thanks
Thanks, Elias, for taking the experiment to a logical end--though it might be no conclusion, I agree.
It has added much to the perception of people like me for whom DIY & listening are easier than the rarefied atmosphere of theory and maths.
I must admit that as someone who had all along felt that "something was the matter with conventional stereo", I did enjoy the great hands-on journey. Thanks again for the insights, and as they say, keep up the good work!
-- UKP
Thanks, Elias, for taking the experiment to a logical end--though it might be no conclusion, I agree.
It has added much to the perception of people like me for whom DIY & listening are easier than the rarefied atmosphere of theory and maths.
I must admit that as someone who had all along felt that "something was the matter with conventional stereo", I did enjoy the great hands-on journey. Thanks again for the insights, and as they say, keep up the good work!
-- UKP
Trifield = Gerzon's algorithms = linear matrix.
It does have divided frequency bands (two?) but it's not a miracle
In fact, if you scroll back this thread you'll see I'm also using similarly inspired psychoacoustic filter for the center vs side elements.
Also, it appears to me that similar concepts in Trifield can be found from Ambisonics as well, the master piece of surround sound theory
For practical experience you can ask diyaudio member optic who is the pro guru in Gerzonic matrixing !
edit: See patents US5594800 and US5671287. Patent Searching and Invention Patenting Information
- Elias
Last edited:
Sounds good, what could be better than finding an enjoyable audio experience via diy road
I believe that's why we are here in this forum.Can you share something about your system please ? We want pics !
- Elias
You have put it so nicely as regards DIY and its pleasures...Sounds good, what could be better than finding an enjoyable audio experience via diy road
Can you share something about your system please ? We want pics !
- Elias
Dear Elias, you wont believe me when I tell you that both my (old) fav speakers (large AR and bookshelf Infinitys) are in cold storage, wrapped up, after I listened to a simple open baffle, and in spite of the reduced FR etc etc, I havent been able to put them back. Then this thing about the problem with stereo and the attempt to build something that will be better than good from most (if not all) angles, though far from perfect, and enjoy listening... My current poison is a couple of front/rear identical driver sets (from about 100Hz to only about 9.5 kHz) in a shallow box connected as a true dipole (in order to alleviate the front/rear radiation inconsistencies etc, and the fond wish is to evolve something like a SSS, but with a leaning to the M-S approach. Everything else is secondary now for me. Perhaps I should not be posting such a personal note here, so anything more via mail to ukpanickar at yahoo dot co dot uk.
My apologies to all again.
Keenly following most things that are happening ...
@ Elias
Dont know if you have seen this thread:
http://www.diyaudio.com/forums/multi-way/169521-open-baffle-vibrations.html
Dont know if you have seen this thread:
http://www.diyaudio.com/forums/multi-way/169521-open-baffle-vibrations.html
Hi Prof,
The triple S or what i call the "Three some" (cause its every males dream). It not far from the M-S approach. I have a technical approach to things. However not a genius. I'm sure there is a perfect combination of the "Three some" and Gerzons work with ambisoncs. Im working on it!
I suggest you work on the same. if you can try what Elias and i believe has potential. Then we can learn more. I would love to hear what you think.
Cheers
Optic
Last edited:
nice name!
however I think that I like Sweet Sixteen (for 4x4 array) even more John Watkinson once asked Does humour belong in pro audio?
Perhaps it is high time to ask do girls belong in diyaudio?
Some unsorted thoughts ...
not discussing what is "right", but just showing variations how one
can modify reproduction assuming you have two broadband "true dipole"
speakers.
I guess Rudolf e.g. may come close to this ideal with his current
"swinging dipoles".
- You can setup the dipoles in a conventional stereo triangle without toe in.
- You can increase toe in until you reach "RLX" configuration
- You can put the speakers even closer together until the stereo base
gets rather narrow ...
- Now you can go from "toe in" to "toe out" ...
and we are IMO pretty close to the behaviour of the MS "family" of
configurations.
Unfortunately there seems a multitude of criteria to be important:
- Believable center phantoms (which was one main motivation for the
single speaker approach, as i understood)
- Tonal consistency of center and laterally shifted pantoms
- Consistency of apparent distance of center and lateral phantoms
(initial time delay gap ? direct/diffuse ratio ?)
- Size of the sweet spot allowing for "proper" or at least "good" imaging
- Size of the enjoyable "easy listening" area within a usual room, mainly
defined by balanced tonality and "spaciousness" and not by proper imaging
- Believabale apparent source width with respect to different
phantom sources and their location.
- Balanced excitation of room modes at lower frequencies
As soon as i have time and opportunity, i'd like to play with
"unconventional" setups and have a look at the above (or similar)
criteria.
not discussing what is "right", but just showing variations how one
can modify reproduction assuming you have two broadband "true dipole"
speakers.
I guess Rudolf e.g. may come close to this ideal with his current
"swinging dipoles".
- You can setup the dipoles in a conventional stereo triangle without toe in.
- You can increase toe in until you reach "RLX" configuration
- You can put the speakers even closer together until the stereo base
gets rather narrow ...
- Now you can go from "toe in" to "toe out" ...
and we are IMO pretty close to the behaviour of the MS "family" of
configurations.
Unfortunately there seems a multitude of criteria to be important:
- Believable center phantoms (which was one main motivation for the
single speaker approach, as i understood)
- Tonal consistency of center and laterally shifted pantoms
- Consistency of apparent distance of center and lateral phantoms
(initial time delay gap ? direct/diffuse ratio ?)
- Size of the sweet spot allowing for "proper" or at least "good" imaging
- Size of the enjoyable "easy listening" area within a usual room, mainly
defined by balanced tonality and "spaciousness" and not by proper imaging
- Believabale apparent source width with respect to different
phantom sources and their location.
- Balanced excitation of room modes at lower frequencies
As soon as i have time and opportunity, i'd like to play with
"unconventional" setups and have a look at the above (or similar)
criteria.
Last edited:
I see you put all the emphasis on imaging and give no value to spaciousness ?
Second point I'd like to make, and which you did not mention and what I think is the main failure mechanism of conventional stereo triangle, is unnatural interaural cross talk. The goal should be to minimize any unnatural cross talk !
The SSSx5 is a step in that direction.
- Elias
Seems you cought me
No, i do not want to put all emphasis on "imaging".
Spaciousness is valuable to me, but in my personal list
believable imaging is a prerequisite to believable spaciousness.
Which may not be in line with common use of the term ...
If i have a zone where there is both, i can live with an even
larger "easy listening zone" where there is just "spaciousness".
What do you mean exactly by unnatural crosstalk in conv. stereo triangle ?
Do you mean the left ear getting a (smoothly low pass) filtered and delayed version of the right speaker's
signal and vice versa due to HRTF ? Which of course differs from a real centered source ?
Or do you also mean interspeaker interference artefacts ?
Some advocates of stereo say, that crosstalk cancellation appearing towards HF is "healing"
the interference artefacts.
(Often one hopes one drawback should correct another drawback, but in this case i
have difficulties in understanding that ...)
HRTF
http://gilmore2.chem.northwestern.edu/images/hrtf.gif
In my view this is hardly a "crosstalk cancellation". Interference artefacts of
coherently radiating speakers cause cancellation or addition of left and right signal
changing within small distances at the listening area (at least when using sine).
The only healing thing here is IMO the quasi stochastic and broadband nature of most music,
also having a power spectrum which is usually LF dominant.
To me the major stereo problem is interference at HF.
But how would you describe the audible effect, that "unnatural crosstalk" has - especially
on centered phantoms ?
By unnatural cross talk I mean, like you say, the same centrally located sound arriving twice in both ear canals, and also the interference pattern.
I'm not sure what perception the twice arriving sound causes ? Maybe localisation of the tweeters ?
But the interference pattern is extremely annoying !
Also unnatural is the attempt of stereo to act as wavefield construction method for a phantom source. While there can be a propagating energy field at the tiny sweet spot at low freqs, the field is unnatural since the propagation speed is less than for a natural plane wave. v = f*l where in the interference field apparent wavelength is smaller and so is the energy propagation speed of the phase isobars. If one turns his head in the interference field, and even neglecting head shaddow, ITD would shift unnaturally compared to identical head turn in a natural plane wave field.
I have encountered that SSSx5 do not have the interference problem, so I'm very happy with it !
- Elias
I'm not sure what perception the twice arriving sound causes ? Maybe localisation of the tweeters ?
But the interference pattern is extremely annoying !
Also unnatural is the attempt of stereo to act as wavefield construction method for a phantom source. While there can be a propagating energy field at the tiny sweet spot at low freqs, the field is unnatural since the propagation speed is less than for a natural plane wave. v = f*l where in the interference field apparent wavelength is smaller and so is the energy propagation speed of the phase isobars. If one turns his head in the interference field, and even neglecting head shaddow, ITD would shift unnaturally compared to identical head turn in a natural plane wave field.
I have encountered that SSSx5 do not have the interference problem, so I'm very happy with it !
- Elias
The matrix is very simple. See post 97.
An externally hosted image should be here but it was not working when we last tested it.
You can wish to try the shelving filter, I find it useful. But it may depend on various issues, like the drivers, rooms etc.
Sure, fun it will be
- Elias
This is an unnatural *absence* of cross-talk.
(..taken to an extreme would be headphone listening without a binaural recording.)
This is an unnatural *absence* of cross-talk.
Eh the centrally located natural sound should appear only once in both ears, one time in left and one time in right ear. If it appears twice it is not centrally located natural sound, but a source located off center, yes the stereo speaker locations.
- Elias
See this recent post of mine:
http://www.diyaudio.com/forums/multi-way/201081-disappearing-act-11.html#post2801744
I am sure our collective approaches will sooner realize the 'male dream' and lead us to audio nirvana of some sort...
In this context, in the light of my own misdirections and follies, and tapping into the sagely observations of many, I would like to once again stress the need for proper tools and source material to evaluate our designs. I surely have in mind the nicely evolved, though old-fashioned, pair of appendages on either side of the grey matter that does a wonderful job of real-time neuro DSP. (I am solidly in the 'hearing is believing' camp, to a large extent.) But surely we need to have suitable and 'proper' source material for evaluating any new topologies that we might be pursuing.
Current recording practices, IMHO, leave a lot to chance as they are all 'optimized' for the stereo triangle and that pain in the neck, the 'sweet spot'. Even many 'purist' tracks are guilty of this, where often an otherwise pristine signal is ruined by the vocalist being close-miked and then 'fed into the system' God alone knows how. :-( Sadly, all 'truth' is lost as a result. With all humility I dare to suggest that even Linkwitz, that very practical modern-day audio 'Guru', has once or twice been misled in his judgment by such industry practices, if I recall some of his comments correctly.
So my take is that if you are going in for SSS, let us have a simple M-S recording (with no fancy electronic wizardry, and in its own component form) of a trio or a quartet, preferably with voices, to judge our humble babies with. That means the same source material could also be used to directly drive your single-point transducers, if you are dabbling with M-S.
Does anybody know of such simple M-S component sources that could be downloaded for our experiments? Samples of a few seconds length are there; we need at least a few minutes, I guess.
So, here is to that ultimate liberation from the strait-jacket of the 'sweet spot' !!