In the digital world, what you want to avoid is processing your audio file too many times because each time, you risk to degrade its quality.
Recording close to zero db was a thing of analog consoles and tapes. Its not necessary today, and in fact I'd advise against it.
For a full band I usually record so each tracks peaks between -16db to -12db. The reason is; the sum of those tracks will be adding up on the master bus of the daw and normalise tracks will make the master bus peak over zero db. The result will be that I would either need to lower the faders or the gain of each tracks (which adds processing). Keep in mind that it is recommanded to send files that peaks around -9db to -6db to the mastering engineer.

When you mix a song it is NOT THE TIME to make it louder !

You're -18db track is perfert to work with like it is. If you must, adjust the gain a bit...

Awesome response! I'll keep'er right where she is!

Thanks a lot pcrecord

-Chuck

Also... you shouldn't dither if it's going to a mastering facility, and if you are the one doing the mastering, dithering should be done as an absolute last step - and is added only to the true master audio file, image file or glass pressing (in the case that a glass master is made for quantity CD duping).

The reason I am mentioning this, is because - as short as just a year or two ago - there were still some DAW platforms - usually previous or legacy versions of software, where the dither defaulted to being on, and if you weren't paying attention, you could end up dithering each time you did an export, or a print-to-file as a .wav, as well as each time you would do a bounce, comp, sum, freeze or track merge. The result was that the audio was being dithered, over and over again, each time degrading the signal further and further.

I think most of the current DAW platforms have switched their default dither settings to being "off" but, just to be sure, you should check your DAW's system settings, and make sure the dither default isn't set for dither to be on each time you render, bounce, or freeze an audio track or audio file; especially if you are using a platform that's older than a few years.

FWIW

d.

To add a bit on level, the general rule of thumb for tracking is:

Set your AD converters to 0 dBvu = minus 20 dBfs and record your tracks at an average level of -20 dBfs and everything will fall into place.

I usually never move up in level until mastering time which Marco put so well.

Thanks all! :^)

FWIW, I don't normalize much...

Even before the extended bit resolutions came about, I still didn't do it, as I felt that - beyond my experience that has shown me that it's not really doing anything "positive" - it also has a tendency to wipe out the dynamics if you're not very careful, and that's kind of the operative word there, because may people who use normalizing don't really know what they are doing with it, and because they aren't familiar with other processing that can accomplish the same thing, they then use it to extremes.
Rookies and newbies generally like to reach for "that one button", either because they like the convenience, or because the principles of gain expansion and reduction elude them and they don't want to take the time to learn better ways to do certain tasks.

Basically, I guess that my own thought is that Normalizing isn't necessarily bad, per se', at least not in the hands of someone who knows how to use it.... it's just that there are other - and better methods - to increase/decrease gain.

I almost always used compression or parallel gain reduction instead, as a way to bring softer parts up while keeping the louder transients in check.

IMO

d.

I normalize when I am doing critical track comparisons. Its the best way to hear change. I normalize to top off the transients full scale, line the peaks up so they match and do nulls , then copy paste a seamless 2 bar loop between the two and listen like a dog as they loop one after the other to hear any sonic differences. I never stop audio when doing comparisons. I always loop two at at time, one after the other on the same track created just for the comparison.
I will flip the phase and use critical EQ and volume to find the freq's that are remainders in a null. Normalization is a good thing for that!

audiokid, post: 426711, member: 1 wrote:
I normalize when I am doing critical track comparisons. Its the best way to hear change. I normalize to top off the transients full scale, line the peaks up so they match and do nulls , then copy paste a seamless 2 bar loop between the two and listen like a dog as they loop one after the other to hear any sonic differences. I never stop audio when doing comparisons. I always loop two at at time, one after the other on the same track created just for the comparison.
I will flip the phase and use critical EQ and volume to find the freq's that are remainders in a null. Normalization is a good thing for that!

...Although not exactly how the average "engineer" would use it. ;)

Exactly. Most use normalization to increase lack of gain. Which I am like you, I think there is a bit of mojo in the white space lol, that has more to do with overload at summing time.. Plus, by the time you get to the master bus, the last thing I want to be doing is "reducing" gain . It should be right around -6 on the master bus . I always think mixing is good when you don't have to go backwards. Always one way to the finish line. >>> Seems to always sound as expected that way.

I tend to mix a bit lighter than that, generally shooting for peaks of around -8db or so, but then again, I don't handle the mastering end like you do, either, so I generally like to give whatever M.E. I'm going to be working with as much room to work as they feel they need to do the best job.

The -6 level may be the standard that most M.E.'s are looking for when they get the 2 mixes, although I've never been told by an M.E. ( at least not a real M.E. (LOL) that at -8db, my mixes were "too shy".
If anything, I've been complimented by them for not sending them a mix that has been pushed balls to the wall, right up to -1. ;)

Generally, I like to shoot for track recording levels around -14 to -10 or so, and then on the master output, around -8 on the peaks. There may have been a few times where I have pushed that to -6db, but certainly never any hotter than that.
For me, -6db is the absolute ceiling on the master output.... and I'm not talking about using that peak value with an RMS of -10, either. LOL

imho .... normalizing is like beating your wife and children, one of those things you can do, but shouldn't. if tracks are recorded well, we don't need it. not rocket science. in fact it's so much simpler now. set peaks to -18 or so and hit the red thing. old school was shoot for -3 and cross your fingers ..... snarf!