One must understand the difference between interchannel time differences contained in the recording, and the low freq ITD generated at the listening position in the interference field from stereo amplitude panning.
The latter exists also without the former. Low freq ITD is present with pure amplitude panning in a stereo triangle. This is how Blumlein intended it !
What comes to Watson, nearfield interference field is close to the ideal stereo field at the low freqs and this allows ITD perception to be more accurate, purely from amplitude panning to ITD.
- Elias
See: British patent 394325 Alan Blumlein 1931
Really, stereo triangle produces low frequency ITD from pure amplitude panning due to interaural cross talk when listener is located at the stereo sweet spot.
The low freq ITD therein produced is more correct if less room influence is present. Thus Watson works being able to lessen the room influence due to nearfield conditions.
- Elias
The action starts at page 2 line 10 onwards.
If intuition is useful, think about ideal crossed dipole mic, X-Y mic: It have no phase difference at low freq (and if really ideal, there is no phase difference at any freq) between the channels. But when such low frequency signal is played through a stereo triangle, listener at the midway between the speakers, and room influence neglected, there will be localisation sensation perceived. Also we know that at low freqs head diffraction is enourmous so there will be practically no amplitude difference between ear signals. Then only way to have localisation cues is ITD produced by cross talk.
- Elias
If intuition is useful, think about ideal crossed dipole mic, X-Y mic: It have no phase difference at low freq (and if really ideal, there is no phase difference at any freq) between the channels. But when such low frequency signal is played through a stereo triangle, listener at the midway between the speakers, and room influence neglected, there will be localisation sensation perceived. Also we know that at low freqs head diffraction is enourmous so there will be practically no amplitude difference between ear signals. Then only way to have localisation cues is ITD produced by cross talk.
- Elias
That's probably a bit of an exageration . . . I'd bet that more than 1% of classical music recordings (all that Linkwitz is interested in) use an ORTF pair (augmented, perhaps, with a spot mic here and there). And those that do (use a near-coincident pair for primary capture) do carry interchannel time differences . . . exactly what Linkwitz is looking (listening) for with the Watson idea. Any additional benefit for synthetic (studio) stereo, if it exists, comes from some other effect(s).
Whether it ever becomes popular . . . well . . . not likely. I'm certainly not interested in "localization effects" that work in one seat only (and don't move your head too much) except as a curiosity, since precise spatial localization of instruments adds nothing to my enjoyment of the music. The phony effects of "multi-channel" (apart from movie sound effects) actually detract from my listening, so I don't go there (for the half dozen recordings of interest that might exist) either. But what Linkwitz is demonstrating is an interesting effect . . . and hints that, by some accident of process, there is more positional information in (some) conventional two channel stereo than we usually extract. That's not something that I would have predicted . . . especially since binaural (headphone) listening to the same sources typically does not produce an externally projected image.
See: British patent 394325 Alan Blumlein 1931
Really, stereo triangle produces low frequency ITD from pure amplitude panning due to interaural cross talk when listener is located at the stereo sweet spot.
The low freq ITD therein produced is more correct if less room influence is present. Thus Watson works being able to lessen the room influence due to nearfield conditions.
- Elias
Sooooo, why not just move the listener closer to the speakers. In my room, 16' x 20', I have me speakers (NaO II currently) separated by about 8', 4' from the side walls and 5' from the wall behind the speakers. I normally sit about 12' from the plan of the speakers. However, I often move to a position about 5' from the plan of the speakers. This has several effects. Among them, the direct to reflected sound ratio increase, the included angle between the speakers at the listening position increases significantly with result that the shadowing of the head increases. The effect of such positioning of the listener seems to me to be very similar to what is claimed for Watson: Greater width/depth of the sound stage, better localization, better voice inteligibility,....
On a related topic, something I have contemplated for a number of years but have never gotten around to trying is an experiment to get the head phone image out of the head. I have theorized that if for the head phone right channel I mixed the stereo left channel, with suitable low pass filtering (related to the HRT), attenuation and delay, with the virgin right channel, and the inverse for the left channel that this should mimic what we hear with stereo, scans room effect, and might fool the brain into thinking it was listening to stereo speakers. The next step would be to sum the L and R, process the sum with delay and reverberation to mimic room effects, and then add them to the L&R for head phone listening. In this manor there would be control over ITD, HRT and a simulated venue reflected sound field.
Last edited:
Like you noted direct to reflection ratio increases in that case. This consequenses the trade off: Imaging or spaciousness. A fine balance between these two is the Holy Grail of two channel audio.
Another issue, in my experience too high direct to reflection ratio makes the speakers localisable. Linkwitz also notes this with Watson (thus highs must be rolled off). I believe speaker localisation is more common problem than generally accepted.
Watson cleverly enhances the freq range where enhancement is needed: low freqs.
- Elias
Sounds like you could benefit from
http://www.diyaudio.com/forums/multi-way/200040-stereophonic-sound-single-loudspeaker.html
- Elias
I don't find that to be the case. Perhaps it is a side effect of dipole speaker and the reflections from the wall behind the speakers. When I move closer I find the sound stage broadens and I don't sense a loss of spaciousness but and increase, nor do I experience that the location of the speakers becomes more evident. On the other hand I don't find that localization of specific instruments in a large orchestral work is real. I certainly don't hear that at a live event. I believe that the conventional sense of imaging taken to mean that each instrument can be pin pointed in a 3-D space is just artificial. Never heard that is a real venue with full orchestra.
I don't know that what I hear is in anyway related to what this Watson thing is supposed to be about, but I agree with Dewardh. I'm not inclined to clutter my listening room with "enhancement speakers" that require a fixed listening position and head in a vice.
You know what comes to mind when I think of Watson is where it could go. I get this picture of Captain Nemo, or Hell Boy fighting the giant squid or that multi-tentacled beast where each tentacle has a speaker on the end of it. No disrespect intended, but that is way comes to mine.Fighting the audio monster. I'll stick with 2 channel and if I want to knock down the walls I'll turn on the DSP-1.
I was talking about 99% of all recordings, not just classical music recordings.
If spaciousness is all you're interested in then you should use a system that adds lots and loud room reflections preferrably from the side and elevated locations. Of course now imaging will suffer. A "Watson" can probably bring back better localization but the price is "head in a vise". Pick your poison.
There's rather more to it than that . . . and there's no *necessary* conflict between "spaciousness" and "imaging" (dipoles in a decent room comfortably give both) . . . the necessary "clues" for both are present in many room/recording combinations. Which is, more or less, what Linkwitz is demonstrating with Watson by extracting even more of the already present "localization" cues.
I've captured and A/B'd different stereo setups with differing reflection patterns and found that there's always a tradeoff between imaging and spaciousness.
Try it for yourself and make some binaural recordings in your room. Compare your dipole with and without the strong backward reflection (use a absorbing panel).
That is the oldest reference I'm aware of. The stereo we know today is based on that. Pure low frequency amplitude panning is transformed to low frequency ITD at the listening position by interaural cross talk of the interference field. What kind of explanation are you looking for ? Are you having any doubts stereo does not work ?
- Elias
That is the oldest reference I'm aware of. The stereo we know today is based on that. Pure low frequency amplitude panning is transformed to low frequency ITD at the listening position by interaural cross talk of the interference field. What kind of explanation are you looking for ? Are you having any doubts stereo does not work ?
- Elias
You're the one that has problems with phantom source localization, not me.
I just asked for reference where this process of "Pure low frequency amplitude panning is transformed to low frequency ITD" is explained.
Ok, in the patent it is stated "It can be shown that..."
But no proof of that 'show' ?Maybe the answer is so simple that no peer reviewed article is not needed ?
How about DIY, within the fanatic spirit of this forum ?
The answer is in the interference field (vector summation, even free field will do at low freqs) around the stereo sweet spot, and then employ the basic principle of diffraction.If it's out of the question, then search articles about vector analysis of stereo field. I'm sure they exist since the early days. I remember papers from Gerzon and Farina et all that deal with the matter.
- Elias