• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Phase shift

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
It is what I have been saying from the beginning. That is fixed, constant ... that phase does not vary with the frequency -> linear, FLAT, ideally 0º in ALL audio band.

It has always been that way and a new amplification technology should respect it.

Imagine an audio chain in which the source, the preamp and the amplifier each vary the phase with the frequency ... We have to eliminate variables from the equation and not to add a new one.

It is impossible to achieve 0 degrees of phase shift in (through) a real system -- there is always some delay unless you have infinite bandwidth. If you delay each and every frequency equally, the signal will arrive all nicely lined up in time, but that requires linear phase.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Yes, but let us minimize the possible causes of phase shift in ALL electronic components, class D amps included. A wave has three parameters: amplitude, frequency and phase. An ideal audio system should modify only the amplitude of the input signal. Unless it is so throughout the electronic part!

Other thing is a constant, fixed time delay. A few weeks ago I am playing with it from JRiver MC 64-bits, DSP. One of the reasons why I am interested in the dominant H2 profile is to have greater depth in my second system, limited by listening in the near field. Playing with the delays I have achieved greater depth, playing with values from 1 ms to 5 ms. I am even able to fine tune a little more, for example 4.2 ms, thanks to which the piano concerts I have heard these days sounded better than with 5 ms (orchestral music).
 
Last edited:

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
Yes, but let us minimize the possible causes of phase shift in ALL electronic components, class D amps included. A wave has three parameters: amplitude, frequency and phase. An ideal audio system should modify only the amplitude of the input signal. Unless it is so throughout the electronic part!

Any audio system modifies the phase of the signal.

Other thing is a constant, fixed timne delay. A few weeks ago I am playing with it from JRiver MC 64-bits, DSP. One of the reasons why I am interested in the dominant H2 profile is to have greater depth in my second system, limited by listening in the near field. Playing with the delays I have achieved greater depth, playing with values from 1 ms to 5 ms. I am even able to fine tune a little more, for example 4.2 ms, thanks to which the piano concerts I have heard these days sounded better than with 5 ms (orchestral music).

So what are you delaying in relation to what?
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
Any audio system modifies the phase of the signal.



So what are you delaying in relation to what?
I believe thinks that different delay settings sound different. When applied to the whole signal, not frequency dependent. Just a simple delay.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
I believe thinks that different delay settings sound different. When applied to the whole signal, not frequency dependent. Just a simple delay.

To a single pair of speakers? If a 5 ms delay brings benefits, think about the hundreds of megaseconds between recording the album and staring playback? :)
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
That's how I've been reading it yes.

I just couldn't believe that is what he meant. The easiest way to generate a 5 ms delay is to push "play" 5 ms later than intended...
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
I just couldn't believe that is what he meant. The easiest way to generate a 5 ms delay is to push "play" 5 ms later than intended...
You need to recalibrate your DRM (delusional ravings metric) settings.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,029
Likes
10,796
Location
São Paulo, Brazil
In my understanding this is a no issue. Bruno already explained this, but again. The delay is constant in time for all frequencies. Only when you plot that constant time as a phase degree x frequency that it shows as a curve. But again these are constant delay in time for all frequencies. No problem.
 
Last edited:

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
Any audio system modifies the phase of the signal.
What does that mean? An audio system can certainly be designed to give no phase modification (it might need to add some latency, however). It will, always be true, of course, that a 1mm length of cable is causing some phase shift. Just not very much. So your statement is literally true, but not very useful :)
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
What does that mean? An audio system can certainly be designed to give no phase modification (it might need to add some latency, however). It will, always be true, of course, that a 1mm length of cable is causing some phase shift. Just not very much. So your statement is literally true, but not very useful :)

As useful as the phase mania of @maty, but I was really talking about delay (as it seems that is what maty also talks about).
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,200
Likes
16,981
Location
Riverview FL
My DSP twists phase based on what it "hears" at the listening position, overruling anything else in the chain (except itself).

1572731139833.png


The MartinLogan EQ must have been set for "sharper" automated correction than the JBL trace - more squiggles... this example was two years ago
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,200
Likes
16,981
Location
Riverview FL
At the listening position...

The MartinLogans throw a measureable "flat" phase over most of the range to the listening position, regular speakers don't, I think, due to their wider dispersion and wall/ceiling/floor reflections.

Here's the listening position phase measurement of my ML in my room vs @dallasjustice M2 setup in his room.

1572732692251.png


Brian Ding (Rythmik Audio) looked at one of my posts and said "That's impossible!"

My cone/dome speakers measure like the M2 above at the listening position. Nearfield, their phase measurement is much flatter (if not entirely flat).
 
Last edited:

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,013
Location
Berlin, Germany
Ray,
switch to linear frequency scale and then note the basically linear slope of the red curve... an underlying pure time delay, point-of-reference type of measurement error. You have to de-embed this for the real phase and then you'll also get a resonable Y-axis scaling (now it's a full 360 degrees per grid line!). The reference point must be set to the top of the first peak of the impulse reponse, for both data sets.
Phase measurements are tricky ;-)
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,200
Likes
16,981
Location
Riverview FL
switch to linear frequency scale and then note the basically linear slope of the black curve...

Do you mean the red curve? (M2 in this case)

1572736406438.png


Makes the 'stats look even better.

an underlying pure time delay, point-of-reference type of measurement error. You have to de-embed this for the real phase and then you'll also get a resonable Y-axis scaling (now it's a full 360 degrees per grid line!). The reference point must be set to the top of the first peak of the impulse reponse, for both data sets.
Phase measurements are tricky ;-)

And what do I have to do to adjust the black line (electrostats)?
 
Last edited:

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Ray, that looks like a much older version of REW... I was looking for a Phase only tab but couldn't find it!

I don't exactly understand Maty's obsession with having perfect phase.

Mucked around with the options and here's what I've got:
1572738883417.jpeg
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Just a question:

As for the group delay view in REW, which is more important to look at for reference? The "Min Group Delay"?

1572740500004.jpeg
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,200
Likes
16,981
Location
Riverview FL
Ray, that looks like a much older version of REW... I was looking for a Phase only tab but couldn't find it!

REW V5.20 Beta 24, Overlays window

1572741053737.png
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Only some words about phase shift, some designers do care, just now: https://www.diyaudio.com/forums/solid-state/344540-alpha-nirvana-39w-8ohm-class-amp.html



Updated with image:

View attachment 37252

- End off topic -

Maty why are you still banging on about this?

Phase shift has importance. The bit you haven't grasped with reference to the Bruno Putzeys comments and his class d designs is this.

If the phase shifts linearly with frequency ie x deg per Hz then it is just a time delay across the whole frequency range which is of no importance at all. When phase shifts non linearly then different frequencies get delayed different amounts from each other. This could be an issue.

Please try and understand the difference between the two situations and stop misrepresenting this as an issue with his class d designs. This is just your lack of understanding of the subject which has been endlessly explained. Its tedious to keep having to correct your lack of understanding.

Also, having said this please take a look at how much passive speakers usually mangle phase and you might come to a different conclusion regarding how important minor shifts are in amplifiers.
 
Last edited:
Top Bottom