What exactly is phase?
I totally get what phase is in the basic sense -- like how balanced cables use phase to eliminate hum. Thats not my issue.
I mean phase like as it applies to EQ. I've gathered that most eg circuits have phase issues, but what is it? How do EQ stages introduce or interact with phase?
Is it a (very) slight delay caused by the circuit itself that causes some degree of phase cancelation (excitation)? or is it something else? I firmly believe that phase is important (at a fundamental level), but I'm having difficulty explaining it.
I feel like there is a more profound relationship between phase and tone than most people acknowledge. I think it stems mainly from the "simpler is better" logic of amplification (little walter) and how it relates to conventional ideas of EQ response,
Can anyone explain to me the relationship between equalization circuits and "phase" ?"
Phase.
Moderator: Shoshanah Marohn
-
Dave Mudgett
- Moderator
- Posts: 9648
- Joined: 16 Jul 2004 12:01 am
- Location: Central Pennsylvania and Gallatin, Tennessee
Mathematically, phase is the parameter phi of a sinusoidal signal s(t) = A*sin(2*PI*f*t + phi), where A is the amplitude of the signal, f is the frequency, and t is time. phi may be time-varying, but in many situations, it is a constant that marks the time where the sine wave crosses zero and immediately after, becomes positive. Sinusoidal signals are important because it is possible to expand any 'real' physical periodic signal as a sum of sinusoids using a Fourier expansion, and when engineers talk about phase, they're generally referring to that constant phi.
In terms of amplification, the only time you need to worry about phase if you have two or more signals because the human ear can't distinguish differences in a signal with different absolute phase parameters. But if two signals of the same frequency but different phase are added together, there will be partial to complete cancellation of their combined amplitude. In the case of PI radians or 180 degree phase difference, two sinusoids of the same frequency will completely cancel each other out, in principle.
In terms of practical amplification circuits (let's consider analog circuits for now) - an AC amplifier has various components like resistors, capacitors, inductors, and active elements like vacuum tube or solid-state amplification elements. A pure, ideal resistor reacts the same at all frequencies, and doesn't shift the phase of an input signal at all. Perfect capacitors and inductors react differently at different frequencies (they are called reactive elements), and shift both amplitude and phase of input signals in different ways. Other types of circuit elements are nonlinear, but also may shift both amplitude and phase of the input signal.
In general, the less complex the amplification circuit (in terms of number and type of amplitude and phase-shifting elements), the less complex the amplitude and phase response of the amplifier. I guess to some extent one can argue that less reactive circuit parameters in a circuit often leads to an overall more direct and fuller, less phase-canceled sound. But I personally think it's hard to make sweeping generalizations, and especially about what is 'better' or 'worse'.
Myself, if I am looking for a sound that emphasizes the natural response of the tubes in a tube amplifier, I prefer an old-school minimum-components-in-the-signal-path circuit with minimal to zero negative output-section feedback (which also affects the overall frequency response). To me, that's great for some applications - I have a '53 5B5 tweed Fender Pro Amp 1x15" that is about the best jazz guitar amp I've ever heard, and, in a nice way, takes the edge off a Telecaster's lead pickup for some great classic country sounds.
But without control-circuitry, tubes can really be quite nonlinear when pushed, and thus I tend not to use an amp like that for the classic clean-pedal-steel tone or, let's say, a real quacky-clean 60s-later country Tele or Strat sound. Instead, I prefer a more modern (but still old-school) '57 5E6 tweed 2-input Fender Bassman 4x10" amp or some of the later designs. Actually, in most situations, I find that Bassman perfect for both guitar and clean pedal steel, a feat I thought impossible not long ago.
Gradually, as time rolled on, in an attempt to get increased bandwidth and clean volume out of the classic American amp designs like Fender, tube amp circuits got increasingly complex with increasing use of negative feedback in the circuit. The tone generally got thinner but cleaner, especially with the standard-grade speakers most amps had.
In short, I consider what type of circuit to use strictly a matter of personal taste, and of course also a matter of what fits with the music being played. A lot of this is set by precedent - trying to match sounds from classic recording archetypes. Blackface and silverface Fenders have a distinct sound that was really defining for a lot of 60s and 70s music, especially country and other roots styles. Marshall went back to the tweed Bassman circuit for its first designs, and then moved to an even more raw sound that ultimately came to dominate rock. Personally, I have generally preferred amp designs that live on the edge between the earlier, simpler designs and the super-clean later Fenders and uber-clean solid-state amps. For me, the issue is getting that simpler, more direct sound at higher power/volume levels without getting the shrill response that trying to over-control the power tubes causes.
My take.
In terms of amplification, the only time you need to worry about phase if you have two or more signals because the human ear can't distinguish differences in a signal with different absolute phase parameters. But if two signals of the same frequency but different phase are added together, there will be partial to complete cancellation of their combined amplitude. In the case of PI radians or 180 degree phase difference, two sinusoids of the same frequency will completely cancel each other out, in principle.
In terms of practical amplification circuits (let's consider analog circuits for now) - an AC amplifier has various components like resistors, capacitors, inductors, and active elements like vacuum tube or solid-state amplification elements. A pure, ideal resistor reacts the same at all frequencies, and doesn't shift the phase of an input signal at all. Perfect capacitors and inductors react differently at different frequencies (they are called reactive elements), and shift both amplitude and phase of input signals in different ways. Other types of circuit elements are nonlinear, but also may shift both amplitude and phase of the input signal.
In general, the less complex the amplification circuit (in terms of number and type of amplitude and phase-shifting elements), the less complex the amplitude and phase response of the amplifier. I guess to some extent one can argue that less reactive circuit parameters in a circuit often leads to an overall more direct and fuller, less phase-canceled sound. But I personally think it's hard to make sweeping generalizations, and especially about what is 'better' or 'worse'.
Myself, if I am looking for a sound that emphasizes the natural response of the tubes in a tube amplifier, I prefer an old-school minimum-components-in-the-signal-path circuit with minimal to zero negative output-section feedback (which also affects the overall frequency response). To me, that's great for some applications - I have a '53 5B5 tweed Fender Pro Amp 1x15" that is about the best jazz guitar amp I've ever heard, and, in a nice way, takes the edge off a Telecaster's lead pickup for some great classic country sounds.
But without control-circuitry, tubes can really be quite nonlinear when pushed, and thus I tend not to use an amp like that for the classic clean-pedal-steel tone or, let's say, a real quacky-clean 60s-later country Tele or Strat sound. Instead, I prefer a more modern (but still old-school) '57 5E6 tweed 2-input Fender Bassman 4x10" amp or some of the later designs. Actually, in most situations, I find that Bassman perfect for both guitar and clean pedal steel, a feat I thought impossible not long ago.
Gradually, as time rolled on, in an attempt to get increased bandwidth and clean volume out of the classic American amp designs like Fender, tube amp circuits got increasingly complex with increasing use of negative feedback in the circuit. The tone generally got thinner but cleaner, especially with the standard-grade speakers most amps had.
In short, I consider what type of circuit to use strictly a matter of personal taste, and of course also a matter of what fits with the music being played. A lot of this is set by precedent - trying to match sounds from classic recording archetypes. Blackface and silverface Fenders have a distinct sound that was really defining for a lot of 60s and 70s music, especially country and other roots styles. Marshall went back to the tweed Bassman circuit for its first designs, and then moved to an even more raw sound that ultimately came to dominate rock. Personally, I have generally preferred amp designs that live on the edge between the earlier, simpler designs and the super-clean later Fenders and uber-clean solid-state amps. For me, the issue is getting that simpler, more direct sound at higher power/volume levels without getting the shrill response that trying to over-control the power tubes causes.
My take.
-
Joseph Meditz
- Posts: 345
- Joined: 14 Nov 2005 1:01 am
- Location: Sierra Vista, AZ
It's easier to think in differences in delay which are the cause of phase shifts. For example, if you were in the back row of a music hall and the piccolo and tuba were playing the same melody it would be very disconcerting to hear the piccolo before the tuba or vice versa.
Different delays in different frequency bands cause unpleasant smearing distortion. In the very worst theoretical case, a single instantaneous, infinitely short, impulse would be heard as an infinitely long hiss of white noise. In this extreme case, both the impulse and the white noise have identical, perfectly flat frequency responses!
To add to what Georg said, an amp can be "linearized" to improve frequency response at the expense of identical delay, or linear phase. To the listener, especially a musician, an amp with extremely non-linear phase sounds just plain horrible.
Different delays in different frequency bands cause unpleasant smearing distortion. In the very worst theoretical case, a single instantaneous, infinitely short, impulse would be heard as an infinitely long hiss of white noise. In this extreme case, both the impulse and the white noise have identical, perfectly flat frequency responses!
To add to what Georg said, an amp can be "linearized" to improve frequency response at the expense of identical delay, or linear phase. To the listener, especially a musician, an amp with extremely non-linear phase sounds just plain horrible.
-
Donny Hinson
- Posts: 21192
- Joined: 16 Feb 1999 1:01 am
- Location: Glen Burnie, Md. U.S.A.
Ben, first off, rid your mind of the idea that "phase distortion" is a bad thing. 
Most of what we read about phase distortion on the web pertains to hi-fi, the "audiophile world" where the listener wants little or no coloration on a recording, hearing it exactly as it was made. In the world of guitars, of course, things are much different. Guitar players (and most steel players) actually want to avoid the flat response and exact phasing that audiophiles seek. For the most part, they want a lot of tonal lattitude and different sounds, and they actually want to actively change what is heard, as opposed to being "locked in" to just one tone.
Audiophiles should worry about phase distortion, overload distortion, and intermodulation distortion. For guitar players and steelers, though, these are not always bad things, but often very useful tools that expand the tonal pallette that's available.
Most of what we read about phase distortion on the web pertains to hi-fi, the "audiophile world" where the listener wants little or no coloration on a recording, hearing it exactly as it was made. In the world of guitars, of course, things are much different. Guitar players (and most steel players) actually want to avoid the flat response and exact phasing that audiophiles seek. For the most part, they want a lot of tonal lattitude and different sounds, and they actually want to actively change what is heard, as opposed to being "locked in" to just one tone.
Audiophiles should worry about phase distortion, overload distortion, and intermodulation distortion. For guitar players and steelers, though, these are not always bad things, but often very useful tools that expand the tonal pallette that's available.