Help in understanding phase shift

[Mbira]

I'm not a good math person yet. I understand that capacitance can cause a phase shift by holding on to the signal and causing a bit of "smear", yes? What other factors can cause phase shift? If this is too general of a question, does anyone know any good reading (again without too much math) that would explain phase shift a bit more?

Cheers as always...

:guinness:

Joel

 

[jrmintz]

Acoustic reflections can cause phase shift, like inside the headgrill of a mic.

 

[Mbira]

inside the headgrill of a mic.

Wow, will that really be enough to cause a shift? I could see a reflection picked up from the source hitting a wall, and then getting re-picked up in the mic doing it.

I forgot to mention the other obvious cause of two mics recording one signal.

Joel

 

[jrmintz]

I'm not exactly sure of the technical definition of phase shift, but as I understand it (two of the same signals, one slightly offset in time from the other at some frequencies) reflections inside a headgrill could certainly cause phase shift at some frequencies. Gus has posted a fair bit about that.

There's also phase shift between a miked and DI signal.

 

[zmix]

[quote author="Mbira"]I'm not a good math person yet. I understand that capacitance can cause a phase shift by holding on to the signal and causing a bit of "smear", yes? What other factors can cause phase shift? If this is too general of a question, does anyone know any good reading (again without too much math) that would explain phase shift a bit more?

Cheers as always...

:guinness:

Joel[/quote]

In a capacitor the current leads the voltage, this phase relationship is what makes EQ possible.


Don't worry about it too much, it's a law of physics. You cannot change it.

 

[Ethan]

"Phase Shift" refers to a time difference (or shift) between two or more signals (I think Mira is interested in the shift in a signal that occurs between input and output) The difference is measured in degrees. If you have Art of Electronics, I believe the first chapter gives an easy to understand primer, and general perspective of some occasions where "phase shift" occurs at the electrical level (between input and output). Analysing phase relationships of sounds in the real world or even just a room get too complicated too fast. It seems like a simple concept until you start thinking about it too deeply. Just think of it as time and difference.

Think of a sine wave, that starts at the base line 0, proceeds positive, then back to zero, then negative, and back to 0. If another sine wave at the same frequency starts at 0 (or any portion of the positive or negatice cycle) at exactly the same time as the first signal, they are "in-phase". Phase shift is when one of the signals is not exactly "in-time" with the other signal.
When two signals are not "in-phase" certain frequencies will cancel/add to a certain degree.

As for capacitors, like zmix said, the current leads the voltage by 90 degrees (with an inductor, it's the opposite, the voltage leads). This is because they are both storage devices that work in different ways. Dig into AOE, some of it will start to make sense after about a dozen readings...I'm still working on mine--repeatedly... :sad:
PRR's my role model...I wanna be just like him when I grow up :grin:

 

[Mbira]

I guess I understand "what" is phase shift. A more clear question I have is:

What can cause phase shift in an electrical circuit-other than capacitors. Can resistors, etc? I'm not talking about flipping phase in a tube or transistor.

If you are trying to attenuate a balanced signal with a stepped attenuator, can phase be affected by resistor tolerences (not just gain?)

Joel

 

[gyraf]

any capacitive and/or inductive component can affect time domain.

And yes, resistors have both capacitive and inductive properties.

So yes, in theory, any component will influence time domain/phase response.

But the magnitude of influence will depend entirely upon the associated circuit - e.g. load and source impedances at the given point in the circuit.

So no rule-of-thumb here, sorry.

Jakob E.

 

[Guest]

As mentioned previously, phase can be affected by any component in a circuit, but the degree phase is affected is usually neglible in components like resistors and a lot in components like capacitors and inductors.

Since you understand what phase is, and you know how capacitors affect phase relationships, then you can relate a similar effect to inductors. Both capicitors and inductors are reactive devices that are used in electrical circuits to change phase relationships.

So far we have just mentioned passive components. Any active component, (i.e. tube, transistor, opamp,) will affect phase as well.

Depending on the circuit, there are many, many little things that affect phase. Component placement, length of component leads, wire position and type all have some affect. Sometimes these little things are critical and sometimes they have only a small affect, it really depends on the circuit.

I think the META has some information on basic electronics that also may help clear this up further.

 

[Mbira]

Good stuff-thanks guys!

Joel

 

[Guest]

[quote author="Ethan"]"Phase Shift" refers to a time difference (or shift) between two or more signals [/quote]

Exactly "linear phase shift" refers to the time difference betwen
signals.


There are circuits with linear phase, like some digital filters,
or FFT equalizer. Mainly paragraphic (every is done via
parametric eqalizer and other curves are translated).

Phase atach of it is equivallent to delay.


Many analog circuits (eqalizer, shelf,...) are
of the minimal phase.
There phase characteristic is Hilbert transform of modulus characteristic.

That hilbert transform can be done efectively on the Bode diagram,
where one curve have extrem, other have inflex point.

For further reading: Heyser: Time delay spectrometry. It was
printed in the AES antology.
I have pages from it scanned, but what about Author s law,
it is AES own ? But here are many peoples from AES, Can I put
scans here ?

xvlk

 

[DrFrankencopter]

Phase shift does not refer to a time difference....it refers to a phase offset (in degrees, not seconds) relative to a reference. Fourier showed that any complex wave can be broken down into a series of sine waves. Each of these sines is defined by:

y = A*sin(omega*time + phase_angle)

where:
A is the amplitude
omega is the frequency
time is ....ummmm, time
phase angle is the number of degrees of offset of that sine wave (i.e. does it start at zero for time=0, or some other number).

Phase shift would be the difference in phase angles between two (usually correlated) signals. Phase shift can often be interpreted as delay, but generally this interpretation only holds true at one frequency (except if time delay is the actual cause of the delay, or if it's a 'linear phase' process). All 'real-time' processes that change the frequency response (i.e. EQ) introduce phase shift of some sort . The only way around this is to apply an EQ to the signal forwards and backwards...this cancels that phase shift, and doubles the magnitude change (obviously this is not very practical for audio recording, but it's used all the time during data analysis).


Cheers,

Kris

 

[Guest]

All true, but maybe somewhat complex for this thread which is supposed to be sans math. Leaving the subject to it's most basic core, I don't think it matters that much that phase is really an angular measurement. As a first step in understanding I believe it is most helpful to get a feel for what happens when phase changes, and what things in the circuit have the biggest effect on phase. I hope I am close to the mark and not being counter productive.

 

[CJ]

Transformers and neg feedback can cause phase shift.
Circuit design also. (cathode bypass caps, screen bypass caps)

There is a bunch of stuff in RDH4 but the math is kind of heavy.

 

[Mbira]

So for a layman, the important thing in an audio circuit would be balancing the phase shift of both sides of a balanced signal, or two channels of a stereo unit, right?

It seems like phase shift should be an important issue, as not being accurate with that, wouldn't a signal (balanced) cancel out alot of itself if there was inaccurate phase shifting going on-therefore making a signal have less of a "true" sound or maybe even less bandwidth? Is this something that is audibly noticable to anyone other that audiophools, or is this theoretical stuff that doesn't affect sound?

Joel

 

[Guest]

Exactly right. The main issue with phase shift is how it causes cancellations and additions to the signal chain as it goes through your circuit. Sometimes you want this to happen, as in tone control stages and feedback configurations, but mostly you want to avoid instances where phase differences occur.

 

[Kev]

[quote author="Mbira"]Is this something that is audibly noticable to anyone other that audiophools, or is this theoretical stuff that doesn't affect sound?[/quote]

as with all things audio with the right demonstration, any thing can be heard easily
I think this thread should start again as it has had so many examples from polarity to outright delay

Phase shift generaly is looking at the very small delay vs frequency

Phase shift of an op-amp with feed back is a very different problem to the differences between a DI and Mic recording bass guitar
... although both can be described as phase error or shift
Remember even speakers woofer and tweeters have a phase shift and then when you add the x-over it all changes again.

there is a book in all this stuff

can a human hear phase shift or differences in group delay frome only a single sound source ?
this has been debated over and over ... often by the audiofools
:roll:
we will leave that for now

it is easy to demonstrate phase differences when multiple audio signals with difference phase are mixed together
both with stereo image and with mono tracks

NO you don't have to be an audiofool to hear this stuff

so which phase would you like to talk about first ?
:green: