Why EQ a sound doesn't change timbre?

[Nowhk]

Hi all,

I've a question that fly on my minds since lot of weeks
Not sure if thats the correct forum section, but I'll try.

For how is defined, timbre is defined as the spectrum of a particular sounds. This means the amount/group of partials/overtones (manly harmonics) the sound got and theirs relative amplitude.

But... why if I EQ that particular sound I can feel the same timbre? In fact I'm editing amplitudes of some partials. So timbre should change.

Instead (till I don't do "dramaric eq-ing"), I "feel the same"; such as the "qualia" is kept, and only some attributes (high or low) are attenuated...

Any clues/explanations?
Thanks

[slipstick]

Hearing? Imagination? When I EQ a sound the timbre definitely changes. If it didn't I'd find a better EQ.

But of course it depends on the sound you start with. What you can't do with EQ is change the harmonic content of the sound e.g. if it only contained even harmonics then it still only contains even harmonics. You're just changing their relative amplitudes. So some of the sound quality is retained. As an extreme example, if for some reason you EQ a sine wave the timbre stays exactly the same. All you might change is the volume.

Steve

[ghettosynth]

You are probably also associating time dependent attributes of the sound to the timbre, e.g. the attack of a guitar or a piano.

Try using a low pass EQ with a sawtooth with no attack, max sustain, you will definitely here the timbre change.

[Nowhk]

Hearing? Imagination? When I EQ a sound the timbre definitely changes. If it didn't I'd find a better EQ.

So every time you hear your own track on different speakers (which in some way also alter the amplitude of some partials) you got a different timbre?

Try using a low pass EQ with a sawtooth with no attack, max sustain, you will definitely here the timbre change.

Thats of course, lp will remove harmonics content. I mean sort of peak filter, at some frequency, with a bunch of + or - db: thus only increasing/decreasing some range of partials...

[jancivil]

In fact I'm editing amplitudes of some partials. So timbre should change.

Instead [] I "feel the same"; such as the "qualia" is kept, and only some attributes (high or low) are attenuated...

In fact, you're doing something blunter than that. But attributes of a timbre are altered. The qualia, or the 'it's like that' is another matter, subjective. If you're deaf enough, sure you won't feel differently.

[ghettosynth]

Try using a low pass EQ with a sawtooth with no attack, max sustain, you will definitely here the timbre change.

Thats of course, lp will remove harmonics content. I mean sort of peak filter, at some frequency, with a bunch of + or - db: thus only increasing/decreasing some range of partials...

It's an example to demonstrate the error in your thinking. Less drastic filters will change the timbre less drastically. Your ability to hear the change, or that level at which you associate a certain quantification of change with a change in timbre, is purely subjective.

You simply cannot have it both ways. If you define timbre as the relative levels of partials at time t, then any change to those levels will change the timbre. If you incorporate the notion of time into your definition of timbre, then of course a static EQ will not change its impact over time and the relative character over time will largely still be present.

If timbre is defined loosely as the quality or character of a sound distinct from pitch and amplitude, then this does not exclude time from being a part of the definition. In that sense, a piano and a guitar will have very different timbres. Much of that timbre, however, is defined by the attack transients. However, now your question makes no sense because you are trying to define timbre more strictly such that an EQ is supposed to change the timbre.

There's nothing exciting, you aren't on to anything, you simply have a mistaken understanding of things. EQ changes timbre, but only in a static sense, generally. What you perceive as timbre is not purely static.

[DJ Warmonger]

EQ can only boost or cut the partials that are already there. It cannot introduce new frequency content or detune, for instance. Also keep in mind that EQs are indeed designed to be as transparent as possible.

[Nowhk]

There's nothing exciting, you aren't on to anything, you simply have a mistaken understanding of things. EQ changes timbre, but only in a static sense, generally. What you perceive as timbre is not purely static.

I've read this 20 times but I don't get what you mean Is not purely static? EQ will be applied during the whole "envelope", changing setting at the beginning of each try... not during playing.

EQ can only boost or cut the partials that are already there. It cannot introduce new frequency content or detune, for instance.

Of course. But aren't levels of each partials part of what define timbre itself?

[ZentralmassivSound]

Timbre is a combination of the lower partials, starting directly above the fundamental frequency, and the formant content at higher frequencies. The partials change with each note, the formants are basically the same for all notes. So you can indeed change the timbre by using a preferably narrow EQ with large boost or cut in the formant region, or by using it in the partials region, however there you would need to shift the EQ accordingly with each note. BTW that is what Meldaˋs MCharacter does, it's basically a moving EQ, and it can change the timbre of stuff stunningly.

BTW whether partials are time dependent or not doesnt change their role for the timbre, imho.

[ATS]

timber isnt a frequency?

[ghettosynth]

There's nothing exciting, you aren't on to anything, you simply have a mistaken understanding of things. EQ changes timbre, but only in a static sense, generally. What you perceive as timbre is not purely static.

I've read this 20 times but I don't get what you mean Is not purely static? EQ will be applied during the whole "envelope", changing setting at the beginning of each try... not during playing.

You are expecting EQ to change something that it cannot change. The relative change in partials over time is, in part, what you perceive as timbre. It's why the timbre of a guitar is not the same as the timbre of a piano. They are both vibrating strings, why don't they sound the same? It's not just pickups and wood, how you strike the string has a huge impact on what you perceive as timbre. So, EQ will affect the entire sound over time, however, by expecting it to change your perception of timbre, you are asking it to do something that it cannot do.

From wikipedia:

In music, timbre (/ˈtæmbər/ TAM-bər, also known as tone color or tone quality from psychoacoustics) is the perceived sound quality of a musical note, sound, or tone that distinguishes different types of sound production, such as choir voices and musical instruments, such as string instruments, wind instruments,...

You have already agreed that EQ will impact the timbre of a static tone. So, you understand that EQ will change timbre with respect to your perception when the time dependent attributes of the timbre match the time dependent nature, or lack thereof, of an EQ.

For the last time, you are confusing what you perceive as timbre with an overly simple understanding of what it is technically. An EQ will not make a piano sound like a guitar because your perception of the two timbres is heavily dependent on the nature of their attack transients.

An EQ DOES change timbre, that you can't perceive it is simply a problem with your perception with respect to what the EQ is changing. Period!

[ghettosynth]

Timbre is a combination of the lower partials, starting directly above the fundamental frequency, and the formant content at higher frequencies. The partials change with each note, the formants are basically the same for all notes. So you can indeed change the timbre by using a preferably narrow EQ with large boost or cut in the formant region, or by using it in the partials region, however there you would need to shift the EQ accordingly with each note. BTW that is what Meldaˋs MCharacter does, it's basically a moving EQ, and it can change the timbre of stuff stunningly.

BTW whether partials are time dependent or not doesnt change their role for the timbre, imho.

Indeed, but

Many commentators have attempted to decompose timbre into component attributes. For example, J. F. Schouten (1968, 42) describes the, "elusive attributes of timbre", as "determined by at least five major acoustic parameters", which Robert Erickson (1975, 5) finds, "scaled to the concerns of much contemporary music":

The range between tonal and noiselike character
The spectral envelope
The time envelope in terms of rise, duration, and decay (ADSR—attack, decay, sustain, release)
The changes both of spectral envelope (formant-glide) and fundamental frequency (micro-intonation)

The prefix, or onset of a sound, quite dissimilar to the ensuing lasting vibration

Onset or attack transients absolutely are a part of timbre, are time dependent, and may not be impacted by EQ at the fundamental. Consider the timbre of a flute, so much of it is not the harmonic content of the sustained tone but the chiff of the attack transient. The D50 leveraged this fact to great effect.

[jancivil]

But aren't levels of each partials part of what define timbre itself?

Part of it, but note I said EQing is doing something blunter than directly manipulating partials.
Timbre is tone color. Cutting the highs certainly results in a different timbre, for example. But there is no way in itself to determine that you're cutting specific partials. There are synthesizers that facilitate this, some you can draw a waveform in harmonics, but [w. EQ] you are dealing with everything that happens to occur at that frequency, depending on bandwidth (or 'Q'). I mean if it's just one instrument that you know really well and extremely specifically you might have a more educated guess at your EQ device... but if there's more than one thing in a mix, there is no way to parse what derived from what, you know, first partial or fundamental.

[jancivil]

BTW whether partials are time dependent or not doesnt change their role for the timbre, imho.

Why do people care to manipulate transients, then?

[jancivil]

what you mean Is not purely static?
EQ will be applied during the whole "envelope"

Timbre changes over the time of a sound. The attack of a trumpet for instance likely contains 'tone color' that the trumpet is not doing over a sustain or decay. You can't EQ this in by a boost if it isn't there. And conversely you prob'ly aren't going to surgically remove it, or not very well, by an EQ. This is a sound design issue; such as Yamaha FM, you apply the feedback loop on the modulator to help get that blast in the attack.

You can get more transient content in a recording by microphone placement.