A Simple Mastering Question (regarding low output)

[digitaldoom]

Background:
Hey I've written alot of tracks using things like Cubase SX, FL Studio & Reason, but one problem still plagues me.

When I've finished adjusting all the levels in the software, I normally output to WAV.

The Problem:
The volume of the WAV seems very low. I read up on the benefits of normalization and so I've tried that with Sound Forge. With some of my FLStudio and Cubase Tracks I've even ran Compression plugins (like: BBE Sonic Maximizer) on the master tracks.

So I was thinking that this problem could be solved through normalization and compression, but somehow it's not working. Well... I feel that it's not working because when I play my finished track next to a professionaly produced track... mine still sounds low.

Am I just using the wrong normalization or compression settings?

Can anyone help or offer some suggestions?

[vurt]

using the wrong settings mate
im afraid i cant help much more than this as i dont know your actual music but

in cubase say are you on the master mixer using any mastering tools such as compressors maximisers limiters?
you need to boost the output there first before normalising in soundforge(which luckily has a normalise preset for music )try this preset then try soft limiting it (under dynamics in the effects dropdown) then normalise again and repeat the soft limiting

this is what i generally do and for me personally it does the job,but its horses for courses really

[jouni]

Do not use Normalization on a finished track. You just ruin the levels. You'd better use loudness/maximizer plugins for that. However mastering is more than just raising the levels. Mastering involves EQing, compressing (multiband), fixing the stereo image (multiband usually), exciter is good effect as well to bring brightness.

I'm sure this forum is full of mastering tips.. try to look for those.

[Shane Sanders]

Do not use Normalization on a finished track. You just ruin the levels. You'd better use loudness/maximizer plugins for that. However mastering is more than just raising the levels. Mastering involves EQing, compressing (multiband), fixing the stereo image (multiband usually), exciter is good effect as well to bring brightness.

I'm sure this forum is full of mastering tips.. try to look for those.

Jouni! When did you come to KVR?

[Crackbaby]

Before even thinking of mastering .. fix the problem in the sequenser. If you normalise a wav with low level, you loose all the dynamics.

Um.. how low is "low"?

If you look at the wav in an editor, is it half full? More or less? (from top to bottoom of the file)
What is the output in the vstmixer?
If you put a maximiser on the master insert, is the volume still low? Is the master .. the ...the what is it called in english..........:
1... |
0 --|--
-1.. |
-2.. |
-3.. |

/\
|
|
That i mean. where is your --|--? On the masterchannel it should be at zero.
(the "." is there because this forum cannot handle multiple " ")
Are you sure you are saving in the right format? I remember that i once accidently changed the format, everything rendered was distorted (some kind of windows wav, not pcm).


Good luck

[Jeremy_NSL]

Normalizing peaks won't 'ruin' the dynamics. It simply raises the volume of each sample so that the loudest sample is 0db. The problem with this is it raises the noise floor as well. So if you normalize a very quiet recording to 0db it will be quite noisy.

If your levels are very low before normalizing, its best to go back in the sequencer, raise your volume, and re-render. If you end up having to normalize a few dB, there shouldn't be a whole lot of noise added.

[Crackbaby]

um.. hmkay?

If you have a 16-bit wav, and you only use 8 of those bits (half full as i wrote) you will have only half of the dynamics. Therefore the difference in db will be cut in half.

IMHO

[garret]

I took a listen to your track Tranced. Honestly, you're not doing that badly. You just need a few more dB to be respectable. Here's my quick and dirty technique... it's probably not going to please purists, but it works for my purposes.

Get a copy of DFX RMS Buddy. This is an RMS loudness measurement VST plugin.... It's suuuuper useful for quick mastering work.
http://www.smartelectronix.com/~destroyfx/extras/

Make a final mix of your track without any mastering plugins (no multiband compression or limiting or maximizers). Try to set levels so your loudest peaks on the output bus are closing in on 0db, but don't kill yourself. If you only make it to -6db, that's fine... you can fix that with a bit more gain at mastering time. But if you find your mix is way low (like -12 db peaks), push every track up in volume and try to get closer. Don't clip, of course...

Export the mix to a wave file.

Start a new project, and import your mixdown wave file.

Add RMS Buddy to your output bus. Open the plugin window, and you'll see left and right numbers for Average RMS, continual RMS, and peak. They'll be blank until you start playing your track.

Play the track start to finish. When the track finishes playing, the peak numbers will show how far below 0db your track is. Hopefully the left and right channels are similar. If not, you should go back to the mix and correct things to get the balance right.

If the peaks are at -2.5 db, then add about 2db of gain to the track. You don't need to get to 0db, just close... you don't want to clip, so I'd leave a little bit of headroom.

Reset the RMS buddy counters, rewind, and play your track again from start to finish. Watch the average RMS this time. That's a good measure of the loudness of the track...

You might want to aim for -15 to -17db. That's loud enough that your track will sound respectable next to a commercial recordings, but not so loud that things will distort. Your track Tranced measures -20.5 db. So it needs 3-5 db of extra loudness.

The easiest way (not always the best) to get it is to put a limiter on the output bus. (Put it before RMS buddy in the effects chain, so you can measure its effects later).

TLS Maximizer is a good freeware limiter, and the interface is dead simple:
http://hem.bredband.net/tbtaudio/tls.html

In TLS Maximizer, drag the gain increase slider to 4 db, reset your RMS Buddy counters, rewind, and play the track start to finish.

See if you got your RMS closer to your target range, and if the track still sounds good.

-Garret

[BONES]

If you have a 16-bit wav, and you only use 8 of those bits (half full as i wrote) you will have only half of the dynamics. Therefore the difference in db will be cut in half.

So normalising doesn't stuff the dynamics, recording at a low level does. Fair enough but not the same thing.

digitaldoom, you should render your songs out at 32 bits so that you still get full dynamics even if the output volume is low [higher bit-depth gives exponential benefits]. If you've got SF5 or above you should have the WaveHammer DX plugin which is a great way to maximise the volume of your songs. It has plenty of presets so you should get results easily. I usually normalise to about 90% before I run it through WaveHammer so that the results are always similar. It works a treat.

[digitaldoom]

I really appreciate all of the responses! There is a lot of great information here. Thanks for all the links to free software to help, that's awesome! I'm going to take a more detailed look at the levels of the mix in the respective application. In all honesty, I haven't taken a close enough look at my db levels. I will do that now I have some information to follow. Also I tried in the mix to turn up the volume on each instrument while still keeping the overal volume of the track as high as possible without clipping. That has always been my goal. It's just that I never knew what to do afterwards. I also wasn't sure why my songs sounded consistantly lower, even though I had thought that my volume in the application would go as high as possible without clipping.

Now I have some cool suggestions to try out

[digitaldoom]

BONES: Why would I render to 32bit if I was going to burn to a 16bit Audio CD?

[Hypertone]

Because when you run your mix through eq, limiters and compressors, 32 bit will give you much better sound quality from the additional headroom.

You should get your mix sounding as good as you possibly can before you export. Compressing a mix won't give you near a s good results as compressing your individual tracks-- drums, bass, etc. individually in the mix. You can do so much more when you have the individual tracks to work with, think of a mix as the ultimate multiband compresor.

Make sure it sounds nice and loud before you export a wave, and when you do export, try to get the peak levels close to 0db, -3 or better is right on the money. Do not use normalization. Use a limiter on the mixdown to get it to 0db. And remember that most pop music we hear today is overly compressed. Louder is not better. Draining out the dynamics of a recording can make it fatiguing to listen to.

[BONES]

Why would I render to 32bit if I was going to burn to a 16bit Audio CD?

Pretty much what he said. Working in 32 bit just gives you more room to manouvre. When you are finished you just normalise and convert it to 16 bit.

Compressing a mix won't give you near as good results as compressing your individual tracks-- drums, bass, etc. individually in the mix.

You know, whenever I apply a compressor to an individual track I never get any sense that it is anything but louder or softer. If I balance the levels I can bypass them and not hear any difference in 99% of cases. I just don't get it.
Rich has just added a Bass Compressor into the latest ORION beta and it is the first time I've used a compressor on a single track and felt that it was doing something. It's awesome. I can always perceiove the effect of compression on a mix and sometimes on a vocal track but almost never on a single instrument.

[Crackbaby]

You know, whenever I apply a compressor to an individual track I never get any sense that it is anything but louder or softer. If I balance the levels I can bypass them and not hear any difference in 99% of cases. I just don't get it.

I know what you mean. But for some reason you will (..i do) hear the difference when you play back the whole track. Its all those little things on each track that makes it.
If the recordings are excellent tho, you need less compressors on each track.

IMHO

Hehe .. fair enough

garret: Excellnt! Superb! That was some great instructive writing

[BONES]

But for some reason you will (..i do) hear the difference when you play back the whole track. Its all those little things on each track that makes it.

No, that's when I bypass and hear no difference at all. I never compress any tracks, except with this new bass compressor, and I'm completely happy with my final mixes.