Audio loudness, I don't get it.

Hi everyone,

I'm sure youve heard this topic mulled over and over to pieces. I have a few questions, though. I have recorded bands and sounds for over 6 years and noticed that on CD's i listen to, the "loudness" and "presence" of music are so much more significant, when compared to music I have recorded and tried to compress, hard limit... pretty much anything I do, it gets it almost as loud, but the problem is, is that i record in cool edit, say, a song that I want to match from a professional cd, and the waveform is SIGNIFICANTLY LOWER, just by looking at it by eye.  Are there methods im missing out on? I know theres mastering but still the waveform is so much smaller, yet so much more effective. Are there methods of making sure the EQ's are balanced so much to the point that you really can maybe use a statistical frequency graph or what? Or should I be using the multiband compressor setup? Please someone help me out

thanks always (im going to copy this into the syntrillium forums as well, just to warn about redundancy)
Ron Cavagnaro

You said it yourself. It's the mastering that does it. There they compress the music to get it as loud as possible. Nowadays most of the times, too loud, so the quality suffers.

Anyway, I use the multiband compressor most of the times to correct some EQ- or volume mistakes that are left from the mixing process. Not really to make everything louder. With the compressor I rather try to make things smoother or punchier. Or give everything some more warmth.

Then I let a limiter (Ozone's most of the time) do the job to make everything louder.

Another thing with those commercial cd's is that the mix they start mastering with is so good (balanced volume- and EQ wise). So then it's possible to do even more with it during the mastering process. In other words, mastering will always be about enhancing then and not rreally correcting. Plus the fact that we're dealing with professionals also.

I posted this on the former Syntrillium site so here it is here too:

Reference points. Subjective terms like loudness. Pretty hard to get that last 10% going to get a pro sound isn't it ?  

You might consider this. Get the Yamaha 'Sound Reinforcement Handbook' and read the chapters on Sound Levels and Dynamic Range about 100 times - that's what I'm doing. I'm only up to 4 though!

What might be bugging you is something like this. Individual track record levels both average and peak as well as full mix rms and peak levels. You can get a pretty good pre-master of a full mix by making some peak adjustments as well as bringing the rms (loudness) up using 'sane' loudness maximizer techniques.

Assuming you've recorded the tracks correctly and nothing is either too low or crossed the 0dB digital peak ceiling of your soundcard
(for example maybe you've recorded with a track average of -6dB) and you have finished your mix satisfactorily you can try to make a pre-master and adjust the loudness to compete with commercial CDs a little better. If you seriously want to compete with professionals take it to a pro who masters for a living - this is just for fun and demos.

Normalize the peaks of your mix to -0.3dB (to allow for small overshoot) then rip a song from your favorite reference CD, a song that sounds musically similiar to the one you're working on. Get the stats on each song from the Cool Edit Pro single track view Statistics. Look at the average rms (loudness acording to how we hear) and the peak values. The difference between these 2 values is the [clinical] dynamic headroom of the song. You can also look at the average rms power and the max rms power, this is how much the loudness varies.

Now after having normalized your mix to -0.3dB or something and it still does't stand up against your reference song in terms of loudness then you can try to adjust the loudness level (average rms) up while at the same time limiting the peaks (since you don't want them going over -0.3dB (or whatever your ceiling is for a particular media like CD). At the same time you're sacrificing the dynamic range which is OK if you have too much headroom, but not OK if you don't - It'll just sound too squished.

You've already mentioned multi-band compressors and you can use those to adjust peaks and headroom in various frequency bands or octave ranges so you don't disturb the dynamics in areas that don't need adjustment.

There's a whole bunch of stuff to do before that to ensure a balanced mix (dynamics and EQ) and there's a whole lot of stuff to do after that to fine tune a good mix into a great pre-master.

I'm not sure if you have an unbalanced mix or if it's just low in perceived loundess because a couple of peaks are preventing you from getting a true normalize. It's kind of like when you do statistics, you have to knock some of the meaningless transient data out of the way...

You mentioned you've been doing this for 6 years. Are you coming in to digital from analog - like tape decks and things ? You get a lot of loudness and stuff for free out there. It's way different in here for me but at least the physics is the same - I just have to pay attention to it more!

Good mixing and pre-mastering !
kylen

...the waveform is SIGNIFICANTLY LOWER, just by looking at it by eye.  Are there methods im missing out on? I know theres mastering but still the waveform is so much smaller, yet so much more effective. Ron Cavagnaro

Ron, I'm a little confused on that point. Do you mean the pro mix's waveform is smaller, or your own?

I have to admit I'm also a little confused by the number of people who are mystefied at how to get things as loud as the average modern rock monstrosity. I mean, unless I'm missing something, all you have to do is compress the living daylights out of it, and then lop off the top 6 db so there's practically no headroom, right? Will it affect your sound quality to do so? Of course! That's why most current top 40-ish music sounds like distorted, overcompressed crap!

Ron, I don't want you to take this as me being snippy at you, personally. I sympathize with what you're trying to do. I guess I'm just not sure what the mystery is as far as loud mixes. If you look at the average loud mix, you'll see pretty much a solid band of waveform. There's the key. Don't leave any dynamic range and your mix will be loud as hell.

Right? What am I missing? Why are so many people having a hard time with this one?

Chris

Everyone is basically telling you the same thing.  Groucho went a step further and explained the "trade off".  Most commercial  recordings today have the life compressed out of them.  Meaing that by scrunching the peaks down in proportion to the rest of the file, there are no (very little if any) dynamics left.  Some styles of music (rock, rap..) can tolerate higher degrees of compression/limiting than others (jazz, classical...).  If you want to fight the volume wars, then compress the file heavy and then run it through the hard limiter with a setting of -1db.  It will be loud, but it might not be what you really want.  Try it... you might like it.  

BTW:  Make a copy of a file and compress/limit to your hearts content.  If you don't like the end result.... you've still got the original file to fall back on!  

okay... so heres the thing, if i compress things so much, it usually ends up bringing in the noise of whatever that might be left. so if i compress things so theres barely any dynamics, i still dont notice any power. and yes i said that MY MIX or whatnot in cool edit is HUGE and pro sounding fresh good sparky stuff's wav are considerably more dynamic and smaller. That's what I don't get. I want to be able to check frequencies or SOMETHING and try to make it balanced. In the meanwhile im going to play with compression. I do understand compression mostly, but lets say in the graph i'd want to push the higher graph line, (everything) lower on the Y scale, or is that limiting?

gah!
Confused,
Ron

Have you tried hard limiting?

...then run it through the hard limiter with a setting of -1db.

I actually go as high as -0.1dB myself and it seems safe enough.  Interestingly enough, when I look at the statistics for a track off a CD (of the contemporary, compressed-to-death variety, that is ), there will often be hundreds or thousands of possibly clipped samples, and they can be quite audible - by audible, I don't just mean everything sounds overcompressed and lacking dynamics, I mean sometimes I can hear actual clipping.  It's scary that some professional mastering engineers can do such a rough job.

[judicious]

okay... so heres the thing, if i compress things so much, it usually ends up bringing in the noise of whatever that might be left.

I'm not sure I understand this, the way it's been worded.

....so if i compress things so theres barely any dynamics, i still dont notice any power...

Probably because you've neatly squashed the very thing that gives any kind of "punch" at all --- the dynamics!
Heh!  

and yes i said that MY MIX or whatnot in cool edit is HUGE and pro sounding fresh good sparky stuff's wav are considerably more dynamic and smaller.

So I don't really understand the problem... if your mix is already good, why would you want to mangle it?  

That's what I don't get. I want to be able to check frequencies or SOMETHING and try to make it balanced. In the meanwhile im going to play with compression. I do understand compression mostly, but lets say in the graph i'd want to push the higher graph line, (everything) lower on the Y scale, or is that limiting?

Okay, if you need this to be louder, there will be tradeoffs, as has already been pointed out.

I really think some judicious multiband compression is probably all you need. Of course, as with anything "audio", every process you run on your source material should be performed with the idea that the results are always  program-dependent.

If you can't hear what it needs, even after breaking it down into frequency bands and adjusting things independently, I think my best suggestion would be to get some help (meaning; the kind of help that can sit and listen in the room with you).
 

Best of luck!  -Jon

okay... so heres the thing, if i compress things so much, it usually ends up bringing in the noise of whatever that might be left

rpc4493,
It sounds like you have an additional problem if you get noise when compressing. This means that you are introducing noise at some point in your recording chain. What is your recording chain ?

The other thing about adjusting your material to compete against the loudness of professional commercial CD's is to pick one you like and match your dynamics to that. Don't pick the loudest thing out there if you don't like it such as Cher's followup CD to Believe - a very very overly loud in the worst sense of the term distored mastering job - my subjective opinion. Don't compromise youself - but do compare your work to professional works. I use works from the 70's, 80's, 90's and current. Get both your friends and enemies (    )  to review your work also.

There are lists of good reference material out there on the internet for pop rock mastering that actually have headroom. Neil Young's latest 'Passion' and a bunch of stuff. Point being if you like it master to it. Don't master anything to the FM radio of course as they have companders that'll throw you off.

Loudness isn't a dirty word. It doesn't mean overcompress something. It just describes how your ear perceives frequencies and balance at different volume levels. It dosen't mean the overly compressed and clipped thing that can occur if you squish it too much.

Music at it's best has the proper loudness so I can hear it at its' optimal quality, as it was meant to be heard when it was conceived - be it live performance or studio multitracked creation.  

Make it as Loud as you like - but no Louder than you like. dig ?

kylen

Ed: I just read jonrose post above me which is dead-on of course. If you're having trouble getting the 'loudness' you deserve in CEP try Izotope Ozone2 (multiband) or Voxengo Elephant Mastering Limiter (single band VST need a VST adapter).

Also for $75/hr go sit with a mastering engineer for an hour or 2  - I'm going to in my area in the near future.

man lots of information here... i guess i dont really know where im going with all of this... but i think i do want to explore cool edit's multiband compression-- can someone explain what i should do? i split the frequencies up and then what, do i run each band through a compressor? what then? do i toy with each compressor's grid or what? i usually end up just making it sound louder but... i dont feel like i have much CONTROL over it. any help?

Ron C

i think i do want to explore cool edit's multiband compression-- can someone explain what i should do? i split the frequencies up and then what, do i run each band through a compressor? what then? do i toy with each compressor's grid or what?

You've been reading a lot of posts - that's good.
They may be running together a bit.

Cool Edit Pro has a single band dynamics processor that has a band-limiting feature. This means you can stack a few of these in an effects rack and have each one set to be responsible for different frequency band ranges such as - bass, lomid, himid, high. You would have 4 stacked dynamics processors with this approach.

CEP also has a Frequency band splitter which does the same thing that classic hardware frequency splitters do. Split frequencies into seperate tacks. You can then insert effects on those tracks essentially turning former single band effects info multi-band effects (possible caveat on effect latency here).

Seaching the old Syntrillium forum and tutorials will give you some good ideas about where to begin to set compressor controls.

Just try to remember that all sounds have an attack, sustain, and release time. That is the nature of their envelope. Compressors have attack and release controls. Most compressors have an additional control called threshold that tells the compressor when to listen. The compressor ignores the sound envelope below the threshold but turns on when the envelope passes above the threshold setting.

How do you know where to set the threshold using CEP ? On other compressors you can watch gain reduction lights. Well CEP doesn't have gain reduction lights on it's compressor so some say use your ears and listen to the dynamics while adjusting the compressor - this assumes a proper listening environment. Another approach is to gather CEP statistics and make adjustments accordingly.

Another approach is to purchase a multi-band compressor DX plugin with gain reduction indicators on each band.

good compressing
kylen

Just in an attempt to actually say something helpful, I thought I'd offer this:

http://www.digitalfishphones.com/main.php?item=2&subItem=3

Ever tried messing with it?

I use Endorphin only occasionally on mixes - generally on loud, rocking stuff that demands a fairly compressed, "punchy" sound. I, like Zemlin, have never found a need to deliberately mix for volume, but I have found Endorphin capable of a remarkable amount of volume boost without sounding clipped and poopy. As far as tools designed to specifically raise volume while preserving clarity (as much as possible, anyway), I think Endorphin is at least as good as anything else out there, if not better than most.

Plus it's free.

Chris

hey thanks groucho! it actually is a pretty good plugin! (just from fiddling with it and not reading the manual)

any other free plugins anyone?

(man i have to buy that spinaudio vst wrapper)

groucho, zemlin, rpc9943, jonrose, AMSG

Would you be willing to take a moment and list a couple of the commercial pop mixes that you rate in your top 5 list that exhibit good loudness characteristics.

I'd like to consider them in my own list that I use for reference material.

Thanks,
kylen