• Welcome to ASR. 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*)

Nothing to complain at audibly level by my needs and capacities, but numbers are what they are.

123 SINAD of Topping D90 III is better than 113 WiiM Pro Plus, what is audibly or not I don't know.

I found diverse answers in ASR, some of them arguing imperceptible over 120 SINAD and other enough over 110.
Other than very sensitive IEM where the faintest noise can be a problem, I would argue even 70 SINAD is more than ok
(specially if noise is at check and given the other measurements are not horrible)
 
Other than very sensitive IEM where the faintest noise can be a problem, I would argue even 70 SINAD is more than ok
If the faintest noise is a problem then 70 SINAD probably won't be enough for classical music.
 
If the faintest noise is a problem then 70 SINAD probably won't be enough for classical music.
That's why I put the disclaimer.

Edit: that's for the noise element of the SINAD of course, cause if noise is in the -hundreds and it's only distortion that determines this 70, it won't be a problem at all.
 
That's why I put the disclaimer.
I meant, it won't be enough even for non-sensitive IEMs or headphones. For example I'm listening "Fisher, Helmchen / Schubert: Violin Sonata in A Major "Duo", Op. 162, D. 574: I. Allegro moderato" right now (youtube link) on Sennheiser HD650. If I play 12-bit dither noise (-72 dBFS) in parallel, it is easily audible in quiet parts.
 
I meant, it won't be enough even for non-sensitive IEMs or headphones. For example I'm listening "Fisher, Helmchen / Schubert: Violin Sonata in A Major "Duo", Op. 162, D. 574: I. Allegro moderato" right now (youtube link) on Sennheiser HD650. If I play 12-bit dither noise (-72 dBFS) in parallel, it is easily audible in quiet parts.
I have tried similar with Audeze LCD-X and my ancient Audio-Technica with Scheherazade and I have to to go to 80's-90's so I get your point.
Noise must be at check at least to CD quality for peace of mind.

It's the sole SINAD number that does not tell the story I want to point out.
For example this popular amp has a 70dB-80dB SINAD at 5kHz but a good SNR. It won't be a problem at all in its good operational range and no one is reporting such.

(I admit I always tend to think speakers, so... )
 
I'm guessing it's a proprietary scale called "Volume Limit" and it's not the voltage that is reduced by 5%, but the total volume range, and the company makes its own decision on how to scale it. I assumed that Heinz measured it.
I am using a Wiim mini with volume fixed at 95% and the RME ADI2 Pro peak level meter - with my yesterday's Tidal playlist - showed me peaks at -2.1dBFS.
 
Last edited:
Other than very sensitive IEM where the faintest noise can be a problem, I would argue even 70 SINAD is more than ok
(specially if noise is at check and given the other measurements are not horrible)
How in the classical music particular case? I have a DAC dongle of around 90 SINAD and can listen noise at quiet passages in some high dynamic recordings.

That can also be the tendency of many modern recordings to enhance realism by increasing dynamic range on the studio, but this is a topic which is under discussion in other thread.

What in the home theater setups? Can be being on the 120s relevant in action movies with massive explosions and quiet dialogs or tiny effects?
 
Cause it has really quiet passages.
Thanks, I knew that fact. What I really wanted to ask is how much will be the minimum headroom to take an extra dynamic range recordings.

I think not so much, at least my setup peaks at 110 dB @ 1m and my room is not one of the most silent, let’s be optimist and put 35 dB ambient.

I’m listening at 2 meters, so can expect reasonable 105 dB peaks, that means I would be well deserved with the mentioned 70 dB SNR.

Am I right?

(Also I red elsewhere that we can identify organized sounds below the noise floor, but surely cannot go more than -10 dB from the ambient noise, so in worst conditions 80 dB SNR will suffice)
 
Also I red elsewhere that we can identify organized sounds below the noise floor, but surely cannot go more than -10 dB from the ambient noise, so in worst conditions 80 dB SNR will suffice
Maybe also consider that the spectrum of the ambient noise in your room is probably not flat.
 
Maybe also consider that the spectrum of the ambient noise in your room is probably not flat.

Well is a mix between a Russian lady that shouts quite loud by telephone most of the time, a couple whose reproductive noises could merit a place in a National Geographic documentary, the traffic, and seagulls.

The best way to improve my DAC is moving from the town center to the country …
 
How in the classical music particular case?

Bach Violin Sonata No. 1 / Julia Fischer / Pentatone:

1755788388337.png


Michael Jackson Thriller / Wanna Be Startin Somethin:

1755788455311.png


What this means is that the Bach recording will be turned up when you listen to it, bringing you closer to the noise floor. And the quiet passages are really quiet - there are parts where the music pauses entirely and the only sound you hear is the recording studio.
 
Michael Jackson Thriller / Wanna Be Startin Somethin:
Think I'll try a recording off the vinyl with levels set somewhere sensible. See something like what it should look like when not compressed to within an inch of its life.
 
Bach Violin Sonata No. 1 / Julia Fischer / Pentatone:

View attachment 471280

Michael Jackson Thriller / Wanna Be Startin Somethin:

View attachment 471281

What this means is that the Bach recording will be turned up when you listen to it, bringing you closer to the noise floor. And the quiet passages are really quiet - there are parts where the music pauses entirely and the only sound you hear is the recording studio.
I know, and precisely I played this concrete sonata on piano arrangement. My silences were at least as good as Fischer ones…
:)
 
Bach Violin Sonata No. 1 / Julia Fischer / Pentatone:

View attachment 471280

Michael Jackson Thriller / Wanna Be Startin Somethin:

View attachment 471281

What this means is that the Bach recording will be turned up when you listen to it, bringing you closer to the noise floor. And the quiet passages are really quiet - there are parts where the music pauses entirely and the only sound you hear is the recording studio.
Think I'll try a recording off the vinyl with levels set somewhere sensible. See something like what it should look like when not compressed to within an inch of its life.

Loudness wars illustrated:

Screenshot 2025-08-21 at 17.07.01.png
 
Regarding the topic of the thread, can be phase distortion be simulated in a rough way to feel how it seems?
There is something called an all pass filter that shifts different frequencies differently.

Do you have Audacity? Audacity is a free open-source audio editor. If you download it, I recommend "download without Muse Hub" so you don't accidently sign-up for something you may not want.

There is a REAPER plug-in called ReaEQ that has an all-pass function and it works with Audacity. REAPER is not free but the plug-ins are. I had to download and install them all and then I copied reaeq-standalone.dll from wherever it was installed to the Audacity plug-ins folder.

...I don't know exactly what the all-pass function is doing (how much phase shift, etc.) but it's doing "something" to the waveform. You might be able to find some documentation for it.

Note that since an all-pass filter shifts different frequency bands around in time and phase, the frequency components sum-up differently. Some waveform peaks get higher and some lower. That doesn't affect the sound of the dynamics but you can get (digital) clipping. So it's a good idea to run the Audacity Amplify effect at -6db or so before you start and before you A/B.
 
You can't disagree using data you don't have. For 50 years, Dr. Toole and crew did not consider something useful to go after. The few others who researched it said it is only audible with specialized signals and even there, its value for preference can't be established.

Given the above, the best position you can take is neutral and not advocate for it especially since such phase "accuracy" (whatever that means) likely won't come for free.

I thinks it's easy disagree with the idea than once things are "done right", there is no concern for phase......and to be able to do so without having to provide data.

I've not yet seen a phase study that was willing to make a definitive claim that phase/GD is completely inaudible.
And most studies call for further research like the one you posted earlier.
I'd like to pull a snip from that study up again as an ongoing example....
1755799565607.jpeg


Those conclusions do state the authors saw no need for concern with phase on regular music and speech with the test equipment, speakers and headphones that they used.
The conclusions also give examples of clear phase audibility, state more work needs to be done, and that audibility limits might become more stringent as improved reproduction systems become available.
I think we can all agree on the basic "done right" spinorama goal. But for me, as a speaker builder to say ok, phase doesn't matter....... that's as good it gets....
....well I view that as capitulated nonsense. The very studies themselves say to me, go to work. Find some icing for the cake.

From all my experiments, I don't know what to conclude about phase linearity yet, other than thinking that if it matters for regular music, it matters more as frequency decreases.
I can positively say, using linear phase crossovers makes it much easier and more likely to get to the "done right" stage...if for no other reason, I advocate linear phase for that.

??? That is how we determined what frequency response we should aim for in speakers. There are many compromises in speaker design and we need to be clear on what brings the most listener satisfaction, and what not.

Umm...for me, speakers should be designed for best flat response on and off axis. I think speakers need to be like DACs, and amps....as technically correct and flat as possible.
I want to be able to use the speakers anywhere,.... and adjust to whatever room, venue, or outdoors......starting from flat.

Any built-in voicing or response curve adjustments, I see as attempts at maximizing satisfaction to a target market segment.
And as a separate process to be overlayed on technically excellent flat speakers.
So imo, assessing speaker preferences is more about making good business/marketing decisions, than a part of how to make excellent speakers.
 
Last edited:
And most studies call for further research like the one you posted earlier.
There aren't that many studies. I only know of two: one said it was not audible when audiophiles were playing their music. And two, said it requires special situation and signals. That they put disclaimers at the end doesn't give rise to this being a significant issue to worry about.
I'd like to pull a snip from that study up again as an ongoing example....
There are plenty of signals that cause issues for speakers. Take a bookshelf speaker and a 40 Hz tone. It will distort like mad or not play at all. Pretty sure that is audible and not preferred. I suggest spending energy solving that problem instead of wondering about phase where only pathological signals are audible.

If you do care about phase more, then conduct your own formal experiments and come back with that data. Otherwise disagreeing for the sake of what one feels in their belly is not going to get us anywhere.
 
Back
Top Bottom