I understand what you are trying to say Cakey2, but normalizing after mastering is redundant if you are mastering correctly.
Generally in mastering, you eq and compress (if you need to) the one song that you want all others to sound relavtive to on the finale CD. All other songs are then adjusted to be percieved as being as loud as and tonally the same as your reference tune. When you are done, you will find that all the songs will be "percieved" as being the same volume, an important distinction because different songs will sound the same volume with quite different RMS average levels. It is that Fletcher/Munson Relative Loudness Curve thing again...
Another reason that I don't recommend Normalizing after mastering is that the level changes made during Normalizing will create extented bit length which will produce quantiniztion errors during the D/A conversion. Dithering may need to be applied to hide the bad effects. But where the problem starts is that the mastered song was probably already dithered once, and applying another dither to it will produce an obvious "mask" to the sound.
So, you COULD wait to dither after normalizing, if your normalizer has a decent dither scheme that you can apply. But, normalizing after mastering is just a unnecessary step if you are following good mastering techniques.
If let's say that you want to master a tune for several different song combinations, like one for a compilation, one for all the songs by an artist, etc....then the same song probably needs to be mastered each time to fit with the other songs on the CD. Makes sense right?
Mastering is about getting all the songs on the same CD to sound somewhat relative to each other. The same song would need to sound quite different if combined with different songs depending upon the combination of songs that is desired on any particular CD.
Anyway. It all still comes back to comparing all the songs on a CD to a reference song that will be on the same CD. You start by mastering the reference to sound good, then make the other sound somewhat similar.
When you do this, and if you check out the wave form of many professionally mastered songs, you will find that in a lot of cases, you will never get to 0db at all! Yup, a lot of the best sounding masters I have heard peak as much as 2db below 0db and usually have an average RMS that is in the -10-15 range.
Now that is not to say that I have not seen CD's mastered that have an average RMS of around -6 (wow! not a lot of room for dynamics, but great for metal CD's...
)and where 0db is reached all the time. But hey, I have a CD here mastered by Brian (Big Bass) Gardner at Bernie Grundman mastering that has digital distortion on it!!! Yup, a $1000 mastering job by one of the best in the biz (for big low end stuff at least) that has distortion. Oops. On cheaper playback systems I never noticed it before, but a close listen on my studio monitors reveals this. Boy, there is no way my CD's are ever going to be that loud mainly because I don't have the mastering gear to pull it off. But, I have noticed that a lot of the best sounding CD's hardly ever reach the 0db limit. Tracy Chapmans first CD comes to mind, and it is a fine sounding CD, and sounds great on the radio too. So what that it's quieter then Metallica's newest, who cares? It sounds more natural.
So, the point I am trying to make is don't get all caught up in the volume game here. Many of us do not posses the necessary equipment to compete with the big mastering houses anyway, so you will just be hurting your music trying to get it as loud as the big mastering houses can. I am using Wavelab with Waves and
Steinberg Mastering Edition plug in's and it just does not sound as good as the stuff another local mastering guy uses (Weiss digital eq's and compressors, Sadie mastering software, etc...) but, I don't have that kind of money to spend either on mastering gear. So, I live with the shortfall. It is okay with me, the stuff I have works reasonably well, and match's the quality of alot of the product that my humble project studio produces. If the big time game was my aim, I would need the better gear, bottom line. Yes, in time I will learn to maximize the use of what I got but at it's best, but it still will not work as well as the better gear. Of course by the time I get the maximum use out of it, the better gear will just get better and I will be that much farther behind the game...
Oh well.
I am sitting on most of the submissions for the Homerecording.com Compilation CD here. There is stuff here submitted that maxed out and still doesn't sound as loud as songs I have mastered that don't have as high of RMS level. Some of it has very obvious compression that is affecting the songs ability to "breath". Transients are cut off. The "air" is lost. In fact, all the stuff I master also is usually cut at -.3db, so it is not even at digital 0db! So, normalize all you want, it is still a lousy way to increase volume, and not a very accurate way to produce a good relative loudness from song to song.
Peace.
Ed
[This message has been edited by sonusman (edited 05-04-2000).]