I have five identical speakers. Agree that it's a very important thing for good surround sound.
We can see the difference but the point is - can we HEAR the difference?The first loudspeaker is coherent, the second is not.
dave
Most reserach to this topic clearly says no. I didn't have done my own listening tests - so need to trust them meanwhile.
Our ear doesn't detect soundwaves as we see it here in time domain - it detects frequencies with it's hair cells. The brain does the timing - and that's pretty complex.
Possibly or perhaps we have our lines crossed about what is being discussed.
Despite my stating that is was at low frequencies in some speakers? Where would you consider the audibility threshold to lie around the crossover frequency between tweeter and midrange?
I clicked the link but it didn't seem to be about wideband drivers, coherence, audibility thresholds, etc... so I left after about a minute.
Signal processing is unlikely to fix the phase shifts caused by physical offset/s between 2 or more speakers. If the crossover is tuned for perfect summing in one direction, there will be destructive interference at other angles.
This is the other side of the coin, where we were at crossed purposes yesterday. The wavefront implies multiple measurements of the same thing at different locations, whereas the waveform at the listening position implies measuring different things at the same location.
To make a good coaxial system involves dealing with shortcomings like any other system.
gedlee said:
As I've stated that I use DSP to correct phase among other things, so I don't think you should ask me. I don't have to question that. Aside from that fact, I do use wideband drivers, remember? So I don't have any tweeters. But this paper would be a start.
This video was linked because of its relevance to the perception of phase at higher frequencies than most would deem possible.
It gives valid info about the audible results of that and why we as the listener should care. Even if it isn't about speakers, it's still relevant to our hearing abilities.
For the record, about half way he's using a couple of wideband speakers to demo some effects. 😀.
True. But... what Griesinger present is of interest for loudspeaker reproduction too.
It took me a while to get it i must admit as it's a bit of a stretch at first ( at least it was for me).
In fact it* gives a way to deal with the issue you face in small room and in the method Wesayso applied for his room it is brilliant to have made the link between what Griesinger present as a starting point of his observation (capacity some engineer have to locate microphone in location giving the feeling of proximity and the involving effect of enveloppement this location brings too) and loudspeaker reproduction.
It's so obvious for Wesayso he often forgot to talk about it, or it's not easily understood from others: it's only a part in his whole approach.
* What Griesinger is interested in as a whole atm, not nescessarely what is presented in this video even if it's a part of it.
And as a side note: no need to bother anyone i talk about if you don't like me or what i say. I'm no part of any gang, don't have anything against people i disagree with...or whatever make you paranoid.
Thanks for the link. I have only briefly skimmed it and look forward to reading it more fully when I start pulling together this aspect. The quantity they are measuring is not a direct measure of the audibility of the phase/time deviations. It is more severe/exaggerated but hopefully related though how isn't immediately obvious but may become so with further thought. The results for real and simulated speakers are not as expected which the authors seem to recognise as an issue to be addressed/understood. Their conclusions seem reasonable in claiming not to measure the threshold but something tighter. It seems likely to be reasonably in line with my current understanding and what was posted above: pay attention at low frequencies but higher is normally OK.
Thanks that is useful. I will look at one or two of his papers and try to assess competence. He is likely fine but unfortunately there is a long history in the psychoacoustics field of people making claims for the audibility of quantities below accepted thresholds based on incompetent experiments and/or understanding of what has been measured.
As the man behind many Lexicon reverb algorithms and formerly Harman's goto person for anything perception related (responsible for Harman's Logic 7 among other things) I hope he passes your test. 🙂
He's no stranger in the world of audio, yet he seems to have stayed pretty hidden from the HiFi crowd. I think I found him through something Geddes mentioned in a post quite a while ago (about 10 years ago or longer) and have been studying his material ever since.
Last edited:
I still fail to see why a multi-way system would have more trouble in retaining phase relations than a fullrange speaker.
Last edited:
Because with one, you need to be confined to a sweet spot that is a point in 3d space. With the other, the sweet spot spans a 2d plane with flexible height and depth.
Some Griesinger paper, which I'm unable to find at the moment with my mobile, writes in the abstract something along that "coherence is lost due to room reflections, because phase information is scrambled". Anyone interested on the subject might want to look for that paper, please provide link here if find it, I'd like to read it myself too 🙂
If it was that he wrote, and if we relate it to this discussion and to fullrange / multiway speakers then all it says, with my logic, is that narrow directivity works better. Narrow directivity would increase direct to room sound ratio, which would "increase coherence" if one wants to use tge term. Fullrange drivers make easy to achieve narrow directivity, while multiwayspeakers can also be narrow directivity, although not as narrow but more uniformly so. Bottom line is the room ruins it, better keep listening distance short enough to keep coherence up on some required level for good perception 😉
Well, perhaps should find the paper first before hasty conclusions 🙂
If it was that he wrote, and if we relate it to this discussion and to fullrange / multiway speakers then all it says, with my logic, is that narrow directivity works better. Narrow directivity would increase direct to room sound ratio, which would "increase coherence" if one wants to use tge term. Fullrange drivers make easy to achieve narrow directivity, while multiwayspeakers can also be narrow directivity, although not as narrow but more uniformly so. Bottom line is the room ruins it, better keep listening distance short enough to keep coherence up on some required level for good perception 😉
Well, perhaps should find the paper first before hasty conclusions 🙂
Yeah cannot seem to find it.. probably my memory is failing. Although, many of his papers mention phase coherency. I could think about it in a pointsource / multiway speaker context, even if direct sound would be coherent in a way that harmonics arrive as they should directly from speaker, but if its a multiway speaker with stacked drivers and think about very early vertical reflectins then harmonics are split at crossover to two physical distances to ear and the coherency is broken. So, use acoustic treatment on the ceiling to reduce its effect. Or, if you have great acoustics in lateral plane, longer path lengths and so on, attenuation, then just put the speakers sideways and phase coherence is retained with the vertical reflections.
Perhaps studio folks use monitors over console top sideways, very early reflection on the desk doesn't ruin it. Perhaps this is important stuff.
Thinkin a minute longer, also edge diffraction would affect, diffraction has different fingerprint on multiway speaker, which would also mean harmonics of an instrument that is split to the two drivers, although the drivers output would be in phase if crossover is so but if the edge diffraction is different between the two also the harmonics (any sound) would be. Point source would fix this, diffraction would largely be more similar when sources colocate. Although size of the colocating transducers would have different "beaming" and thus different diffracting bandwidth. Multiple entry horn would do nicely I think. Or multiway speaker with low diffraction and ceiling treatment. Or single fullrange driver speaker.
ps. site akutek.info seems to host many Griesinger papers, find them by googling "site:akutek.info griesinger"
Perhaps studio folks use monitors over console top sideways, very early reflection on the desk doesn't ruin it. Perhaps this is important stuff.
Thinkin a minute longer, also edge diffraction would affect, diffraction has different fingerprint on multiway speaker, which would also mean harmonics of an instrument that is split to the two drivers, although the drivers output would be in phase if crossover is so but if the edge diffraction is different between the two also the harmonics (any sound) would be. Point source would fix this, diffraction would largely be more similar when sources colocate. Although size of the colocating transducers would have different "beaming" and thus different diffracting bandwidth. Multiple entry horn would do nicely I think. Or multiway speaker with low diffraction and ceiling treatment. Or single fullrange driver speaker.
ps. site akutek.info seems to host many Griesinger papers, find them by googling "site:akutek.info griesinger"
Last edited:
A fullrange speaker has time/phase discrepancies at the low and high frequencies due to driver stiffness and mass plus cabinet loading resonances at low frequencies. A multiway speaker also has time/phase discrepancies introduced by the crossover filtering.
A crossover/signal processing can be used to correct phase anomalies but this tends to involve altering the signal over a range of time not just at a single value. If this alteration starts adjusting the signal during the silence before something like a click/transient arrives this is a mechanism for introducing audible anomalies despite the phase being "perfect". Additionally for, say, a tweeter and midrange if the crossover filtering is designed to move the cones in opposite directions before the click arrives in order to sum to zero on-axis this generally won't happen off-axis for non-coaxial drivers where the path lengths from the drivers to the listening position will differ changing the relative phase and breaking the cancellation.
The ear/brain isn't particularly sensitive to timing anomalies like this because the cochlea effectively works more in frequency space than time by selectively waggling hairs based on frequency (it then sends nerve impulses to the brain acting more in the digital space then analogue if you want to tease audiophiles further). However there is a limit and heavy handed use a few decades ago gained something of a bad reputation and I suspect may have had a role in the relative lack of commercial success of products like the K&H (now Neumann) O 500C.
A fullrange driver becomes an array of sources at higher frequencies. In cases where the lobing pattern changes with frequency, the harmonic structure is going to vary with angle. This is without considering breakup.
There are many many papers linked on his own site, if you're willing to scroll down enough.
A flawless speaker means nothing, if it can't work with (or in) the room you have. The commercial offerings can't cater to your room specifically, they have to cater to all, make something that they believe is made with the right compromises. But the DIY crowd that make their own can look at their own room and use that knowledge, to make sure the room + speakers can work together, whichever concept you may prefer.
So when we're talking about the flawless speaker, it could be a different one for all of us, due to our room and personal preferences.
So you argue that the off-axis behavior of fullranges is much better? It could be, but in the horizontal plane, it almost always is worse than that of a multi-way system. And I hesitate to believe fullranges, especially bigger ones (>3”), do a better job off-axis, when I look at my measurements of common midranges. I have some fullranges lying around, could put them on the test some time.
Found it, kinda, Griesinger paper/lecture about phase and coherency and audibility, also about what planet10 posted few days ago.
The audibility of direct sound as a key to
measuring the clarity of speech and music"
Edit time over, gotta add that this is about the most important "hifi" aspect on my system I've noticed, and happens only up to about 2.2m listening distance in what is quite normal living room without any acoustic treatment, with quite narrow directivity speaker system. If listening distance is any longer, sound gets worse as described in Griesingers paper.
I don't know how others have this aspect in their systems, but its night and day difference in my opinion on the stereo effect side of things what goes to "hifi". Unless you, yeah you reading this, haven't heard this transition ever I challenge you to experiment with your listening setup, find out if you can perceive the trasition the griesinger paper is about and post about it here or somewhere else about it 😉
Well, this could be just some problem in my setup so all of you might be fine already.
But, if my setup is fine it would mean the phenomenon is common in typical living rooms and many people woukd not have very good sound at home, without acoustic treatment / narrow coverage speakers, or quite short listening distance / small stereo triangle.
It would also mean that hifishow room setups with 10+ chairs in untreated rooms few meters away about none of them have a good sound unless its basically a horn speaker system. Oh well, perhaps dipoles, cardiod systems with some waveguide, perhaps arrays, have high enough DI so that listening distance is about 2meters, perhaps more. Maybe big enough fullrange drivers, big baffles could extend it too. Well, anyway 😀
Something to think about. Want better sound? skip purchasing new stuff untill you've found/heard the effect and be sure to have listening distance short enough to have good sound. If listening distance ends up shorter than you'd like, then its time to purhase higher DI speakers, not new amplifier or DAC. Nor better tweeter or woofer, but a waveguide.
I don't know how others have this aspect in their systems, but its night and day difference in my opinion on the stereo effect side of things what goes to "hifi". Unless you, yeah you reading this, haven't heard this transition ever I challenge you to experiment with your listening setup, find out if you can perceive the trasition the griesinger paper is about and post about it here or somewhere else about it 😉
Well, this could be just some problem in my setup so all of you might be fine already.
But, if my setup is fine it would mean the phenomenon is common in typical living rooms and many people woukd not have very good sound at home, without acoustic treatment / narrow coverage speakers, or quite short listening distance / small stereo triangle.
It would also mean that hifishow room setups with 10+ chairs in untreated rooms few meters away about none of them have a good sound unless its basically a horn speaker system. Oh well, perhaps dipoles, cardiod systems with some waveguide, perhaps arrays, have high enough DI so that listening distance is about 2meters, perhaps more. Maybe big enough fullrange drivers, big baffles could extend it too. Well, anyway 😀
Something to think about. Want better sound? skip purchasing new stuff untill you've found/heard the effect and be sure to have listening distance short enough to have good sound. If listening distance ends up shorter than you'd like, then its time to purhase higher DI speakers, not new amplifier or DAC. Nor better tweeter or woofer, but a waveguide.
Last edited: