, but i thought that was only true in cases wherein
dithering
> and/or noise-shaping processes were repeatedly applied.
> ?

You're completely right.
Normalizing doesn't harm the sound quality in the slightest way (well, it
brings your peaks up to 0dB and that might harm your ears when they were
at -30 or wherever before, but that's got nothing to do with the sound
quality per se...).

Everybody can do a test by themselves (can be done completely inside Logic):
- Take a loop or whatever soundfile. Peaks should be pretty much below 0dB.
If not, reduce the level, let's say to -5dB.
- Copy the file and phase invert it.
- Play them against each other.
=> Until now the playback result should be (digital) silence.

- Take the inverted file and add 3dB (normalizing doesn't do anything else
but adding some level).
- Reduce the level by 3dB.
- Add 3dB.
- Reduce...
- Add...
- Do that as often as you like, just make sure to finish with reducing so
you have the same level as in the beginning.
- Play the files against each other.
=> Still silence!

This is the defenite mathematical prove that normalizing doesn't alter your
sound quality at all (unless your normalization algorhythm is a bad/wrong
one).

However, some people allways only normalize after any mastering was applied.
It seems to me that some people even tend to finish their mixes a bit lower
than 0dB to keep a bit of headroom for mastering (some mastering effects
apparently might sound different).
I'm the last one to tell if this is a good or bad idea (I'm allways trying
to bounce down my final mixes as hot as possible to keep the maximum
resolution, but that's just because I don't have a clue), but in the end it
defenitely shouldn't matter at which stage you normalize (proper mastering
effects should have some input control in case they need some headroom to
work properly).
Maybe someone with a bit more of mastering experience could jump in and give
an explanation, sound quality reduction while normalizing defenitely isn't
it.

Regards,
Sascha

sorry to tell you: you are wrong, every digital calculation brings you
quantization noise due to quatization / rounding error (error which comes
when you calculate. Due to the reduced stepes (65536 at 16 bit & 16777216 at
24bit), the calcualation process always has to round to the closest step.
Gilles

> >See my last mail:
> >more processes, worse quality.
> maybe i'm wrong, but i thought that was only true in cases
> wherein dithering
> and/or noise-shaping processes were repeatedly applied.
> ?
> best,
> dt / s-c

Sorry to tell you Sascha: wrong, see my other post. if you don't believe me
check out Ken C Pohlmann, Principles of Digital Audio. You might not hear
the mistake, but it comes up & you will see it when you analyse it. More
steps & you will start to hear the mistake (I speak about A LOT MORE
PROCESSES).  It becomes eventually audible in Fade outs or low volume
passages.

OK. Imagine this: you have a 3 bit soundfile, which gives you 8 steps. The
volume of your sine tone is "7". You reduce the volume by -6 dB (normalizing
does nothing else than changing volume), though half as loud. 7/2 = 3.5, but
how will it be showed, as 3.5 does not exist? It is either 3 or 4 in the
digital domain & voila our rounding error & quantization noise. Well with 16
bit & 24 bit, the resulting noise is of course of much lower volume, because
the steps are far smaller. Of course you are also working with a complex
wavefile & the result will be different for every cycle & at very low
volume, but believe me, it has a reason why Mastering Engineers do not
normalize!
Sorry, I don't have time to get more into this, but If anybody has a
question, they are welcome to contact me privately (i have to unsubscribe
for a while, I am too busy).
Hope this helps (although my explanation might be a bit confusing, in the
morning my english is not that great :-))
Gilles

Gilles Ruppert
gilles@...
www.latower.com
ICQ: 113166277

> -----Original Message-----
> From: Sascha Franck [mailto:saschafranck@...]
> Sent: 16 November 2001 06:41
> To: logic-ot@yahoogroups.com
> Subject: Re: [L-OT] normalize before mastering or not
>
>
> <texture444@...> wrote:
> > maybe i'm wrong, but i thought that was only true in cases wherein
> dithering
> > and/or noise-shaping processes were repeatedly applied.
> > ?
>
> You're completely right.
> Normalizing doesn't harm the sound quality in the slightest way (well, it
> brings your peaks up to 0dB and that might harm your ears when they were
> at -30 or wherever before, but that's got nothing to do with the sound
> quality per se...).
>
> Everybody can do a test by themselves (can be done completely
> inside Logic):
> - Take a loop or whatever soundfile. Peaks should be pretty much
> below 0dB.
> If not, reduce the level, let's say to -5dB.
> - Copy the file and phase invert it.
> - Play them against each other.
> => Until now the playback result should be (digital) silence.
>
> - Take the inverted file and add 3dB (normalizing doesn't do anything else
> but adding some level).
> - Reduce the level by 3dB.
> - Add 3dB.
> - Reduce...
> - Add...
> - Do that as often as you like, just make sure to finish with reducing so
> you have the same level as in the beginning.
> - Play the files against each other.
> => Still silence!
>
> This is the defenite mathematical prove that normalizing doesn't
> alter your
> sound quality at all (unless your normalization algorhythm is a bad/wrong
> one).
>
> However, some people allways only normalize after any mastering
> was applied.
> It seems to me that some people even tend to finish their mixes a
> bit lower
> than 0dB to keep a bit of headroom for mastering (some mastering effects
> apparently might sound different).
> I'm the last one to tell if this is a good or bad idea (I'm allways trying
> to bounce down my final mixes as hot as possible to keep the maximum
> resolution, but that's just because I don't have a clue), but in
> the end it
> defenitely shouldn't matter at which stage you normalize (proper mastering
> effects should have some input control in case they need some headroom to
> work properly).
> Maybe someone with a bit more of mastering experience could jump
> in and give
> an explanation, sound quality reduction while normalizing defenitely isn't
> it.
>
> Regards,
> Sascha

--- In logic-ot@y..., "Gilles Ruppert" <gilles@l...> wrote:
> Sorry to tell you Sascha: wrong, see my other post. if you don't believe me
> check out Ken C Pohlmann, Principles of Digital Audio. You might not hear
> the mistake, but it comes up & you will see it when you analyse it. More
> steps & you will start to hear the mistake (I speak about A LOT MORE
> PROCESSES).  It becomes eventually audible in Fade outs or low volume
> passages.

It is true that you should avoid normalizing, when afterwards you are going to apply A LOT of processing. But I wonder if it's not the processing that will mangle the original sound more than the error due to calculations.
 
> OK. Imagine this: you have a 3 bit soundfile, which gives you 8 steps. The
> volume of your sine tone is "7". You reduce the volume by -6 dB (normalizing
> does nothing else than changing volume), though half as loud. 7/2 = 3.5, but
> how will it be showed, as 3.5 does not exist? It is either 3 or 4 in the
> digital domain & voila our rounding error & quantization noise. Well with 16
> bit & 24 bit, the resulting noise is of course of much lower volume, because
> the steps are far smaller. Of course you are also working with a complex
> wavefile & the result will be different for every cycle & at very low
> volume, but believe me, it has a reason why Mastering Engineers do not
> normalize!

This is a wrongly explained example that is used to justify to not use normalization. When normalizing, floating point calculus is used, and not the way you explain it above. When using floating point calculus, the mistakes will be much lower.

One of the reasons why not to use normalize is mentioned on the digido.com site. Here is a quote from one of the articles:

"The Myth of "Normalization"
Digital audio editing programs have a feature called "Normalization",
a semi-automatic method of adjusting levels. The engineer selects
all the segments (songs), and the computer grinds away, searching
for the highest peak on the album. Then the computer adjusts the
level of all the material until the highest peak reaches 0 dBFS.
This is not a serious problem esthetically, as long as all the
songs have been raised or lowered by the same amount. But it is
also possible to select each song and "normalize" it individually. Since the ear responds to average levels, and normalization
measures peak levels, the result can totally distort musical values.
A compressed ballad will end up louder than a rock piece! In short,
normalization should not be used to regulate song levels in an
album. There's no substitute for the human ear."

It's a much more pratical reason not to normalize.
Hope this clears s things up a bit. Mastering is still an art on itself, just like mixing is, or composing. Tell me about it, I've been trying to master a track of mine for a couple of months; still don't like the result, but I'm getting there, :-).
Yoonchi.

Sorry if I missed this, but my understanding of normalization is that you
find  the highest value "a peak" then multiply the bits uniformly by a
factor that will bring the peak to 0db. I don¹t see why multiplying would
cause errors since the magnification factor is uniformly applied to the
sample data with respect to time.

It would cause a slight increase in noise - either quantization noise or
dither noise if dithered.  Point is, the mastering engineer is further
processing the signal so why do it.

> From: LogicBaby <basharar@...>
> Reply-To: logic-ot@yahoogroups.com
> Date: Sat, 17 Nov 2001 03:25:11 +0800
> To: "logic ot yahoogroups.com" <logic-ot@yahoogroups.com>
> Subject: Re: [L-OT] normalize before mastering or not
> 
> Sorry if I missed this, but my understanding of normalization is that you
> find  the highest value "a peak" then multiply the bits uniformly by a
> factor that will bring the peak to 0db. I don¹t see why multiplying would
> cause errors since the magnification factor is uniformly applied to the
> sample data with respect to time.
> 
> 
> 
> 
> 
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>

>It would cause a slight increase in noise - either quantization noise or
>dither noise if dithered.  Point is, the mastering engineer is further
>processing the signal so why do it.


If a signal were maxed out...and a mastering engineer were going to EQ 
and send it over the top, why can't he simply turn the signal down before 
it hits the EQ? There's a thing called a 'fader' that can be set 
pre-send, and these mastering engineers are surely clever fellows.

>Sorry if I missed this, but my understanding of normalization is that you
>find  the highest value "a peak" then multiply the bits uniformly by a
>factor that will bring the peak to 0db. I donpit see why multiplying would
>cause errors since the magnification factor is uniformly applied to the
>sample data with respect to time. 

Its seems preferable to normalize a track then push the fader down, to 
minimize noise .... but maybe that is ancient thinking, since you're 
going to bounce digitially, maybe no further noise is introduced from 
preamps or whatever.

However, its easy enough to see why there are problems if you limit, and 
then apply any kind of affect. In the Waves manual they talk about using 
EQ (boosting or cutting) and how that can cause clipping. Also modulation 
affects like phase shifting, flanging, etc where there is a circular 
motion which adds and subtracts can cause this.

A related issue is that in limiting its often the stage were one dithers 
from 24b to 16b, and the dithering creates noise in inaudible regions - 
which can become audible after processing.

Of course, I barely know what I am talking about here, so I hope some 
more knowledgable person steps in.

So what number of db corresponds to multiplying the volume by exactly 2? 
That will preserve the bit relationships, right? Is 3 db "twice as loud"? 
I think I read that this is very subjective and there are different 
opinions, but there should be one definite answer as to sliding every bit 
over one place from least toward most significant bits. For example turn 
7 = 0111 to 14 = 1110.

>OK. Imagine this: you have a 3 bit soundfile, which gives you 8 steps. The
>volume of your sine tone is "7". You reduce the volume by -6 dB (normalizing
>does nothing else than changing volume), though half as loud. 7/2 = 3.5, but
>how will it be showed, as 3.5 does not exist? It is either 3 or 4 in the
>digital domain & voila our rounding error & quantization noise. Well with 16
>bit & 24 bit, the resulting noise is of course of much lower volume, because
>the steps are far smaller. Of course you are also working with a complex
>wavefile & the result will be different for every cycle & at very low
>volume, but believe me, it has a reason why Mastering Engineers do not
>normalize!

At 05:08 AM 17/11/01 -0500, you wrote:
>So what number of db corresponds to multiplying the volume by exactly 2? 
>That will preserve the bit relationships, right? 
 Is 3 db "twice as loud"? 

6 dB is twice as loud. Two violins are 3dB louder than one violin, it takes
four to be twice as loud or 6 dB louder.

>I think I read that this is very subjective and there are different 
>opinions, but there should be one definite answer as to sliding every bit 
>over one place from least toward most significant bits. For example turn 
>7 = 0111 to 14 = 1110.

Arithmetical operations on 16 or 24 linear audio file produce rounding
errors in the Least Significant Bit. 
Normalising is an example of this type of operation. It does not increase
the resolution of your file -- it boosts the noise with the signal and adds
more noise. It is often suggested therefore,  that you do all your
processing of the file -- FX ,EQ, compression, reverb etc and boost only in
the final stage to get maximum level  to take best advantage of the limited
resolution of the CD format (16 bit linear, 44.1KHz sampling). 

The internal processing of Logic's native effects and VST effects etc is
all at 24 bit float so the absolute level of the file that undergoes that
processing doesn't matter beyond the usual caveat that the highest
resolution is obtained by recording as near as possible to full scale
without clipping.

Regards,
Murray

> Re: normalize before mastering or not

definitely NOT!!! bad idea.

-- 
GO SEE 
http://www.teddybut.com and http://www.mp3.com/teddybut
Kumpelstiltskin Music, Inc./EMI Music Publishing
Burning Beagle Studios, Brooklyn, NY

HEY, wake up.... don't take offense personally but....
you guys are anal retentive. please get over it, don't make red lights go
off and stop driving each other crazy please. What does this have to do with
anything musical??? nothing. that's right. you have much better things to
worry about.

tk

> At 05:08 AM 17/11/01 -0500, you wrote:
>> So what number of db corresponds to multiplying the volume by exactly 2?
>> That will preserve the bit relationships, right?
> Is 3 db "twice as loud"?
> 
> 6 dB is twice as loud. Two violins are 3dB louder than one violin, it takes
> four to be twice as loud or 6 dB louder.
> 
>> I think I read that this is very subjective and there are different
>> opinions, but there should be one definite answer as to sliding every bit
>> over one place from least toward most significant bits. For example turn
>> 7 = 0111 to 14 = 1110.
> 
> Arithmetical operations on 16 or 24 linear audio file produce rounding
> errors in the Least Significant Bit.
> Normalising is an example of this type of operation. It does not increase
> the resolution of your file -- it boosts the noise with the signal and adds
> more noise. It is often suggested therefore,  that you do all your
> processing of the file -- FX ,EQ, compression, reverb etc and boost only in
> the final stage to get maximum level  to take best advantage of the limited
> resolution of the CD format (16 bit linear, 44.1KHz sampling).
> 
> The internal processing of Logic's native effects and VST effects etc is
> all at 24 bit float so the absolute level of the file that undergoes that
> processing doesn't matter beyond the usual caveat that the highest
> resolution is obtained by recording as near as possible to full scale
> without clipping.

-- 
GO SEE 
http://www.teddybut.com and http://www.mp3.com/teddybut
Kumpelstiltskin Music, Inc./EMI Music Publishing
Burning Beagle Studios, Brooklyn, NY

> 6 dB is twice as loud. Two violins are 3dB louder than one violin, 
it takes
> four to be twice as loud or 6 dB louder.


6dBm [or i think dBV] is twice as much SIGNAL
but twice as loud to the ear [accoustically] is 10dB spl

though of course the dB is not a fixed amount.
Noticable relative changes in level vary depending on the actual 
loudness and frequency. Actual loudness should be measured in phons 
and sones.
But who cares?
Most rock clients will notice it if they're song appears quieter than 
someone elses.

Maak Bow

At 05:59 AM 20/11/01 +0000, Maak Bow  wrote:

quoting me:
>> 6 dB is twice as loud. Two violins are 3dB louder than one violin, 
>>it takes four to be twice as loud or 6 dB louder.

Replying to me:
>6dBm [or i think dBV] is twice as much SIGNAL
>but twice as loud to the ear [accoustically] is 10dB spl
>
>though of course the dB is not a fixed amount.
>Noticable relative changes in level vary depending on the actual 
>loudness and frequency. Actual loudness should be measured in phons 
>and sones.

Hi Maak ,

I fished out the old textbooks to look this up -- getting rusty I guess.

You are quite correct that a change of 1 bel or 10 dB is "twice as loud" in
terms of our perception. My mistake.

A 3dB change is a doubling or halving of power . eg If you switch off one
channel of a stereo system playing a mono signal the volume drop is 3dB.

A doubling in signal voltage is a 6 dB change -- that's where I got confused.

Regards,
Murray

> I fished out the old textbooks to look this up -- getting rusty I 
>guess.

Yeah , i had to think back a while to remember all that dB stuff too

Maak Bow  [dip Aud Eng] <Hahahaha