• 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!

Does Phase Distortion/Shift Matter in Audio? (no*)

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,874
Likes
6,672
Location
UK
As a starter, go here: http://www.silcom.com/~aludwig/Phase_audibility.htm
That polarity inversion / phase shift is generally audible really is a knowledge many decades old. In the Lipshitz paper from 1982 Lipshitz states "Some of these effects have been known for over 30 years, and are quite audible on even very simple signals."
It really escapes me we are still discussing this, the issue is settled once and for all and therefore any further research won't happen.
Thanks for the link. I'll let Francis Vaughan see if that's what he meant by "peer reviewed modern research", as I'm not an expert on this area, but thanks for the link for sure.

I did the absolute polarity test and thought I noticed a slight difference between the samples, with perhaps there being more emphasis in the bassier parts at the start of the sample on the original sample, and I felt that some of the tinkly high frequency trails of the guitar sounded a bit different in the inverted sample.....but the more I tested myself blindly the less correct I became. I tested myself in 3 seperate "sessions" and each time I got the answer right about 3-5 times in a row at the very start of the test, but got worse as the test went on.....ended up in the region of 58% success or something. I used my Harman EQ'd K702 headphones and turned up the volume to louder than I normally listen.
 
Last edited:

JohnYang1997

Master Contributor
Technical Expert
Audio Company
Joined
Dec 28, 2018
Messages
7,175
Likes
18,292
Location
China
I think he means that the minimum phase for the particular case of a single driver with wide, flat amplitude response is linear (zero) phase.(?).
While that would be true, it just doesn't sound right when said along with 'single driver' and 'by default'.:confused:
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
Single driver is minimum phase.
Sorry for the sloppy wording, but that was addressed in a later post.
I should have written heaphones / single drivers don't have excess phase as they are minimum phase. And with a flat FR that means linear phase for most of the passband.
 

Andrew s

Member
Joined
May 9, 2021
Messages
69
Likes
127
As a starter, go here: http://www.silcom.com/~aludwig/Phase_audibility.htm
That polarity inversion / phase shift is generally audible really is a knowledge many decades old. In the Lipshitz paper from 1982 Lipshitz states "Some of these effects have been known for over 30 years, and are quite audible on even very simple signals."
It really escapes me we are still discussing this, the issue is settled once and for all and therefore any further research won't happen.
Certainly not peer reviewed. It would cost me $33 to read the Lipshitx paper but I will read it.
The reason it's still discussed is because speakers vary in relative phase between drivers, polarity from recording to speaker vary and we don't have a clear differentiation into a "gosh that's horrid" and "my that's right" by speaker, record or system. That is most people don't notice it in most music in the wide variety of equipment they use.
Regards Andrew
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
The reason it's still discussed is because speakers vary in relative phase between drivers, polarity from recording to speaker vary and we don't have a clear differentiation into a gosh that's horrid and my that's right by speaker, record or system. That is most people don't notice it in most music in the wide variety of equipment they use.
Regards Andrew
This is all true, but doesn't touch the point of audibilty. The changes are audible but irrelevant for the most part, and that's what I said early on.
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
To make it as easy as it gets for you all, I've made an example using a normal music piece you can use for ABX. The difference is striking, listen to the first note is all that is required. And looking at the waveform again it is clear why it sounds different when inverted.
 

Andrew s

Member
Joined
May 9, 2021
Messages
69
Likes
127
This is all true, but doesn't touch the point of audibilty. The changes are audible but irrelevant for the most part, and that's what I said early on.
It touches on audibility in normal use as opposed to special attempts to make it so.
I accept it is audible in special cases but I am yet to be convinced for music as normally listened to but will read the paper.
Regards Andrew
Edited to make my point clearer.
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
Neither is symmetrical.
Second one is topologically identical when flipped upside down (for strict identity you have to reverse in time).

I accept it is audible in special cases but I am yet to be convinced for music as normally listened to but will read the paper.
You might want listen first to the many examples available, notably the one I just provided. Once you hear it, no more need to study anything ;-)
 

Andrew s

Member
Joined
May 9, 2021
Messages
69
Likes
127
Second one is topologically identical when flipped upside down (for strict identity you have to reverse in time).

You might want listen first to the many examples available, notably the one I just provided. Once you hear it, no more need to study anything ;-)
I will, but specifically looking for an effect is not what I intended by "normally listened to" I never approach a work of music by focusing on it in that way. Regards Andrew
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
I will, but specifically looking for an effect is not what I intended by "normally listened to" I never approach a work of music by focusing on it in that way. Regards Andrew
Well, once you fully aware that there can be a systematic effect (and often is), you will also note it in normal recreational listening, assumed you have access to a polarity flip switch (many DACs have, SW players as well, or just use a plugin).
 

Tom Danley

Active Member
Technical Expert
Audio Company
Joined
May 6, 2021
Messages
125
Likes
580
I am not sure "does phase shift matter" identify s what conditions are causing the phase shift, or what frequency range it is in so I vote "it depends".

For example the phase shift one see's caused by the time of flight to the microphone effects all frequencies equally and so, does not alter the wave-shape the speaker produced and so is delayed in time but not audible.

On the other hand, the acoustic phase is what the driver did compared to the input signal when all the fixed propagation delays are subtracted. Maybe better to say that acoustic phase shift shows the degree the loudspeakers internal delays are different at different frequencies.

Dick Heyser (the fellow who first figured out a way to measure it) described acoustic phase as the change in physical depth position so far as time was concerned, the source moving forward and back.
It is this acoustic phase shift which alters or not, the input wave shape or preserves it in sound pressure, if you have two or more sources radiating, what you get is normally also location dependent as the sources are different distances depending on position. Headphones would be entirely different case than two ears. Two ears can often localize the depth / distance to the sources (and this identity is present or not while playing a stereo image).

It is my observation that if one looks at an "equal loudness curve", you see where your hearing is most acute and sensitive, most likely to be sensitive to things in this range and there is not doubt in my mind this shape is also related to your head's shape like your ability to locate position tied to your outer ear's shape etc.

Absolute phase is a tough nut too, say you have a fairly broad band but asymmetric wave shape (a 1 mic recording of a firework is an example), play this through a speaker that already has several hundred degrees of acoustic phase through that spectrum and then reverse it, you may or may not hear a difference or they may sound different but not one better than the other.
Take a simple speaker, correct the phase and mag (not the room) with FIR, listen close in the near field and you will likely hear the difference. To the degree the speaker disappears in the stereo image, the effect may be more audible.

Best,

Tom

A suitable test "signal" I made some years back, a tough signal to get right with speakers but usually good with headphones.

https://www.dropbox.com/s/eik72wzv5hptq3r/fireworks 2013 last 6 min cd.wav?dl=0
 
Last edited:

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
@Tom Danley , your fireworks track is already legendary. It's a pity that the main peaks are all clipped, though.
1623444426924.png
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
Rats, i have the original somewhere which shouldn't be
The clipping was already present in the file you hosted on your original site where I downloaded it many years ago. Fingers crossed that you may locate the unclipped original.
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,012
Location
Berlin, Germany
Top Bottom