I understand that the timbre (tone) of a sound depends on its overtones, not just the fundamental frequency; and based on years of experience and a little study, I usually know roughly which frequencies to boost or cut to alter a sound—or least I have some idea of the starting points, even though I am no expert.
 
Still, something has always baffled me: the frequencies we boost and cut often seem to go against logic.
 
For example, to add definition to a voice, or to reduce its harshness, boost or cut around 2.5-3 Khz.  But those are the frequencies of the highest E to G on an 88-note piano and that seems awfully high.  Adding air by boosting at 10-12K is odd since that is way off the musical scale.  Removing boxiness from a guitar usually means cutting around 800 hz, but that is around the A above high C--nearly two octaves above middle C.
 
On the lower end, the numbers seem more reasonable, since we HPF a bass just below its lowest musical note to remove non-musical content.
 
You see where I am going with this?  I know these numbers and they work, but they seem to defy common sense.
 
Ideas or comments (other than just calling me dumb)?
 
 

One factor would be that the overtones above the source frequency define our comprehension of detail more than the source frequency itself.
EG the 7k sibilance issues.

If you have a few extra bucks the EQ Explained vid over on Groove3 is pretty awesome. I'm pretty sure the explanation to your question has to do with the overtones and simply getting certain not so necessary frequencies out of the way of other more important frequencies. The track on it's own may not necessarily sound better after EQing but in context of the mix it will. I used to try and give my guits fat bottom end but after watching that I realize I was doing the exact opposite of what intended because the bass should be dealing with those frequencies. Seemed counter intuitive to me because that's not how I'd set up my amp but I guess that's how it's done.

Oh and if memory serves you have your main note, say A 440. So the main sound of that note resides at 440hz obviously but the overtones will cause frequency spikes at 220hz, 110hz and in the opposite direction 880hz. Basically the overtones appear at double the hz (octave above) or half the hz (octave below). So by adjusting those points of the frequency range you are in fact adjusting part of that overall tone... just the more subtle parts of it that contribute to the timbre.
 
Kind of a mangled interpretation but I can pull out my notes on the matter if you like. Cheers.

Sorry... I was a little off about the octave thing. Those overtones can actually fall under the frequency range of other notes that are related to the main tone like a perfect fourth or a perfect fifth as well as the octaves. So it makes sense that you would manipulate seemingly unrelated frequencies. I dug out my notes on the subject. Maybe they will help. Cheers.
 
Groove3 EQ Explained

FREQUENCY SPECTRUM

Frequency is measured by the number of cycles per second air vibrates when a sound is generated.

The lower the pitch the less cycles per second. The higher the pitch the more cycles per second.

These cycles per second are measured in "hertz". "hz" is the common abbreviation. One thousand cycles per seconds is a "kilohertz" and is abbreviated as "khz".

The generally accepted range of human hearing is 20hz-20,000hz (20khz). Most humans cannot hear frequencies outside of this range.

Musical notes are simply set points within this range. For example: A 440 which is often used to tune instruments to. A is the note name and 440hz is the number of cycles per second that generates that note. If you play the A string on a guitar and it produces a tone at 338hz the tone will sound flat. If it produces a tone at 442hz it will sound sharp.

You can use a Frequency Analyzer to see the hz value of a tone (in the tutorial he uses the Blue Cat's Freq Analyst made by Blue Cat Audio). A guitar tuner is a type of Frequency Analyzer. A strobe style tuner is a type of Stroboscope.

When working with an EQ you have two main variables. Frequency measured in hertz (hz) and Amplitude (loudness) measured in Decibels (dB). On a frequency analyzer or a graphical EQ you will generally have the Frequency Spectrum (hertz) displayed horizontally from left (lowest) to right (highest) and Amplitude (decibels) displayed vertically up (louder) and down (quieter).

The following is a Wiki link that has note/frequency chart for the keys on a piano:

http://en.wikipedia.org/wiki/Piano_key_frequencies

When you play a note of the same letter name the cycles per second will multiply by 2. For example if you played the A note one octave higher than A 440 the frequency would be 880hz. If you played the one above that it would be 1760hz. If you played the A below A 440 the frequency would be 220hz. However if you follow these exact multiplication factors the the notes will start to sound slightly out of tune because there is a slight tonal curve. A piano tuner will compensate for this. The resulting tuning is referred to as "tempered tuning" and the compensation curve is called the Railsback Curve. It is necessary for tuning instruments with a long scale length such as a piano. If it is not used when tuning a note played on the low keys of the piano would sound out of tune with notes played on the higher keys. An example of what can happen when you follow the exact frequency values to perform music is old musical greeting cards or toys that would play a digitally created melody. The notes would sound awkward and unpleasant. This is referred to as Equal Temperament and should be avoided.

Here is the Wiki entry on the Railsback Curve:

http://en.wikipedia.org/wiki/Railsback_curve


When using electronic tone generators to create specific notes your frequency analyzer will display one spike on the Frequency/Amplitude graph. However these types of notes are uninteresting and unpleasant like the holding pattern on a television after the broadcast day is finished. When you look at the graph while it is processing the note from a musical instrument such as a piano there will be multiple spikes across the frequency spectrum with the main note being the tallest amplitude spike. This main spike is the "fundamental tone". The extra spikes represent the Harmonics or Overtones of a sound and are why different instruments (and sounds in general) don't just sound like a long, robotic beep. This is also sometimes referred to as "Timbre" (pronounced tam-ber). Without these harmonics/overtones music and life in general would be very tonally boring. When using an equalizer it is important to be aware of these overtones. If you turn these frequencies up or down in the wrong way you can make things sound worse when the goal is to make them sound better. Also if you pay close attention you may see that the more prominent overtones may fall under the frequencies of other musical notes that relate to the fundamental tone (like a diatonic 4th or 5th interval).

One fascinating study I did was to find the source freq. of a bass (sweeping a narrow Q) and then did a narrow Q boost on it AND it's harmonics (EG 50,100,200,400,800,1600) and pulled the rest down.
Then I pulled those freqs down on other trax.
It makes an incredibly clear, punchy bass without masking or gain.

Probably should have included the following as well and in fact I'd like to post the my entire entry on EQing but it is pages upon pages of material so I'll just leave it at this for now....
 
EQ FILTERS

An EQ allows you to "boost" (increase the decibel level) or "attenuate" (lower the decibel level) of specific sections of the frequency spectrum. By doing this you can alter the sound of the audio signal to accentuate desirable tones and their related overtones or reduce/remove undesirable tones/overtones. This is especially important when dealing with many sounds/instruments playing at once because when there is overlap of two signals on the frequency spectrum they can interfere with each other. This can make things sound muddy or completely hide parts of the audio signal. Sometimes an unneeded frequency of one instrument will interfere with a desirable frequency of another. So you would reduce or remove the unneeded frequency by using the EQ to "filter" out that part of the signal's frequency range.

These filters are also called "bands" and are essentially like volume knobs that control a specific part of the frequency range. You can turn them up or down. Some EQ's only have one band (like a Tone knob on your stereo). Many EQs however have multiple bands (like a multi band EQ module you'd put in your stereo system's chain) and those multi band style EQs will generally be what you'd use while mixing audio. This way you have far more control over the signal.

M Glenn... Yes indeed. It is such an amazing subject and EQ is so full of possibilities I get excited just thinking about it. Looking at this document I made while watching the G3 vid is making me want to jump right into mixing my album but I have a ton of time correction, editing and overdubs to do before I get to that. *sigh*
 

It is that we are talking about logarithmic scale instead of a linear one that makes it confusing? I can't follow what logic or common sense is being employed in the original post.

I understand the confusion... or I think I do. I think konradh is looking at the "fundamental tones" and their place on the frequency spectrum and wondering why adjusting seemingly unrelated frequencies would have such a drastic effect on the sound. The simple answer would be that they are very much related because there is more than just that fundamental tone in play. If it were just the fundamental tone then all we'd hear is just a mechanical sounding "beeeep".
 
I may be terribly wrong but I know that used to confuse the heck out of me as well.