Song Dynamics

[jaybriggs84]

I'm recording a project with my friend and he mentioned about certain parts of the song being different volumes. What ways are there to preform this other than a write automation?

 

[studiomaster]

Compression.

 

[mixsit]

There's no way from your post whether you want more dynamics or less.


These days though, lem' me guess..

 

[TelePaul]

compression is one of those things that is kinda maxed out these days...Its compressed on the way in, when ya hear a song on the radio, its compressed...what I tend to do when I've finished a mix is normalize then add a broad ranging compression plug in...though i think full automation and part editing might better serve your purpose if theres alot of this going on.

 

[studiomaster]

Paul, do you use the normalizer on the Master bus as insert? Will it be better if I use Wavelab to normalize the entire .wav file instead?

 

[NYMorningstar]

A sure way to kill your song is to normalize. Instead, you want to look at the project and ask yourself why you think it needs to be normalized and then fix it.

 

[TelePaul]

A sure way to kill your song is to normalize. Instead, you want to look at the project and ask yourself why you think it needs to be normalized and then fix it.

I just feel i HAVE to normalize when I need to convert to MP3...seems its way too quiet for posting on the net otherwise! Im not sure I get what ya mean by killing the mix but I'm pretty new to this, a friend recommended it to me as the last step in mixing...when the tracks done, i import it as a new audio file, normalise and add a compression/dynamics insert...like I said, I'm new and it kinda seems to work okay for me.

 

[TelePaul]

Paul, do you use the normalizer on the Master bus as insert? Will it be better if I use Wavelab to normalize the entire .wav file instead?

Uhmmm I'm not confident enough to give ya a definitive answer! But yeah i normalize and add compression as an insert with the final .wav file.

 

[bdemenil]

normalizing doesn't affect the dynamics of the mix. All it does is increase the volume of whole whole mix until the loadest part hits 0db. It's a very good idea to normailize any final mix. Of course, if your mix needs a lot of boosting, you should do it in the mix rather than waiting until after mixdown. Normalizing raises the noise floor as well.

 

[NYMorningstar]

It's not a good idea to normalize a final mix especially if you are going to have it mastered. If you're using it to regulate the volumes of your whole album you should stop and use your own ears. Normalizing is an unnecessary step that doesn't change the signal to noise ratio and adds quantization distortion.

 

[Kevin Deschwazi]

It's a very good idea to normailize any final mix.

Bad advice, NYMorningstar has it right.

 

[TelePaul]

It's not a good idea to normalize a final mix especially if you are going to have it mastered. If you're using it to regulate the volumes of your whole album you should stop and use your own ears. Normalizing is an unnecessary step that doesn't change the signal to noise ratio and adds quantization distortion.

so when would you use it? I do it when ive finsihed mixing the entire track.

 

[Kevin Deschwazi]

so when would you use it? I do it when ive finsihed mixing the entire track.

When you're in a rush and the quality of the audio isn't that important (radio ads, background music for corporate events etc.).

 

[TelePaul]

what exactly does it do then? haha seems funny, when Im in a rush id add a step.

 

[mixsit]

what exactly does it do then? haha seems funny, when Im in a rush id add a step.

It just turns it up. But it generally does so by permanently processing the file (unless you do it to a copy).
But if this is simply about getting the volume up where you want it, why not just look at peak hold the last time you ran the song and add the gain at one of the volume controls you're all ready using? Or if stray peaks are keeping the volume down, control them with a limiter and set it's output to -.5 ish' (or -.3 or what ever bit of spare you like) and be done with it.
So yes it can be a redundant step.

I say we re-name it-
'Abnormalizing'

 

[Kevin Deschwazi]

what exactly does it do then? haha seems funny, when Im in a rush id add a step.

A normaliser scans the source, finds the loudest peak and raises the overall volume until that loudest part reaches a predtermined level (often 0dB). So everything gets increased by that amount (including the noise floor).

I guess it's quicker to click normalise than to use those long slidey things.

Normalise isn't an extra step as such, on most recordings there will be steps taken to get more volume (usually at mastering) but normalisation is a particularly destructive way of doing that compared with other ways of raising volume.

Here's a Bob Ohlson quote on the subject:

"Performing extra math on digital audio only degrades the quality. It's a misguided practice left over from the days of 8 bit samplers ...."

http://recforums.prosoundweb.com/index.php/mv/msg/1507/0/0/0/

 

[bdemenil]

normalizing not destructive

normalization is a particularly destructive way of doing that compared with other ways of raising volume.

Normalizing - used in a reasonable manner - does not distort audio. It applies a very simple calculation which is basically identical to pushing up the master volume fader to the highest possible level without having any part of the song clip. I can't think of another way of doing this which is less destructive (maybe pushing up the faders on individual tracks prior to mixdown - but negligible difference).

Normalizing is particularly useful and particularly harmless when you are starting with a file with high bit resolution and will end with a file with lower bit resolution. Say for instance you mix down a 24 bit file. Then you normalize it, then you dither it down to 16 bits. The normalizing in this case will actually reduce your noise floor - because it insures that the full 16 bits are used to encode your music (if the music is not at its max sub-clipping amplitude, bits are being wasted). In the same way, you could mixdown to a 32bit file, normalize, and then dither to 24bits.

If the mastering studio is going to master digitally with a 32bit file, then no reason to normalize. If they are going to use a 24bit file, or if they are going to go analog, normalize - because you want to use full bit capacity, and i don't think any DA converters handle 32 bits.

Of course, extreme normalization will raise noise-floor. If normalizing is raising your mix 30dbs for instance, you're probably cutting too much volume on the tracks in the mix. Any situation where you're cutting volume and then boosting it again will result in more noise - though in small amounts this is negated completely by the reduction from 32 or 64 bit to 24 or 16 bit.

If you're doing things on the cheap and mastering yourself (and this is homerecording after all), then normalizing has an additional benefit. It helps with the mastering because if you normalize every track before mastering you will have more consistent levels going into the mastering - which means it's easier to guess a starting point for compression thresholds.

The worst thing about normalizing is the nightmarish name - who wants to be forced to be normal!?