normalizing before or after tracking

[jeromejello]

so i am tracking out this mmw show i caught on my jb3 the other day and i was thinking about raising the volume a tad (normalization, right).  my question is should i normalize the full wav of the whole set or can i normalize each chunk that i tracked out.  is there a difference is quality or does it cause any additional defects in the sound this way.

thanks.

[cleantone]

Your gonna want to normalize the full file before splitting it...

is there a difference is quality or does it cause any additional defects in the sound this way.

Some would argue that any time you do digital processing to a file their would be artifacts or somthing but for the most part your okay with this. If your only bumping it up .5db I wouldn't bother.

[SparkE!]

Always normalize before tracking.  Otherwise, some tracks will be amplified more than others and you will get jump discontinuities between tracks.

[§†∑∫åµÞ≥¥]

If you're using Audacity, I would recommend "Amplify" if you're trying to boost levels.

[jeromejello]

thanks for the info.. i kinda thought it would be better before.  i chose not to do anything to it.  if you are interested in checking it out, here's the info:

mmw 3-12-05 house of blues, orlando, fl
hop on if you can.

not sure about my upload capacity, but i will leave the computer on until i get a couple of other seeders.  feel free to add any comments its my first time for taping/seeding.

http://bt.etree.org/details.php?id=12476

enjoy.

[heath]

Always normalize before tracking.  Otherwise, some tracks will be amplified more than others and you will get jump discontinuities between tracks.

qft

[Chris K]

im not a big fan of normalizing. i prefer to adjust gain, or amplify as stated earlier.

as for tracking and normalization. if you normalize the entire set or show (if one set) after you track and PRIOR to splitting the tracks then you should be ok.

ymmv

[jaguaracer]

im not a big fan of normalizing. i prefer to adjust gain, or amplify as stated earlier.

as for tracking and normalization. if you normalize the entire set or show (if one set) after you track and PRIOR to splitting the tracks then you should be ok.

ymmv

yes, it sounds like there may be a little bit a confusion in this thread.
There is gain which will boost every single second by a constant rate of dbs.
Normalizing is the process of 'smootihing out' the sound. ie, it will raise the real quiet parts and soften the real loud parts.

[SparkE!]

im not a big fan of normalizing. i prefer to adjust gain, or amplify as stated earlier.

as for tracking and normalization. if you normalize the entire set or show (if one set) after you track and PRIOR to splitting the tracks then you should be ok.

ymmv

yes, it sounds like there may be a little bit a confusion in this thread.
There is gain which will boost every single second by a constant rate of dbs.
Normalizing is the process of 'smootihing out' the sound. ie, it will raise the real quiet parts and soften the real loud parts.

No, normalizing is amplifying the whole thing by an amount that will cause the largest sample to barely hit either the top rail or the bottom rail.  What you're talking about is compression.

[jaguaracer]

I thought I was talking about this:
normalizing simply means to adjust the peak (top and bottom) volume of a selection to a known value.
Normalizing increases voume 'across the board' too? 
Normalizing has does two meanings:

Normalizing, as far as Sound Forge or other digital audio editors are concerned, simply means to adjust the peak volume of a selection to a known value. Generally the recommended maximum is -0.5 dB (I think that's 94.49% or something). Doing this is a no-brainer.

Normalizing a set of tunes to be burned to CD, however, means something slightly different. Here it implies that you're adjusting the average volume (ie, 'across the board')of those songs so that they will all sound about equal. Doing this is an art (whereas we can consider mastering a fine art).
-- Dragon

http://www.homerecording.com/normalizing.html
It sounds like you are talking about the later case, no?
edit: yes, I think my original term of 'smoothing out' can be interpreted as compression but I think I'm talking about the first type of normalizing. I'm sorry for the confusion. I should have been clearer.
edit 2: hmmm. although here's another definition which seems more in line with what you are saying:

Normalizing increases the level of the entire sound file so that the loudest part of the sound is at the maximum playback level before distortion; it then increases the rest of the sound proportionality.

[momule]

Are we referring to the peak value or the RMS value..

It seems kinda crazy to me to Normalize the entire file (set) as your not getting a accurate values.. Some songs are written to be softer than others.
say you had a two track Acoustic set as the opening tracks there gonna be much lower in RMS value than the amplified tracks.

I open the master in SF7
track: (to give regions to each track)

master:  Eq,fades,Normalizing, ect..
you can double click on the track/region and SF will highlight just that track.
then I can normalize just that track. ( I use -16db on most recordings and apply Dynamic compression if clipping occurs)  But depending on how it sounds it may go higher or lower.

resample:  using the highest Interpolation setting and the anti alias filter .

extract the regions : (export separate tracks)

and wammo you have one set ready to flac
YMMV
Nick

[Brian Skalinder]

Are we referring to the peak value or the RMS value..

If necessary - which is rare - I'll amplify the entire file set so the single most peak value approaches 0.

It seems kinda crazy to me to Normalize the entire file (set) as your not getting a accurate values.. Some songs are written to be softer than others.  say you had a two track Acoustic set as the opening tracks there gonna be much lower in RMS value than the amplified tracks.

On the contrary, normalzing the complete fileset is the only way to ensure accurate relative values.  Amplifying on a track-by-track basis is inaccurate and destroys the original dynamic range of the recording.

[jeromejello]

Are we referring to the peak value or the RMS value..

If necessary - which is rare - I'll amplify the entire file set so the single most peak value approaches 0.

It seems kinda crazy to me to Normalize the entire file (set) as your not getting a accurate values.. Some songs are written to be softer than others.  say you had a two track Acoustic set as the opening tracks there gonna be much lower in RMS value than the amplified tracks.

On the contrary, normalzing the complete fileset is the only way to ensure accurate relative values.  Amplifying on a track-by-track basis is inaccurate and destroys the original dynamic range of the recording.

bingo, this is the info i was looking for (as a confirmation of what i thought).  i would rather not do anything to the wav if possible (and definately nothing that i would seed - too many people get picky about it)

[momule]

If necessary - which is rare - I'll amplify the entire file set so the single most peak value approaches 0

On the contrary, normalizing the complete fileset is the only way to ensure accurate relative values.  Amplifying on a track-by-track basis is inaccurate and destroys the original dynamic range of the recording.

Why not just raise the Amplification or Volume of the entire recording .. seems as it would do the same thing.

I think as long as you dont Raise and lower each track by huge amounts the dynamic range will still be intact.. (Such as dont normalize one track to -20 and another to -5) .  I usually dont have but a db or two difference in each track . which is about where it was to start with..

Seems to my ears that Normalizing the entire "master wav"  is "sorta" like selecting the Normalize all files in your burning program, as your gonna raise the Loud parts but the softer ("acoustic" if ya will) of the master are still prolly gonna be below a comfortable listening Volume..  and you are missing the fine detail of the recording. (such as the plucking of the strings)

I use a general rule of thumb that you should never have to turn your volume knob on your stereo over Half way or you simply need more power or in this case more db ..  Seems like the same argument here.
Most the AUD tapes I hear are VERY  low and you need to crank your stereo to get a decent volume level .
After watching a few tapers in the last year or so at shows I have taped,  I find that most folks seem to be afraid to run there Pre a bit hot , for instance At the last show I was taping at the guy next to me Had his Ua-5 set so low that the clip light never came on The entire show. 
Which would lead me to believe there recordings will suffer due to loss of the original dynamic range at the show that night..

For instance One the most recent tapes I pulled came in at -26db , Thats crazy low.
that means you need to crank your Stereo to compensate for the Low recording volume. 


Still confused
Nick

[Brian Skalinder]

Why not just raise the Amplification or Volume of the entire recording

Sorry for the confusion, that's what I meant by normalizing the entire fileset - perform the operation against the master recording all at once, not individual tracks or files.

I think as long as you dont Raise and lower each track by huge amounts the dynamic range will still be intact.. (Such as dont normalize one track to -20 and another to -5) . I usually dont have but a db or two difference in each track . which is about where it was to start with.

I think your comments are a little mixed up:  in order to NOT raise or lower each track by huge amounts, we must utilize different RMS values for each track in order to maintain relative dynamic range, i.e. do exactly as you suggest we should not:  normalize one so it's RMS is -20dB and another to -5dB.

Let's take an example:  We have two tracks, [1] with an RMS of -21dB, [2] with RMS of -9dB.  That's an RMS difference of 12dB across the two files.  Now, let's normalize each track independently to the same RMS value, -8dB.  That entails raising the RMS on [1] 13dB, and raising the RMS on [2] only 1dB.  So now, both tracks have an RMS of -8dB.  The RMS difference across the two files is now 0dB.  Dynamic range:  gone!

In this case, in order to preserve relative dynamic range - but still increase amplitude - we would want to normalize each track to very different RMS values, say increase [1] so it's RMS value becomes -16dB, and [2] so it's RMS value becomes -8dB.  This brings the amplitude of track [1] up to a better level for listening (without having to adjust volume during playback).  The dynamic range is reduced, but not removed.  However, I believe there's a better way to "amp up" the recording but still maintain relative dynamic range...

I use a general rule of thumb that you should never have to turn your volume knob on your stereo over Half way or you simply need more power or in this case more db .. Seems like the same argument here.

The way to accomplish this, IMO, is to apply compression to the entire master WAV and then raise the dB across the entire master WAV.  That way, you maintain the relative dynamic range, if not the actual dynamic range proper.