Skip to main content

Hello to everyone, this is my first post here. :)
I have been reading about mastering lately, bought Bob Katz's Mastering Audio-The art and the science and I really find it exciting!

Now, maybe my question has been asked a thousand times here and maybe you are too tired of hearing it but I couldn't help it, sorry! :oops:

The thing is, lately I have been importing tracks from commercial audio CD's into Wavelab 5 to see and visually as well their waveforms.
What I have found out is that in most commercial CD's the chorus part looks something like this:

http://dl2.rapidshare.de/files/1854578/18897716/how_do_they.JPG

The verse's part breathes a bit more but the chorus parts's waveform is almost squashed and it seems as it has no dynamics at all.
And maybe it doesn't!
But even if it is that way when you listen to it does not sound squashed or overcompressed or anything.
In fact it sounds pretty natural.

And now the burning question:
How do they do it? :)

Thanx in advance for any answer!

Comments

anonymous Thu, 05/19/2005 - 06:49

Michael Fossenkemper wrote: what you are seeing are the peaks hitting full scale. The average level is still moving and dynamic, hopefully.

At least its not all squared off as much as "Vapor Trails"....

Basically, its peak limited so that digitally everything is hammered up to 0dBFS why the dynamics remain fairly similar before limiting.

anonymous Fri, 05/20/2005 - 01:26

Thanx for the replies guys!

So do you mean that they achieve this by using a limiter?
I have been a Waves L2/L3 Ultramaximizer user for sometime and I can tell you I have not been able to get this kind of result. If I pushed the limiter hard I could hear an apparent degradation to the sound.
I recently got MD3 for my Powercore Firewire. I got somehow to this result but only by applying compression before the limiter.

So, if it's mostly limiting and (little?)compression which tools they use in order to retain the dynamics of the program material?
From my experience(I am no mastering engineer) when you push a limiter that hard the music sounds squashed and unnatural.

Can you enlighten me on this?

Thank you! :D

anonymous Fri, 05/20/2005 - 01:50

Its not just by using a limiter.

That sample looks like the brickwalled the peaks to bring the volume up a bit. If you had a more detailed sample, it would show what really had been done. To keep the dynamics, you would also use a compressor before the limiter. There would probably be some EQ'ing done to. However, don't try to go louder thinking its better. Go louder if it is musical.

http://www.prorec.com/prorec/articles.nsf/files/8A133F52D0FD71AB86256C2E005DAF1C

anonymous Fri, 05/20/2005 - 02:35

BrianAltenhofel wrote: Its not just by using a limiter.

That sample looks like the brickwalled the peaks to bring the volume up a bit. If you had a more detailed sample, it would show what really had been done. To keep the dynamics, you would also use a compressor before the limiter. There would probably be some EQ'ing done to. However, don't try to go louder thinking its better. Go louder if it is musical.

http://www.prorec.com/prorec/articles.nsf/files/8A133F52D0FD71AB86256C2E005DAF1C

OK, here are some more detailed screenshots, I don't know if they will help...
http://rapidshare.de/files/1871178/detail1.JPG.html
http://rapidshare.de/files/1871182/detail2.JPG.html
http://rapidshare.de/files/1871199/detail3.JPG.html

anonymous Fri, 05/20/2005 - 03:10

You can see from your most detailed shots where its limited at (where it's squared off at the top). That where things are hitting the brickwall. If you looked at the Vapor Trails article, you would see where they did it too much and ended up with almost-square waves. Your screenshots show that the sine wave still exists. It also shows that it still has some of the dynamic range.

anonymous Fri, 05/20/2005 - 03:30

BrianAltenhofel wrote: You can see from your most detailed shots where its limited at (where it's squared off at the top). That where things are hitting the brickwall. If you looked at the Vapor Trails article, you would see where they did it too much and ended up with almost-square waves. Your screenshots show that the sine wave still exists. It also shows that it still has some of the dynamic range.

I read the article.Thanx!
So the way to achieve this result is by applying more compression(higher ratios maybe?Bob Katz in his book says that in mastering the typical ratio is at about 2:1- Lower thresholds maybe?) or more limiting?I am a bit confused here :oops:

Michael Fossenkemper Fri, 05/20/2005 - 05:54

stop thinking formula. it seems you are looking for a general setting for everything. the ratio is whatever sounds the best. ratio is only 1 parameter on a compressor. it's a combination of everything. there are also different kinds of compressors and each one does different things. don't forget about eq. how you shape a mix before compression and or limiting will get you a completely different sound and result. there are a million different combinations. How something is recorded and mixed will change how you use all of these tools. If you are working with a mix that is too compressed, adding more compression might not be the ticket.

Massive Mastering Fri, 05/20/2005 - 09:56

I'll add to that - Trying to get a mix to "look" like another mix is an awful, horrible idea. There. I said it.

Mastering isn't about capping the volume of a mix - In many cases, it's unfortunately part of it, but it's not the end game.

It's not about setting up a chain to run everything through - What you need to do is listen to the mix in question, visualize what it's "supposed" to sound like when it's done, and set up a chain based on what it's asking for.

The next mix will probably be asking for something completely different.

anonymous Fri, 05/20/2005 - 10:59

Yeah to get a mix sounding decent and be all squashed to hell like your first example, it's more to do with how it's mixed then mastered. If you try to compress so much on the 2-bus, instruments will be fighting each other like crazy, but if you mix it so each track has no dynamics and is hitting close to the peak, the final 2 track won't really require much compression to get it 'hot'. I don't recommend doing this though, because it more then likely will sound worse then if you went with less compression, but obviously that's up to you. That are some compressor/limiters that can squash dramatically and still sound natural, they cost $5000 or more. Take what I say with a grain of salt, as well as the idea that it's all worth getting yourself involved in.

anonymous Sun, 05/22/2005 - 15:40

Thank you for all the replies guys!
By the way, I never asked for specific "fixed"compressor settings and stuff. I understand very well that noone can give settings that will work on every piece of music.
It's just because I am trying to make a master of a rock/metal song that begs for that high level in the chorus like the screenshots I posted.

jamiey wrote: That are some compressor/limiters that can squash dramatically and still sound natural, they cost $5000 or more. Take what I say with a grain of salt, as well as the idea that it's all worth getting yourself involved in.

Just out of curiosity,can you name some of them? :)

anonymous Sun, 05/22/2005 - 20:38

just to note, im not like a master mixer, but ive learned some things which i hope will help.

waves: L2 will work fine. actually, the waves 'master' bundle is pretty much all you need to 'master' your songs. no, its not going to sound as good as if a pro studio did it, but it will give good results, if you know what youre doing it can be good enough. it certainly is good to start learning on and for home made type stuff.

basically what i do to beef up my mixes.

first off, make sure your sources are good. get the right mics, make sure everything going in sits well with everything else. once you do your thing and your mix is good (everything clear and nothing is fighting with another instrument, bass and kick are good examples of instruments that share ranges and can cause trouble).

mastering usually has 3 parts to it, EQing, compression (usually multiband), and finally a limiter. ive heard of people putting reverb in it (i dont get why.. acoustic tracks?). they also use frequency analyzers and RMS/peak meters.

what i do in polishing my mixes (really cant call it mastering)...

my chain usually looks like this
EQ -bypass everything flat
multiband compressor -bypass, set to mastering setting (explain later)
limiter -bypass
spectrum analyzer with RMS/peak reading

first i look at the analyzer, see where all the action is, see where the hats and the bass/kick are dominating, this gives me an idea of how to set up the multiband later on and i can see what frequency range is lacking and whats overdominating (usually the lows).

i set EQ to somewhat flatten the spectrum out. if theres not much around 500-1k i might put a small broad boost on it, if bass is dominating i might put a slightly tight lowpass to try and tone down the sub action (which gets pretty crazy on my mixes, your results may vary).

then to the multiband, the low band captures the bass/kick as best it can, the high band captures cymbol action and very SLIGHT presence from other tracks (usually S type consanants and the very top of the electric guitars).

i have a setting that is automatically 1.2:1 ratio for all, about 20 attack 300 release and thresh at 0. i bring up thresh to where it is just barely capturing kick hits, cymbol action, and middle barely limiting stuff down. the end result should be knocking a few dB off of the main mix while leaving it sounding very very close to how it did before hand. your settings may vary, read up on multiband compressors and experement with them.

then onto the limiter. -0.1 celling (sometimes 0.0 celling may clip). basically, songs in my genre are around -10 to -5 dB RMS on loudest parts (usually hitting about -12), so even though i cant match that with L2, i can get close enough to where my mix sounds about as loud as theirs. i shoot for -10dB RMS on choruses and about -15 at softer parts.

the results arent great, but they bring the volume up, which adds clarity.

EQ - you want to flatten your sound out. if your song sounds perfect in the mix, it will sound like ASS on a stereo system because of the users EQ settings. usually mastered songs sound pretty muddy and beefy and not terribly bright. try loading a song in, then flatten the EQ. carefully EQing your mix will make it sound good on other systems as well as bringing up underused frequency ranges.

multiband compressor - well, it compresses it without having trouble with the kick ducking out the rest of the song and such.

limiter - this is where you see all that clipping action. its amazing how many cut off peaks you find on commercial songs, i really dont get it because even with L2 i get really close to the same volume yet i havent found a single clip point. but, what its basically doing is bringing up soft parts, not just verses, but small small unnoticable level changes within the song. brick wall limiters act very fast. compare an unmastered mix with a mastered one, it looks more squished but zoom in and youll see that the mastered waveforms look 'fatter' than the unmastered forms. (ill get a pic in a sec).

thats basically what they do. for acoustic genres you really dont want to compress much, for orchestra type you really dont want to compress at all, for stuff like metal you want to pump that stuff UP.

oh, and as to why choruses are just about a solid line, its because choruses are where the action is. its the buildup of the song. the guitars go full in, the vocals are loudest, the harmonies kick in, and the drums are pounding. truely dynamic songs will fluctuate a LOT even WITH a limiter. if a song is compressed the whole way through, its not JUST bad mastering, it could very well be bad mixing/song composition.

btw.. have you ever seen the waveform for ricky martin's livin la vida loca? ill grab some screenshots.

anonymous Tue, 05/24/2005 - 06:39

I'm sorry but if you "flatten your sound out" especially using a frequency analyzer to "see" how your mix sounds, then it will almost unfailingly end up sounding like ass. Fletcher Munson curves and all:

http://hyperphysics.phy-astr.gsu.edu/hbase/sound/eqloud.html

Ears always before eyes should be the golden rule of mixing and mastering.

Also, you dont "add clarity" by increased volume from a limiter or clipping. You loose it along with the dynamics.

anonymous Tue, 05/24/2005 - 14:02

lucidwaves wrote: I'm sorry but if you "flatten your sound out" especially using a frequency analyzer to "see" how your mix sounds, then it will almost unfailingly end up sounding like ass. Fletcher Munson curves and all:

http://hyperphysics.phy-astr.gsu.edu/hbase/sound/eqloud.html

Ears always before eyes should be the golden rule of mixing and mastering.

Also, you dont "add clarity" by increased volume from a limiter or clipping. You loose it along with the dynamics.

my mixes sounded weak in the mid range, and i could tell why in the analyzer. not everyones mix is the same, but mine sounded like it was coming through a stereo system, and through a stereo system sounded horrible.

maybe clarity was the wrong term, it sounds more full when you add volume.

they describe this crap better.

EQ or equalization - this is the first step in the mastering process. The frequency spectrum is the level of different frequencies present in the recording. During the mastering process the frequency spectrum is analyzed and equalization is used to even out the spectrum.

anonymous Tue, 05/24/2005 - 15:14

Rider wrote:

http://www.sbomagazine.com/sbomag/Apr05/Recording.html

EQ or equalization - this is the first step in the mastering process. The frequency spectrum is the level of different frequencies present in the recording. During the mastering process the frequency spectrum is analyzed and equalization is used to even out the spectrum.

Rider,

This article seems like writing from inexperience probably from students considering it is a school, band, orchestra site. They list listening as the final step in mastering when I'm sure any MEs here would agree that listening should be the first and foremost attribute of mastering. They list the tools that are commonly used to sweeten a mix but fail to mention listening environment or the importance of an objective ear, two key factors in mastering.

I've just started in this field though I have been doing recording and pseudo-mastering for about 8 years. If you want to learn about mastering, that article is not a good start. They describe some tools, and I think they are poorly described however and the article gives false impressions and perpetuates misconceptions about mastering. The best way to learn about mastering (other than an apprinticeship) is listening.

Listen to as many different types of music on the best speakers you can afford.

You already have a better source of information, this forum. I would also suggest Bob Katz' book: Mastering: The Art and the Science. But above all, listen and tweak your skills as you train your ear. One of the worst things you can do is try to rely on a frequency analyzer to "show" you where your mixes are lacking.

EDIT: Turns out that the article was written by music sales staff.

Michael Fossenkemper Tue, 05/24/2005 - 19:22

what is happening when you add "volume" is that you are bringing out the midrange in your mixes because the low end and high end hit the limiter more than the mid and any heavy compression or limiting will tend to bring out the midrange. The problem is your monitors. The problem with using a limiter or compressor to do this work is pumping and fluttering. you would have much better results if you properly balance the mix before this stage. I would look into your monitoring situation, this tends to be the weakest point. Gear in this situation is only a band aid and will only get you so far. In order to go higher, you need to take a few steps back and invest in some room treatment and monitors. Before you start in with "I know my room or my monitors are good", I'm not knocking anything. half of my investment is is monitoring and room treatment.

anonymous Tue, 05/24/2005 - 19:43

okay i give advise on to something that might help, that has proven to help MY mixes, and i get hell for what i do to my own mixes? it works for me, im pretty sure i said in my post results will vary, but boosting 600-1500 has worked for me. my spectrum analyzer made me aware it was that frequency range that was lacking and causing crap sound.

poor mix? im using a shure 57 beta on a great sounding amp angled to a very sweet spot, i doubt the source is bad.

monitoring? okay im not afraid to say, i use mid grade headphones. yeah yeah headphone mixing is the devil, but its given me the EXACT results i want to get. stereo imaging is a little messed up yes, but i know how to adjust it. monitoring bad? well the results ive gotten from my headphones are exactly what i expect when i listen through apple pros, listen in my cars (not sucky) stereo system, friends stereo system, grandparents mid grade stereo, friends good computer system. i dont judge my tone on what it sounds like through my phones, i compare my phones to real world applications and i am very comfortable with them in adjusting tone. all the music i listen to is through them.

my latest mix i did, took it out to the car, it sounded GREAT.

i know i could mess around and do a little better in my mixing. slight compression and EQing on most tracks does a great job, they just lack volume and peak easily.

fuck it, do what you want with your mix, and ill do what sounds best on mine. theres no set way to mix, and what i do comes out what i want. if you need to crank more bright on your final mix, good for you. i dont, it sounds like complete ass.

anonymous Wed, 05/25/2005 - 05:42

Rider wrote: so its better to have a mix that is weak sounding and harsh than smoothed out and doesnt sound overEQd?

doesnt make sense.

Where in my post did it sound like I was saying that? The point of my post was that you should rely on your ears to tell you what your mix needs and adjust from there.

Noone is trying to slam you here, just trying to give you a little friendly advice. I'm relatively new to this game as well but alot can be learned here. I've learned a ton from people here who are more experienced, from Bob Katz' book and most of all from listening.

anonymous Wed, 05/25/2005 - 11:35

If you go the route Rider explains (completely valid if that's the style you are going for btw) You should always compare the before and after at the same apparent loudness. So, take that squashed mix and turn it down until it sounds as loud as the original, and then just go back and forth between the two. This way, it's much much easier making judgments about whether or not compression/limiting/distorting the mix is actually benefiting your song. Again, sometimes, depending on the style, it'll work.

anonymous Wed, 05/25/2005 - 14:19

i rely on both. i see where the frequencies are hitting so i have an idea of what needs EQing. its not like im taking off my headphones and using my analyzer to master it then exporting it, im listening to what im doing and ABing every plug i set.

yeah, you dont always need strong limiting (sometimes you might only need to throw up a limiter just to get peaks over 0dB, or not at all).

listen to songs like yours, or at least what youre going for. my style goes along more popular type music, metal mostly, so yeah i go for a more commercial sound.

but the original question is how do THEY do it, and provided a screenshot of it. in response (since its how i do my music), i said what _I_ do to make my forms like that. as i said, "basically what i do to beef up my mixes. "

my question is, why is just about every mix i see mastered clipped like hell? ive never seen my mixes clipped after limiting.

anonymous Wed, 05/25/2005 - 17:04

jamiey wrote: You should get one of these special meters, you will become more pro.

Toys don't define who is pro and who isn't. Pro is short for professional meaning you derive a significant amount of your income from that particular field, in this case audio.

Anyway, meters don't tell you jack your ears can't hear.

anonymous Thu, 05/26/2005 - 06:02

Rider wrote: im looking at the waveform itself, and the 'sawd off mountain' effect that runs rampant throughout the entire song.

Most likely clipping on the a/d converters or a brickwall limiter with a low threshold. Can anyone tell just from looking at it? Brickwal limiting and clipping look pretty much the same to me.

anonymous Thu, 05/26/2005 - 09:53

BrianAltenhofel wrote: [quote=jamiey]You should get one of these special meters, you will become more pro.

Toys don't define who is pro and who isn't. Pro is short for professional meaning you derive a significant amount of your income from that particular field, in this case audio.

Anyway, meters don't tell you jack your ears can't hear.

haha, I was joking with that last sentance. But there are meters that don't show clipping until 2 or more consecutive samples run over 0dBFS, and that may be part of the problem with engineers thinking they aren't clipping but they actually are. I thought I was being clear, sorry.

anonymous Thu, 05/26/2005 - 10:10

lucidwaves wrote: Right, but is there any discerning features of the waveform where you could tell the difference between a song that was clipped at the converter or brickwall limited and normalized? They dont look different to me.

Well, we all know that clipping looks like a perfectly straight line up near, or at, 0dBFS. But there is also the "slanted clipping" which looks like a not-so perfect line running downward, a sure sign that clipping was done prior to some other processing. It may be due to an engineers attempts at restoring some life into the track, but I doubt it. I think people just don't know what they are doing sometimes - um, yeah, I think you all know THAT! :)