Normalization

[Rusty K]

Hello,

I'm sure this is an often discussed topic. Generally I haven't used normalization to this point but I've noticed that my masters are not playing well on certain commercial players. Some tracks seem to have slight distortions at max volume.

I'm assuming it is true that if you run the volume at max on a commercial player you might hear the speakers crapping but you still wouldn't want to hear distortions that are coming from your track?

With this in mind I have a couple of questions about normalization and DC bias.

1. Should I wait to normalize till I'm mastering i.e. the last process in the
chain?
a. If yes to what db should I mix and master? Also as a little
sidebar to this question. I track at 48000 usually. I've gotten
as much as +.5db bounce in volume whenever I resample and
convert (with dither) back down to 16bit 44100. I've posted
treads about this before with no real answer as to why this is
occuring. Nobody else seems to have encountered this.


2. I've noticed that I'll get some DC bias at mixdown but my mastering
sofware always centers the wave automatically. Should I bother with
DC bias adjustment at earlier processes?

Thanks,
Rusty K

 

[mshilarious]

1. Should I wait to normalize till I'm mastering i.e. the last process in the
chain?

Yes, but instead of thinking of it as normalizating, you will want to consider the process of mastering. Are you going to use a limiter? If so, normalization is an unnecessary step. Are you mastering a whole CD? Perhaps every track might not peak at 0dBFS.

a. If yes to what db should I mix and master?

You should mix wherever you have tracked, and if you track properly, you will have plenty of headroom for mixing. -6dbFS peak is really a bare minimum for a raw mix, it's easy for a seemingly simple EQ change, including cuts, to eat up 3dB of that. There is no penalty for working even quieter than that, although you should set your monitoring levels appropriately.

Also as a little
sidebar to this question. I track at 48000 usually. I've gotten
as much as +.5db bounce in volume whenever I resample and
convert (with dither) back down to 16bit 44100. I've posted
treads about this before with no real answer as to why this is
occuring. Nobody else seems to have encountered this.

If an SRC does that, I'd look for a new SRC. That's a pretty big swing. Dither can change the average level, but it should be very slight. It shouldn't ordinarily cause clipping, at least of any significance.

2. I've noticed that I'll get some DC bias at mixdown but my mastering
sofware always centers the wave automatically. Should I bother with
DC bias adjustment at earlier processes?

It's worth investigating why there is DC offset. Do you have DC offset on tracking? If it's small, I wouldn't worry about it until mastering. After seeing one bizarre example of wandering DC offset on this board, I now use a 10-20Hz high pass rather than a DC offset tool.

 

[Rusty K]

mshilarious,

Thanks for your attention. Now that I have it...I'm going to pump you for info...ha!

The mastering process has been killing me. I A/B my results with professional tracks in my editor. I find the visuals helpfull. It's actually amazing the wide array of visual results one finds in otherwise good sounding CD's.

I'd like to be able to push the volume envelope like many of the pros do but I don't want to start a discussion about the pros and cons of "squared waves"....it depends on the song. I do usually use compression when mastering. I have a master limiter but I'm not really sure how best to use it.

One thing that's driving me crazy is that I'm getting clip like distortions when referencing with my headphones. These are not audible with near field monitors and there is no actual clipping. It has been suggested this could be that running my soundcard trim at -4dBv means my headphone amp in not able to handle the level. The distortions do in fact disappear when I change the setting to -10dBv however I can reference a track from a pro CD mastered to 0db and not hear the headphone distortion. This really has me confused and distroys my confidence in my process. If it weren't for the fact that I can hear some of this distortion on cheap boom boxes etc. at high volume I'd ignore it....but I digress.

I had been mixing to -1db but I've cut that back to -3db considering that perhaps no "headroom" was an issue with this problem. I'm still getting the headphone stuff. Even with the output section of my mastering software turned back to -1db I'm getting the distrotion. I could be pushing it too hard but compared to some of the "push" I see on pro CD's I don't think I'm overdoing that much. I haven't tried -6db but I will.

SRC?...software conversions? or hardware?


Most of my DC bias is minimal so I'll just let my mastering sofware level the wave in the master.


Maybe I'm expecting too much from my equipment. It's a home recording setup so of course it's not state of the art "big recording house" gear. There are aspects of the process that I'm still "green" on like saturation and limiting but my mixes are beginning to sound pretty good on a wide array of media. It's just this level problem that's bugging me.

Hypothetically speaking if I got one of my tracks played on the air right behind a pro track I'd at least like the level to be comparable.

Am I expecting too much? I don't mean about getting played on the air...ha!

Thanks for your time,
Rusty K

 

[BigRay]

Fwiw...I use R8 Brain Pro from Voxengo for SRC tasks, and have found it vastly superior to anything I have ever used .For dither, I use POW-R , which is fantastic.
www.voxengo.com (there is a free version of r8brain that is good too)



as far as levels..

read this post

https://homerecording.com/bbs/showpost.php?p=2244207&postcount=35

mshilarious,

Thanks for your attention. Now that I have it...I'm going to pump you for info...ha!

The mastering process has been killing me. I A/B my results with professional tracks in my editor. I find the visuals helpfull. It's actually amazing the wide array of visual results one finds in otherwise good sounding CD's.

I'd like to be able to push the volume envelope like many of the pros do but I don't want to start a discussion about the pros and cons of "squared waves"....it depends on the song. I do usually use compression when mastering. I have a master limiter but I'm not really sure how best to use it.

One thing that's driving me crazy is that I'm getting clip like distortions when referencing with my headphones. These are not audible with near field monitors and there is no actual clipping. It has been suggested this could be that running my soundcard trim at -4dBv means my headphone amp in not able to handle the level. The distortions do in fact disappear when I change the setting to -10dBv however I can reference a track from a pro CD mastered to 0db and not hear the headphone distortion. This really has me confused and distroys my confidence in my process. If it weren't for the fact that I can hear some of this distortion on cheap boom boxes etc. at high volume I'd ignore it....but I digress.

I had been mixing to -1db but I've cut that back to -3db considering that perhaps no "headroom" was an issue with this problem. I'm still getting the headphone stuff. Even with the output section of my mastering software turned back to -1db I'm getting the distrotion. I could be pushing it too hard but compared to some of the "push" I see on pro CD's I don't think I'm overdoing that much. I haven't tried -6db but I will.

SRC?...software conversions? or hardware?


Most of my DC bias is minimal so I'll just let my mastering sofware level the wave in the master.


Maybe I'm expecting too much from my equipment. It's a home recording setup so of course it's not state of the art "big recording house" gear. There are aspects of the process that I'm still "green" on like saturation and limiting but my mixes are beginning to sound pretty good on a wide array of media. It's just this level problem that's bugging me.

Hypothetically speaking if I got one of my tracks played on the air right behind a pro track I'd at least like the level to be comparable.

Am I expecting too much? I don't mean about getting played on the air...ha!

Thanks for your time,
Rusty K

 

[mshilarious]

SRC = sample rate conversion.

As far as your other questions, it's hard to say without hearing a clip. I think you still have too high a level too early in the process. I am not sure what your mastering software is doing when you say you set it at -1dBFS. Presumably that is a peak limiter, but what that does to your audio is very dependent on the level going in, the chain before the limiter, and the gain in the limiter. Just because the output is -1dBFS and technically is not clipped doesn't mean it isn't horribly distorted.

 

[Rusty K]

I'm very grateful for the help....

BigRay,

I'm a musician trying to get a handle on the tech side of this. I know from the post you linked that I am definitely recording too hot, usually but not always -3db. This probably explains why I'm getting these anamolies. I thought that perhaps I'd been clipping my preamp. It has no meters, just a light the begins lighting red but I can hear no clipping/distortion when I monitor the individual tracks. I record into a Grace 101 pre and RNC compressor then into a LynxOne 24bit sound card. I use Adobe Audition to track. Unfortunately at this point I can't recut all these tracks but I'll try backing all my levels substatially down until the final master. That may not fix the problem for this CD but at least I won't make that same mistake next time.

The info on the 24bit vs 16bit was very helpful and I'll check out the voxengo link. Free is good!

mshilarious,

I use TRacks to master. There is a clipping stage in the output with a level control and sat control. The output control is like a fader for final adjustments to 0db. Precious little help in the manual on these for an intermediate like me. I'm pretty sure I'm just too hot too early in the process and perhaps overloading plugins or something like that as aluded to in BigRays post. I thought of this possibility when racking my brain over this problem.

I did use a different set of master plugs eq and compressor with simpler controls and got similiar results.

Chalk this one up to "Pilot Error"....isn't that usually the case?


Thanks again for your help,

Rusty K

 

[SouthSIDE Glen]

Hypothetically speaking if I got one of my tracks played on the air right behind a pro track I'd at least like the level to be comparable.

That's the broadcast engineer's job, not yours. While hypothetical, that "test" is not a relaible way to think about it because there is so much more crap going on in the broadcast chain, well outside of your control, as to be not worth considering.

As far as the rest, though, think of it this way. While headroom is important, another concept to consider is the overall dynamics "budget". There is only so much that can be squeezed out of any given sound. The earlier and harder you compress, the more of that "budget" you're eating up early in the process, with not much left over to "spend" by the time you get to polishing the two mix.

Having plenty of headroom in the overall two mix is meaningless if the individual tracks that comprise the mix have been pressed flat as a pancake before they even got there.

This is also true of signal quality. Concentrate on keeping things clean and sharp at the major domain transition points. Mic technique and preamps at the transition from sound to electricity, quality monitoring and trained ears at the transition from electricity to sound, and - just as important - good converters at the transition between analog and digital. Every weakness at any of those points will be amplified as problems when you get to the two mix, by which time they will be too late to fix. The tighter you geep everyting earlier in the process, the more wiggle room you'll have in "mastering" (the longer I am here, the more tired I am of being forced to use that term erroneously.)

And how you arrange your mix and how it's mixed tgether matters a great deal as well. get as much right in the mix as you can and don't count on fixing it in the master. Use track ompression only when needed to construct the mix or to acheive special desired sound and try not to compress as a "given" to get tracks to "sit" right in the mix until you have tried other techniques first. This goes back to the first paragraph and the "dynamics budget". It also goes, however to the overall robustness of the mix. The more that track compression is used, the more fragile the mix actually is. Use manual peak shaping, track level automation and even tongue and groove EQing to weave together the tracks better instead of compression to lay the tracks flat on top of each other, and you'll probably find yourself able to push the mix harder before it breaks.

And, finally, yes the music genre does make a difference. A wall of Les Pauls with infinite sustain and more distortion than a fun house mirror will go to a higher RMS than a cowboy with a Martin acoustic sitting around a campfire. You're right. So what? Just because you can doesn't mean you should. And just because "the big boys" do it doesn't mean you should either. Pretend you're making a movie instead of an album. Just because you're holding DVDs of "Waterworld", "White Chicks" and "Gigli" in your hands doesn't mean that's how you should make your movie.

Just let the volume handle itself in the mix; make yourself a mix that sounds as good as it can sonically, and then push it with just some light limiting (mabe shaving 2-3dBs max off of the top at -0.3dBFS or so), and be happy to move on.

If you want to compete for volume against a commercial CD, then get your stuff commercially recorded and mastered.

G.

 

[Rusty K]

SouthSide Glen,

Yes, I know of what you speak....Less is more and getting the tracks down right in the first place is most important. I'm getting there. These techniques you speak about are learned over time with experience and frankly I'm just working on my own material and not interested in doing this kind of thing professionally for other people. That puts me at a disadvantage because working with varried sessions is part of that learning experience I would imagine. This is true in becoming competent in any field of endeavor.

I guess I implied that I wanted to "compete" with the big boys. I didn't mean it exactly like that. I see my current project as nothing more than a demo because I respect what the "big boys" can do, however I live in a town with ecclectic musical tastes that are all over the place. It's not that hard to get local airplay and I'd like to leave the listener with a good first impression. Yes I could "demo" this project to a real pro engineer/producer and of course I'll pursue that angle. On the other hand some DJ might just like it enough to play it as is. While it may be the DJ's job to optimize the impact of my debut, in reality that's often not what I hear coming over my car radio speakers.


Didn't mean to bore you with personal info you didn't need. I appreciate your time and I will try to implement your suggestions.

Thanks,
Rusty K

 

[flatfinger]

Well.
I'm fairly imppressed w/ what the current generation of "home" gear can do , but as far as competeting with the pro sound production(big budge), thats very tough, not only do they have pro (not "semi-pro") gear; like 2 or 3K per rack unit, but they have years of cutting their teeth and problem solving experience. Give yourself sometime and you'll be able to do some damm good demos, but as far as (non-"indie") commercial quality at home????


Normalization is only going to scan the file , look for the peaks , move THEM to 0dbfs and the rest of the stuff in a proportional fashion: no increase or decrease in the peak to average ratio. Thats what the limiter set to stun is for!!!

 

[SouthSIDE Glen]

Normalization is only going to scan the file , look for the peaks , move THEM to 0dbfs and the rest of the stuff in a proportional fashion: no increase or decrease in the peak to average ratio. Thats what the limiter set to stun is for!!!

Well, there are normalizers (e.g. Sonic Foundry/Sony's) that give the choice between peak normalization and RMS normalization, with the RMS normalizer having the further option of whether to limit/reduce at a peak level.

Peak normailization is what you describe, and is usually pretty harmless for self-mastering purposes as long as the mix dynamics are already pretty solid and if you have no plans to send the mix on to an ME. I'd personally rather see a home rec'r/self-masterer put their effort into getting the mix right and then peak normalizing the mix than to slap together a mix and then try to force it to sound good and sound loud by throwing heavy MBC against a limiter. The former will almost always sound better than the latter.

RMS normalization, OTOH, is a real weapon of mix destruction. It's kind of like limiting, but it's different in two important ways. First, it often gives you the choice of whether you actually limit or not. If you think a mix sounds bad when you push it too hard against a limiter, try pushing the mix too hard without limiting before saturation .

Second, is that RMS normalization actualy lets the user select the target RMS value, that is the driving factor, with any limiting (or not) done if and wehn required to acheive that target. This is rather opposite of limiting in that you set the peak limiting values as the target and the RMS level will follow as the effect of the limiting. What IMHO is extremely dangerous and ideologically wrong about this is that a) dialing in an RMS number without conideration for the content is ludicrous. Every mix has an RMS it "wants" to be at, and trying to force it to be something that it's not is just bad engineering no matter how you spin it, and b) most rookies are either unaware or don't care about a), they just want to get their crap mix up to -9dBFS RMS because thy think that's a good way "to compete". Their getting their hands on an RMS normalizer is akin to a jihadist getting their hands on a shaped charge; even if their mix is as solid as a Humvee, it's sunk.

G.