How's sound supposed to sound?

[gotpop]

I'm new here, so, Hello!

I've seen plenty of books that have lists of instruments and which frequency ranges you might try boosting or cutting to make them sound better, but I was wondering if there was any book or website that would show _exactly_ how the energy would be distributed in an idealized recording of a particular instrument (i.e., mapped out on a graph so you could see the amplitudes of all the frequencies).

-Steven

[RefD]

the amplitudes of all the frequencies will vary greatly depending on so many factors that the mind boggles...even when measuring the same instrument/singer/box of oatmeal cookies.

[JGriffin]

What Ref said, though I prefer chocolate chip. I don't think there's any such thing as a definitive "idealized" recording of anything.

[gotpop]

the amplitudes of all the frequencies will vary greatly depending on so many factors that the mind boggles...even when measuring the same instrument/singer/box of oatmeal cookies.

True, but aren't there some tambres that are time tested and could be used as a reference? As you probably guessed, I'm not a professional engineer, so I've only had limited opportunities to learn how all the parts fit together in a great mix.

I've always been told to just listen critically to a lot of stuff and try to imitate it, and while I've come a long way in the last five years, I still wouldn't consider anything I've done to sound remotely "professional" sounding. Nowadays everything's so compressed it's hard for me to break it all down.

Just to give you an example, my myspace page (http://www.myspace.com/gotpop ) has a song called "Easter Island". It was recorded by a really great engineer on tape, but since I ran out of money and recording time, I've been trying to mix it myself in Logic just for the educational value of it. I feel like I just don't have a clue of how much to add or subtract from each instrument (especially my vocals).

I'm just looking for some perspective, I guess.

[JGriffin]

the amplitudes of all the frequencies will vary greatly depending on so many factors that the mind boggles...even when measuring the same instrument/singer/box of oatmeal cookies.

True, but aren't there some tambres that are time tested and could be used as a reference?

In general, yes. But you asked for "exactly," which is a horse of another color.

I know there are folks on this board who can answer this question more accurately than I can, and I need some time to consider a more thorough response anyway, so for now I'll let them say some stuff and just leave you for the moment with: Welcome to the board!

[cgarges]

Hey Steven,

Like most things in life, there are generalizations that can apply, but rarely do they fit even most circumstances and oftentimes, the exact opposite of the generalizations are true. (Well, I suppose there really isn't a triangle with a fundamental of 9Hz,but, you know...)

Take ANY track and mess around with it. Take an EQ, boost one of the bands all the way and then sweep the freqencies and listen to what you hear. Take note. Then, take that same EQ and cut that same band all the way. Sweep the frequencies again. What do you hear? Take note. Do the same for each band. Then, take another recording of the same TYPE of instrument and do the same thing. What was similar? What was different? Take note.

Do this enough and you'll start to see some of the generalizations appearing, but you might be surprised at what's different.

I did a session with with Mitch Easter yesterday. We were recording some guitar through a Leslie cabinet and something sounded really peaky to me. I was startigng to assess the situation when Mitch said, "Sounds a bit harsh, doesn't it? Just reach over there and roll out some 800 or something." To me, what sounded harsh seemed like it was going to be much higher than 800, but I engaged the EQ and cut at 800Hz and I'll be damned if it didn't EXACTLY do the trick.

I've been getting paid to do this for ten years and I've STILL got plenty to learn.

Chris Garges
Charlotte, NC

[RefD]

the amplitudes of all the frequencies will vary greatly depending on so many factors that the mind boggles...even when measuring the same instrument/singer/box of oatmeal cookies.

True, but aren't there some tambres that are time tested and could be used as a reference? As you probably guessed, I'm not a professional engineer, so I've only had limited opportunities to learn how all the parts fit together in a great mix.

I've always been told to just listen critically to a lot of stuff and try to imitate it, and while I've come a long way in the last five years, I still wouldn't consider anything I've done to sound remotely "professional" sounding. Nowadays everything's so compressed it's hard for me to break it all down.

Just to give you an example, my myspace page (http://www.myspace.com/gotpop ) has a song called "Easter Island". It was recorded by a really great engineer on tape, but since I ran out of money and recording time, I've been trying to mix it myself in Logic just for the educational value of it. I feel like I just don't have a clue of how much to add or subtract from each instrument (especially my vocals).

I'm just looking for some perspective, I guess.

there are a number of books and online resources that give a graphical depiction of what frequency areas certain types of instruments generally occupy, if that's what you were getting at.

[Professor]

there are a number of books and online resources that give a graphical depiction of what frequency areas certain types of instruments generally occupy, if that's what you were getting at.

The catch with those sorts of charts like the famous Carnegie Hall chart is that they really only show the fundamental frequencies of the notes the instrument can play. A chart which showed the harmonics above those fundamental pitches and the subharmonics or possible combination tones below would simply end up showing the full 10-octave range of our hearing and quite a bit beyond.

There isn't really anything that fits the bill for what Steven is looking for, but there's a really good reason why there isn't and indeed, why there can't be anything.

Consider for a moment an instrument as varied as the human voice. What frequency do I cut to make the voice less sibilant (less 'S' and 'T' sounds)? Well a trite and simple answer might be to cut between 7-10k, and in all honesty, the offending sound probably resides somewhere in that neighborhood. But will it be the same for your voice, and my voice? What about a high soprano voice? Or a deep baritone radio announcer? Or a child whispering softly? There are far too many possibilities to give a definite fix-frequency for "the voice", and the side effects of cutting too much or the wrong place could make the voice muddy or cause it to get lost in the mix.

So what about an instrument that is fantastically standardized like the piano. They all have 88 keys with the lowest A at 27.5Hz, and the highest C up around 3500Hz, it seems pretty straight forward. But are we talking about a small spinet upright or a 9' concert grand? Why do those instruments even sound different from one another? Why especially if they are both playing the same A-440 in the same room at the same volume?

For that matter, what of any instrument playing an A-440 at the same volume as another? Why does one piano sound "dark" while the other is "bright" when it's the same pitch? Why does the piano sound different from the violin or the guitar or the bagpipe or the alto voice on the same note?

The difference is what we call "timbre" but lots of engineers, musicians, and other folks out there don't know what is going on that causes the difference. As engineers, we learn to manipulate the timbre of various instruments through all sorts of means. The most obvious first step is the selection of instruments, followed by their placement within a room, our choice of microphone, placement of the microphone, choice of mic amplification, and then all the wide and varied methods of recording, processing, or otherwise altering the reality of what sound was actually produced in "reality". Yet we're still often at a loss to explain what on earth we're actually doing to that sound.

Well the timbre of any instrument is determined by the ratio of overtones produced above each fundamental pitch. There is also an effect from the 'attack envelope' which is how fast each individual note attack, decay, sustain and release happens. But the sustained sound of different instruments have their particular timbre because of the overtones.
Not sure what an overtone is?
Real quick, an overtone or "harmonic" is the linear progression of frequencies above a starting pitch. Most of us are familiar with octaves being a doubling of the frequency, so from A=220Hz, the next octave is A=440, then A=880, then A=1760, and so on. But if you take the linear addition of frequencies from 220, to 440, 660, 880, 1100, 1320, 1540, 1760, etc. then you get the harmonic above the note or the overtone series. I usually like to start examples at A=110 because the numbers fit a little easier, but wherever you choose to start, the numbers are simply a linear addition. From 10Hz, octaves would be 10, 20, 40, 80, 160, etc. while harmonics are 10, 20, 30, 40, 50, 60, 70, etc. When we check to see what those particular notes end up being, they end up as the intervals we know and love that form the basis of western music. A=110, A=220, E=330, A=440, C#=550, E=660, G=770, A=880, and so on. You see the octaves included along the way, and the interval of a 5th from A to E, then a 4th from E to A, then a Major 3rd from A to C#, then a minor 3rd from C# to E, and the intervals get progressively smaller.
(For those of you checking the math against charts or something, please remember that the equal temperment system used in our music skews the numbers a little bit so that instruments don't sound perfectly in tune with their harmonics in any single key, but they can play any of the 12 keys interchangeably.)

OK, back to timbre.
When you play a flute at A=440 you get a really strong fundamental pitch at 440Hz, but you also get a little bit of harmonic content going up the line of harmonics. When you play a piano at A=440, you get a very strong 440 and an almost equally strong 880, and also additional content on the way up. Vibraphone is almost all fundamental, and so sounds very much like a sine wave. Violin highlights a different combination of harmonics.
In a sense, you can define an instrument by the way it "lights up" different harmonics. Imagine being able to define an instrument as 100% Fundamental freq. (F1), plus 50% next harmonic (F2), plus 10% F3, 25% F4, 2%F5, 5%F6, etc. and then being able to differentiate that instrument from another one which is 100%F1, 5%F2, 30%F3, 10%F4, etc.
The whole system is referred to as "Fourier Analysis" and there are even devices which 'synthesize' sounds by allowing you to build them up one frequency at a time called "Fourier Synthesizers" and I found a really cool one online here.

So why am I getting into all of this crap as an answer to Steven's question?

Well, you see not only does every type of instrument highlight a different combination of overtones to produce its unique timbre, but individual examples of any instrument will highlight different overtones. That is to say that when we hear a "bright" piano we may be hearing one which produces a little more 4th and 8th harmonic than another. By contrast, a dark instrument may ring a little less on the 2nd, 4th, and 8th harmonics. And then to make matters more confusing, there are instruments which change their proportions along the way, like many reed instruments which can have a different tone quality in their upper range than their lower range.
Then we engineers come along and use various tools to try to capture these instruments.
Imagine a microphone with a 3dB dip in its response at about 1kHz (remember that's 1/2 the energy). Well that 2dB dip at 1k happens to coincide with the 4th harmonic of instruments playing at or around 250Hz. If the instrument's timbre is defined by its ratio of overtones, then we have just altered its ratio in that range. But then at around 500Hz the 4th harmonic is OK and the 2nd is lower, and if the instrument can play as high as 1k, then the fundamental is lower while all the overtones are at their correct levels. And in case that ain't confusing enough, that might actually mean that our interpretation of a microphone with a 3dB is that it is bright! on that particular instrument because while those lower harmonics are dipped, all the hiigher ones ring through as they should.
Then we start considering things like tubes and transformers in our gear. Why exactly does a tube make things sound "warmer". Well, a lot of it has to do with the fact that the tube is "ringing" with each note that comes through and adding a little harmonic resonance up through the series with a little boost on every other one. The tube is altering the ratio of the harmonics and so changing the timbre of the instrument.

Hopefully, if you've made it this far, that helps to shed a little light on why we can't really have some simple chart that says, "for more clarity in vocals, boost 2dB at 1.2kHz" or any similar kind of statements. No voice is the same to start with, and when you factor in the voice, the room, the microphone, and all the other factors effecting the sound from when it leaves the instrument to when it arrives at our ears, there is no particular guarantee of how all those ratios of overtones are going to be effected. Instead, if we want 'more clarity in the vocals' we just have to select turn up the gain a little on the upper midrange and sweep the frequency until we find the right spot, or better still, bring down some of the low mids, or maybe dip the upper mids out from some of the other instruments...
or maybe just push the overall volume up a little...
or maybe compress...
or perhaps a little less reverb...
or a different microphone...
and on, and on.

There are lots of variables in there to consider, so you're going to just have to trust your ears and do what sounds best.

-Jeremy

[RefD]

there ye go!

[Professor]

yeah, sometimes I get carried away.

-J

[scott anthony]

What sound sounds like to someone is based upon in part what that person grew up listening to. The engineer who grew up in an audiophile home would have a tremendous head start over the person who rocked his clock radio.

Check out live performances with an analytical ear, especially if the artists features instrument you've never heard. Live music (esp. acoustic) is neither as bright or compressed as modern recordings. You'll start to notice things really hard to do on a recording and in reaching for them, your own sounds will change. Repeat.

The EQ sweeping Chris recommended is a good idea. You can also learn volumes from really messing with a mulitband compressor...

[gotpop]

Hopefully, if you've made it this far, that helps to shed a little light on why we can't really have some simple chart that says, "for more clarity in vocals, boost 2dB at 1.2kHz" or any similar kind of statements. No voice is the same to start with, and when you factor in the voice, the room, the microphone, and all the other factors effecting the sound from when it leaves the instrument to when it arrives at our ears, there is no particular guarantee of how all those ratios of overtones are going to be effected.

I agree completely. I didn't want a magical EQ recipe--I was looking for spectrum analyses of well-recorded instruments. I suppose it'd be even better to hear examples of what a good sounding (insert instrument name here) sounds like outside of a mix. I actually have iTunes playlists of songs that have a solo Vocal, Guitar, or bass section to better understand how recordings of these instruments typically sound. I try to figure out what's been taken out to make room for other stuff, what's been boosted, etc. I only have a few of these, however, so maybe some of you have listening recommendations where there are instruments that are completely solo'd in places?

[thethingwiththestuff]

what you refer to in the last post could only be applicable in solo recordings of instruments anyway. in the context of a mix, individual elements cant all "sound their best."

[gotpop]

what you refer to in the last post could only be applicable in solo recordings of instruments anyway. in the context of a mix, individual elements cant all "sound their best."

I'm talking about songs where all but one instrument drop out for a few seconds, not recordings where there's only one instrument all the way through.

[ludwig_van]

what you refer to in the last post could only be applicable in solo recordings of instruments anyway. in the context of a mix, individual elements cant all "sound their best."

I'm talking about songs where all but one instrument drop out for a few seconds, not recordings where there's only one instrument all the way through.

He's saying that the proportion of frequencies which will make anything sound best is dependent on context. So listening to a brass section by itself, for example, will tell you how to make a brass section sound good by itself, but not how to make a brass section sound good as part of a mix with other things. This is the problem of mixing in solo that a lot of people run into. Oftentimes you have to make adjustments that will cause a track to sound weak when soloed, but sound good in the mix.