Some Test Tracks to check for Absolute Phase!

[gregorio]

540 - 360 : 180
imo 160 is the better one

Sure but what about your worst 210° example, wouldn’t it actually be better in my example?

Well sorry, actually if you talk about 2 mic setups you may flip one mic's polarity, so the recorded phase lines up with eachother because of different distances from mic to sound source, and i guess you could in theory line them up "perfectly" instead of just flipping the polarity "to make them somewhat lineup"

There’s three serious problems with that: Firstly, just flipping the polarity of one mic will not “line the phase up with each other” because the phase difference to start with is almost never 180°. Secondly, if you line up the phase due to different distances from mic to sound source, then you’ll destroy the psychoacoustic panning effect which absolutely relies on that phase difference and that would be completely catastrophic in the case of a binaural 2 mic setup. Thirdly, in the two examples I gave we’re not talking about a 2 mic setup, we’re talking about a roughly 9 to 40 or more mic setup and if you line up the phase between any two mics then you’ll make the phase differential worse with the various other mics.

Why is artistic use of phase important here? let them do their thing

If you were a music or sound engineer you would know the answer to that question! For everyone else: With music, the most common use of flipping the phase (inverting the polarity) of a mic/channel is with a drum kit, where it’s quite common (though not always the case) to use two mics on the snare drum, one next the top/batter head and one underneath next to the bottom/snare head. This always results in a phase differential to the point of some amount of audible phase cancellation but just as the phase differential in this scenario is never 0°, it’s also never 180°. So one of the first things an engineer will do when commencing mixing the drum kit is to flip the phase of one of the snare mics to compare the resultant freq response, which is virtually always different. Typically the difference can be heard most obviously in the amount of low-mid freqs. One of the states, flipped or unflipped, will provide “better” phase coherence (be closer to 0°) and there will be significantly more low-mid freqs. However, it is sometimes not desirable to have significantly more low-mid freqs in the snare sound as this can overlap/interfere with the same freqs produced by other instruments or the vocals. Therefore, it is sometimes the case that the “worse” phase state is chosen, the one which results in more phase cancellation.

the point is absolute phase is audible, no matter how messed up the actual phase is of a particular song

No, the absolute phase is not audible. It’s audible with test signals but almost never with actual music recordings because the phase is so messed up anyway. There maybe some very rare outliers to this general rule, under certain specific unusual circumstances, but even if someone ever actually comes across such a rarity, the audible difference is not more than very subtle anyway, there are much bigger fish to fry.

just make sure you listen to it like the artist intended

But you cannot make sure of that. It is extremely unlikely you will have the same speakers and you definitely will not have the same room acoustics. These and other factors are bigger fish to fry, have far more effect than the absolute phase polarity even in those outlier cases where it might be detectable. However, “like the artist intended” (and the engineers/producer) is actually a range, because no one expects that consumers will have the same or even particularly near the same speakers/acoustics as the recording or mastering studios where the musicians and engineers created the recording and indeed, that’s why we have the mastering process in the first place. The difference with absolute phase is either non-existent, inaudible or on those extremely rare examples where it is audible, so subtle that it falls within the “intended range” anyway. And this is why, contrary to your assertion, it is not “a controversial topic”, it’s an irrelevant topic and, an actual engineer would know that!

G

 

[The Jester]

I have KEM speakers with a radial design that supposedly improves phase. I can’t discern any difference in that regard as opposed to my regular box speakers.

KEF’s “UniQ” design creates a point source but phase coherence between the drivers is still down to crossover design, I’d assume (always a concern) that KEF’s experience in speaker design would minimise that …
Also playing the devils advocate and assuming (there we go again) that the listening tests were done in a level matched DBX environment and not merely “how they sounded “ … ?

 

[The Jester]

I tend to measure a few things in a loop when I get new gear, to verify that nothing is obviously wrong and that I didn’t pay for fake marketing performances. If after the most basic sweep in REW, I see the impulse looking down, I’ll do something about it(so far it’s been some cable wrongly soldered and that happened twice in a decade of purchasing way more than I needed to, mostly among the cheapest offerings on the market).
I also got one IEM someone sent me to measure, where the left and right bass drivers were out of phase. No idea if that was on purpose or absent-minded soldering on that one pair. It sounded strange but oddly nice.
That is the full extent of my concern. It's one of the 1000 and 1 audible but irrelevant factors of my audio life. In this case, I can confidently say that I look it up as a maniac and not as a concern for a better listening experience.

Found out that too …a while ago I bought a 3 metre headphone extension cable to reach my favourite listening chair, not expensive but looked reasonably well made, cable covering, plugs etc and something sounded “off” … quick check with a DMM revealed it was reversing the left and right channels …

 

[bigshot]

I don’t have an anechoic chamber, so there’s no way to isolate phase effects to the speakers. I can only compare the fronts to the rears, and the main difference I can discern is that the KEFs have very wide dispersion.

 

[Ghoostknight]

Also playing the devils advocate and assuming (there we go again) that the listening tests were done in a level matched DBX environment and not merely “how they sounded “ … ?

Both, i just did the audiocheck.net blind test again but i cant get 10/10
first try was in a post before with 15/21 i think

 

[danadam]

Such tests should:

• either hide the current score until you decide you are finished,
• or ask for the number of trials before starting the test and then tell that you passed only if you did that many trials (and reached the confidence level, ofc)

Just continuing the trials only until you reach the confidence is cheating, imo.

Here, some random clicks :

 

[Ghoostknight]

You are free to invalidate the tests i did with thinking i did it with random clicks...

beside this, i could just do as many DBT till i got a 10/10... the only advantage you got with seeing the score is that you know when you have to reload the page lol
so i guess even with a proper DBT result you guys would questioned it...

you know i was into dice rolling some time ago with these scheme where you double your money if you lose, its not hard to get 15 coinflips in a row wrong with 50%/50%

i mean i did the test two times now legimately, atleast i know this (with some playing around before each (longer) test i posted) with 96%, i think the chance is pretty low that it were just guesses but do with it what you want

 

[bigshot]

The score is an average of all trials, you don’t just do it until you stumble across doing it right. Record all your tests and do a bunch.

 

[Ghoostknight]

https://github.com/jaakkopasanen/ABX

i might setup a blind test here tomorrow, i just have to see how i switch polarity of one song and where to upload these so it works since i dont have dropbox

 

[gregorio]

You are free to invalidate the tests i did with thinking i did it with random clicks...

Firstly, if a test can be passed several times with >95% confidence with just random choices, how is it not invalid?
Secondly, I’m not sure what the point of this test is anyway. No one is disputing that one can’t tell a difference in absolute polarity with a test signal or in certain cases with a mono (single mic) recording of certain instruments which produce an asymmetrical waveform. An acoustic guitar sometimes produces an asymmetrical waveform and a trombone is another instrument that also does sometimes. However, in a professional/commercial recording how often do you get an unaccompanied acoustic guitar or trombone that has been recorded with only one mic and is unprocessed? Is that common in your mind or do you think the number of occurrences is zero or very close to zero?

G

 

[Ghoostknight]

Firstly, if a test can be passed several times with >95% confidence with just random choices, how is it not invalid?
Secondly, I’m not sure what the point of this test is anyway. No one is disputing that one can’t tell a difference in absolute polarity with a test signal or in certain cases with a mono (single mic) recording of certain instruments which produce an asymmetrical waveform. An acoustic guitar sometimes produces an asymmetrical waveform and a trombone is another instrument that also does sometimes. However, in a professional/commercial recording how often do you get an unaccompanied acoustic guitar or trombone that has been recorded with only one mic and is unprocessed? Is that common in your mind or do you think the number of occurrences is zero or very close to zero?

What about electronic music? for example, specially bass of any kind, i dont think, unless intended, that the phase is inverted that often

i think a blind test is still a good idea, to proof to you guys that you not only can hear it with "test signals"

 

[bigshot]

In commercially recorded music, phase is generally all over the place. When you play it through speakers in a room, it gets even more scrambled. You don’t want phase cancellation, but that requires a perfect 180, which is pretty easy to avoid. I think you’re assuming that phase deviation is always bad and the ideal is for everything to be in absolute phase, but in music you pretty much only get that with mono tracks, which obviously aren’t ideal.

Just because you can hear something with a test signal that reveals it, that doesn’t mean that it’s an issue with music. Music is its own kind of signal. Too often audiophiles focus on things that are audible with test tones and never relate that to the real world playing music on their own equipment in the home. Ultimately, that matters a lot more than with test tones.

 

[Ghoostknight]

Just because you can hear something with a test signal that reveals it, that doesn’t mean that it’s an issue with music. Music is its own kind of signal. Too often audiophiles focus on things that are audible with test tones and never relate that to the real world playing music on their own equipment in the home. Ultimately, that matters a lot more than with test tones.

imo it depends on the music YOU are listening to, i have plenty of tracks where its audible

is it justified to, because of a few bad example, to say every music is inaudible related to absolute phase? i dont think so

 

[bigshot]

No it doesn’t matter what music you are listening to. You seem to think electronic music is like test tones and it isn’t. It’s just like music with all the same reverb and creative use of phase shifts that make these small insignificant things completely inaudible. These aren’t “bad examples”. They’re well engineered music. The goal isn’t to have all the phase in music neatly lined up. As I said before, the only music that does that is mono recordings. And that’s because with one channel, there is only one phase.

 

[Ghoostknight]

No it doesn’t matter what music you are listening to. You seem to think electronic music is like test tones and it isn’t. It’s just like music with all the same reverb and creative use of phase shifts that make these small insignificant things completely inaudible. These aren’t “bad examples”. They’re well engineered music. The goal isn’t to have all the phase in music neatly lined up. As I said before, the only music that does that is mono recordings. And that’s because with one channel, there is only one phase.

No i dont think "electronic music" is like test tones.... good (proper produced) music CAN BE, for example, quite impulsive bass hits, they are not that far off from the test tones... but anyway