Another way to put it: make the field more diffuse at high freqs.
It's easy for you to test by yourself
Only you can answer the question how you will perceive it.For me it's no high frequency localization at all like in a diffuse field. But this does not matter in a phantom imaging sense, because the stereo phantom are more strongly perceived at midrange. But it prevents perception of the two tweeters.
And also because of:
The rest of the 50 % will not lose phantom imaging because it is mostly a midrange phenomena anyway.
The high part of the sound mostly coming from the two tweeters, of course !
- Elias
Please define the frequency range you're talking about.
It's easy for you to test by yourself
How would a suitable test look like?
If you look at the pictures of his room set-up, its clear that he also advocates speaker placement that puts the speakers well out into the room, a long way from any walls, and he has a fairly large room in which to do so, with primary seating locations relatively near to the speakers. He also has a lot of natural diffusion in the room.
This will lead to quite a long time delay to first reflection, as well as a relatively high direct to reflected ratio even with very wide dispersion speakers, so it doesn't contradict anything I said about a high direct to reflected ratio generally being more revealing of what's on the recording. The only real difference is he's used room placement within a comfortably large room to achieve it rather than high directivity.
The fact that he can't easily distinguish between the Orion and Pluto despite rather different polar patterns is a clear sign that they're well away from any walls and he's sitting quite close to them.
This is a fairly different situation than many of us are faced with. In my current room I can only get the front axis of my speakers slightly under 1 metre from both side-walls and front wall - the left speaker is 90cm from driver centre to the side-wall, right speaker 70cm, both 90cm from front baffle to front wall, yet I am listening 2.5 metres away.
I would not get much in the way of precise imaging from that distance with either a Pluto or an Orion, I doubt an Orion would sound good less than a metre from the front wall due to its dipole radiation, likewise I don't think a Pluto would sound very good only 70 and 90cm from the side-walls.
As the listening distance increases and the distance between speakers and room boundaries decrease, narrower dispersion becomes more optimal. Any polar pattern can sound good if you have a fairly large room, speakers are a long way from the boundaries, and the listener is relatively close to the speakers.
When talking about the frequency range from 2Khz to 8Khz, i can see
advantages in a diffuse radiation by lowering the risk of
- pinna cues of direct and early reflected sound contradicting
phase information derived from bass (above Schroeder frequency)
and midrange predominantly (in broadband sound sources like an orchestra e.g.).
- interference artefacts between the stereo speakers messing up
intensity information between channels which is important for phantom
localization above 2Khz
Howewer, if the diffusivity is introduced by room reflections
(and suppression of the direct - non diffuse - radiation), the
system's performance will be very dependent from the
setup in the room:
L-R symmetry, delay of reflections, absorption coefficients, ...
Which would contradict the idea of proper "reproduction".
So i propose diffusivity to be builtin to the speaker in a defined
manner.
Last edited:
I think it is individual dependent ! Here in this thread also we can read peoples comments varying from 1.5kHz to 6kHz upwards. You got to find your own limit.
A starting point could be a good dome tweeter able to play down to 1.5kHz and it being toed-up. The toe-up angle needs to be verified, but I find good results with toe-up between about 30 - 60 degrees, depends on the listening distance (and of course ceiling height).
Try to rotate your head at the listening position while tuning the high pass corner freq of the tweeters. When the sound does not change much with head rotation, the freq limit is found.
- Elias
Good Oliver ! Finally someone who understands !
Maybe make it more clear in other words that the contradiction is between the midrange phantom image at the center and high freq pinna cues at the speaker locations (at +/-30 degrees).
Yes the challenge is there when using the room as an integral part of the reproduction, however it is not too great of a challenge since leaving ceiling and side walls reflective is enough ! Do not try to absorb early reflections.
But is the human perception defined ? Since there is great variance between individuals in perception, optimally the system should be tunable.
- Elias
Maybe make it more clear in other words that the contradiction is between the midrange phantom image at the center and high freq pinna cues at the speaker locations (at +/-30 degrees).
Yes the challenge is there when using the room as an integral part of the reproduction, however it is not too great of a challenge since leaving ceiling and side walls reflective is enough ! Do not try to absorb early reflections.
But is the human perception defined ? Since there is great variance between individuals in perception, optimally the system should be tunable.
- Elias
I don't get it. This will create strong early reflections from the ceiling. Why would this be desirable? I thought you wanted a diffuse field? I don't see how to achieve this in a practical manner. The sound field within a acoustically small room is always highly directional with strong early reflections.
Last edited:
I haven't noticed that, however perhaps I just don't have this:
of Yours, not being used to those OB-arrays
Then I suggest you second time to DIY some dipole line arrays !
To at least have a reference what you'll be missing with the flooder approach. I don't claim you will like them or that better, but to know more about what is possible.But the efect of them is not "being used to", because then many years ago when I switched the signal to my dipole line arrays for the first time, and I remember it like yesterday, the amount of small detail from the recording was so enormous it could be perceived at once by every unbiased listener.
Yes FE126En (and not the idiotic enabled but the unique new fostex line En) may not be directive enough, but it was the biggest of En series without whisser, and I hate whissers ! (It's psychology I know)
Certainly if one aims to prove a failure of a concept, such a conditions can be found. As in this case certainly a such music where they are perceivable do exist. But flooder is better than direct firing tweeter sure for most of the music in order to not to perceive the speakers.
- Elias
My purpose is not primarly to make a strong ceiling reflection, but it comes as a side effect, but to make tweeter more 'omni' in lateral domain. Please also don't consider dome tweeter as the ultimate solution, I suggest it as a starting point of exploration. Secondly, no absolute diffuse field is necessary, but as I said to make the field more diffuse than direct sound.
Happy experimenting !
- Elias
My purpose is not primarly to make a strong ceiling reflection, but it comes as a side effect, but to make tweeter more 'omni' in lateral domain. Please also don't consider dome tweeter as the ultimate solution, I suggest it as a starting point of exploration. Secondly, no absolute diffuse field is necessary, but as I said to make the field more diffuse than direct sound.
Happy experimenting !
- Elias
Well I've been through all of that and it all boils down to single strong early reflections creating increased ASW and image shift.
For unknown reasons you do have a problem perceiving high frequency phantom sources and the simple solution is more speakers or speaker crosstalk cancellation techniques.
Forget about the pinna cues for stereo
Both from "On the role of spectral pinna cues in human sound localization":
http://repository.ubn.ru.nl/bitstream/2066/18958/1/18958.pdf
Both from "On the role of spectral pinna cues in human sound localization":
http://repository.ubn.ru.nl/bitstream/2066/18958/1/18958.pdf
Attachments
Why forget ?
The data of your reference is based on 7 subjects in anechoic chamber.
They even conclude : "spectral cues may contribute to azimuth hearing".
And then they just present an presumption about the weighting.
- Elias
I did but obviously you think of something different than I do when picturing a "sparse floor-ceiling line array". What does it look like?
Yes, level and time cues are strongest and the "problem" with stereo isn't wrong pinna cues (although they might play a role) but speaker crosstalk. That's why the ITD and ILD numbers in stereo don't match those in perception of real sound sources.
Don't tell them se secrets ,hehe
My tweeters in the coax driver, positioned 1m above the floor, fire upwards by 15deg aiming to just below the upper opposite corner, whereas listening angle is about 10deg down but "otherwise straight on-axis", so I'm listening to a total 25deg off-axis sound and additionally have most of the HF energy directed to the upper third of the room (3.2m height) where no damping and diffusing materials are used.
This increased both lateral sweetspot stability and balance between direct and diffuse HF. Some of the effect might come from the rather jumpy HF response close to 0deg on-axis angles typical for a coax driver, and toe-up migitates the ill-effect by using a range of incident angles (at L.P.) that have less shift in the responses.
With Markus' test signals, none of them falls apart although the sweetspot narrows down as predicted, only when I filter above 10kHz two distinct sources become more and more apparent. Above 14k any sensation of direction is lost. My system is optimized to excel specifically in this regard (early IACC above 90%), so no magic and no deafness involved here...
- Klaus
As I find it interesting I took deeper look at the paper. I notice I made an error: the data in azimuth test is not based on 7 subjects but on 4 subjects ! The two authors, and two laboratory stuff. It was a poorly financed research team, but that is so typical
Tests were anechoic, yes.
However the test did not test for phantom imaging but is based on localisation of a real sound source.
With this setup, together with cross-talk, they can only conlude that pinna localisation is not more dominant than ITD or ILD. However, the absolute level of ranking or importance was not tested because they did not test the oppisite where cross talk would have been removed and localisation is based only on pinna cues. It could have been possible with other arrangements, but they did not do it.
So we can only say that removal of pinna cues (with molds) did not effect the localisation of a real sound source in binaural condition in an anechoic chamber using wide band noise signals
- Elias
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.