Never mind the protocol, just go ahead and ask. You can even hint at the fact there's competition lurking. AFAIK, there's still strong demand in the technical/scientific field; another field worth looking into is AI.
Yes I am completely serious and was willing to help you whilst going along with your little game of playing stupid, I now realise that you are not pretending. Are you saying that you are unable to distinguish acoustic absolute polarity ?.
Dan.
Quick Links -> Subscribed Threads
Select the thread(s) then under Selected Threads, choose an option
Farewell.
I'm happy to keep playing if you are? Absolute Polarity Blind Listening Test
Imho those 'special' signals are more common that you give credit for. Vocals, especially spoken word in non reverberant environment are very clearly audibly asymmetrical waveform and these and other natural asymmetric impulsive sounds are perfectly distinguishable on simple but clean speakers like good full range. Multiway loudspeakers can introduce phase anomalies that serve to confuse timing information in the crossover overlaps.
Dan.
JN/Zung thanks. I at least know what to say to her even if I have no grasp of the research topics.
I now return you to the 'phase is obvious you must be deaf/have a rubbish system if you can't hear it' discussion.
She's welcome to PM me as well.
OK, back to the witch hunting...
Pavel, I just ran a headphone test where I put three copies of each file into VLC Player on repeat and listened and noted differences and mentally labelled the two files as 'sharp' and 'dull'. I found the one second file time is too short but three consecutive plays of the same file is useful. With my eyes closed in a dozen experiments I was able to reliably decide which of the two file types was in play. This test confirms for me that this change is audible, and that the one second or so sample time is too short for reliable detection. Real impulsive sounds with natural ringing and natural decay are completely different to reference square pulse wave train. Try applying your all pass filtering to a kick drum or tom recording and let's see how it sounds, ditto spoken word recording. Pavel, I am not setting out to rain on your parade but I find your too brief one second sample time and the signal content are generators of false negatives and hence your test is not revealing of the truth for naturally generated asymmetric waveforms.
Dan.
Dan.
Last edited:
I wouldn't call him a witch, a fool maybe to call me stupid because I queried his inverting one speaker test set up as having any useOK, back to the witch hunting...
I used to think the same for many years but once I did rather rigorous tests I concluded otherwise.
At low freqs (< 300Hz or so) our ears are reacting to the actual waveshape even for pretty natural steady state signals. Try this: Add the first overtone to a sine, equal level, but detune the overtone's (or fundamental's) frequency by 0.5Hz and listen (a signal similar to the decay portion of a bass guitar). The detuning rotates the components slowly against each other, the phase difference spanning 360° every two seconds.
If you don't trust this shortcut technique to shift phases, create a bunch of those signals with exact H2 but shifted in phase in discrete 45° steps. Listen and you will find a set of two with a 180° difference between them where one will sound weaker and the other will sound stronger "than average". Then take the orthogonal set (both shifted by 90°) and re-test, these will now sound the same. Which set will cause the greatest differences depends on the actual ear signal, which may have undergone additional phase shift from high-passes and crossover filters (it's best to use good headphones, which should be low distortion as well, and use a very moderate to low listening level).
Note that with actual music signal it depends on both the content and on the phase response of the total system whether a polarity inversion is audible from this effect. If flipping polarity results in the H2 component appear switched between -90° and +90° @eardrum there's little chances this is audible, but when if we happen to get the full 0° to 180° waveshape change it can be quite audible.
Music with soft but full sounding bass drums, and bass guitar with rolled off treble works well.
And this is only one aspect of phase distortion and polarity inversion. With most recordings that contain a natural "room sound" and reverb there are changes in the soundstage depth perception, size of phantom sources, things like that.
Certainly not "worlds", but funny enough I have to sort of agree with Dan that identifying phase/polarity switching is sort of of baseline test I almost always use at the beginning of A/B comparisions, ABX runs etc. Whenever I have a hard time getting this right I know I don't need to try listen for subtleties as I'm not ready for it at the moment, physically and mentally. Which underlines we're talking about very low-level minor effects, no big deal at all.
And of course the worst-case changes that a cable direction swap could actually produce are many orders of magnitude smaller than these gross changes to the signal....
- Status
- Not open for further replies.