I doubt that a real head reflects as much sound as that applet shows. Boundary conditions in these kinds of simulations is a difficult thing to get right and little differences can make large differences in the results. Impulse response measurements of the head show that the incident wave wraps arround the head and arrives at the other ear, but it does not appear to make it back to the incident ear again, it is diminished too much.
It does not work ! Stereo works only when very specific situations are fullfilled (sweet spot, geometrical symmetry etc..) and even then the illusion is very fragile (no head turning etc).
Sure ! Absolut majority out of earth population of 6e9 do not sit in the sweet spot while listening.
Then it is questionable if the phantom imaging capability shoud be the driving force of a stereophonic system at all ! I would say that majority of the people would enjoy more of they system if other aspects of the reproduction is emphasised. One of the aspects could be spaciousness. Even Toole says generally people like added spaciousness. To achieve spaciousness, one do not have to sit in the sweet spot !
- Elias
The solutions is (again) to use a center speaker
If only my TV wouldn't sit there already.
I forget to mention: No mono (one speaker) will not be optimal for spaciousness, because lower IACC can be achieved with more speakers with decorrelated signals.
I believe for majority out of 6e9, stereo just provides more exciting sound experience than mono, and it's nice to have some ping-pong between the speakers too. For most out of 6e9, the reason to listen to a stereo in the first place it's not about the phantom imaging, or accuracy or whatever, but to listen to the music they like, for a feeling.
- Elias
The solutions is (again) to use a center speaker
To what extend center speaker can correct grossly off sweet spot listening positions ?
Ok, thanks for the explanation. How about using bigger drivers, like 6" or 8" ? Their inherent beaming properties would lessen the intereference and floor bounce. Did you try ? (for a moment forget about their potentially lousier high freq direct response).
- Elias
"Grossly" like in listening to a concert while still standing in front of the concert hall? There's always a thing called "the best seat". Besides wave field synthesis there is no technique that will prevent the breakdown of phantom imaging.
Have you ever tried a center speaker?
I made similar configurations using even smaller drivers, where the array
had less height and all drivers running fullrange ... they worked quite well
subjectively but that was many years ago.
My very personal definition of "fullrange driver" kind of excludes conventional
drivers >4 inches diameter. But feel free for experimenting, i think it is an
easy to test setup.
If you want to try such setups i really recommend you to use some small
and well behaved el cheapo drivers to get an impression by "rapid prototyping",
also trying frequency dependent power tapering. It is superior to the untapered
version IMO.
Last edited:
Not at my house. No problem turning the head or bobbing back and forth. But my speakers are very large. However, Geddes speakers are much smaller and they don't collapse the image if you turn your head or move a bit.
Stereo never works perfectly. It works better if the listener is in the sweet spot. It still adds a lot when many things are pretty wrong technically. I think the ear-brain mechanism does more analysis and adapting than we realize.
AR came out with a speaker many years ago (the LST2) that got rave reviews, apparently because it sounded real good anywhere in the room. Imaging may not have been the best there is, but as a tradeoff with horizontal dispersion, I think they designed it for most people rather than techno nerds like me who are obsessed with image localization clarity at the sweet spot. It has angled side panels that have additional mid and tweets aimed outward about 45 degrees to the sides. The center front has woof, mid and tweet (see pic below). (Thanks DB).
Attachments
Last edited:
If you have the image on your computer just click "Manage attachments" below the message composition section to upload up to three images as attachments. That's all I do...
I'm glad someone thinks a center speaker is a great idea, since I'm in the middle of building such a beast right now. I'm actually RE-building my center speaker, after finding that the first one sounded too confined (just an 8 inch and a ribbon tweeter sitting on top). When dialog or whatever would get steered to the center only, it didn't feel like it integrated in with the rest, which always has a spacious sound to it. My new center speaker is a bit like the AR LST2 in that it has angled front/side panels (only 10 degrees in my design) with a five inch driver on each angled panel, and the front center panel has two fives and a 2 inch dome. A five tweeter curved horizontal array will sit on top, handling 7kHZ and above (3/4 inch domes). I'm using five "super tweeters" instead of one or two, so any cancellations due to mixing with it's other tweeters, or with the tweets at the left or right speaker system, will more likely be filled in by the physically displaced other tweets putting out the same thing with their different delays to my ear. My goal with this center speaker is wide horizontal dispersion. As with all my projects, it's an experiment.The solutions is (again) to use a center speaker
However, it's my belief that although it will help imaging in the upper midrange, it may actually hurt imaging in the lower midrange, since it will put out yet another timing cue that has little to do with any embedded timing cues. But I think it will help the off-axis sense of continuity enough to be a plus on the bottom line.
Last edited:
Markus
I've always liked the idea of a center speaker, but here is my concern and I wondered if you knew the answer. Deriving a center channel by simply summing the left and right is wrong, the center channel should be uncorrelated with the left and right, at least above some frequency. But deriving a unique uncorrelated center signal from the LR signals is not trivial. How is it typically done and does anyone actually derive a unique center channel signal or is it just summed mono?
I've always liked the idea of a center speaker, but here is my concern and I wondered if you knew the answer. Deriving a center channel by simply summing the left and right is wrong, the center channel should be uncorrelated with the left and right, at least above some frequency. But deriving a unique uncorrelated center signal from the LR signals is not trivial. How is it typically done and does anyone actually derive a unique center channel signal or is it just summed mono?
Someone linked a paper lately that had a good summary. Not sure if it was in this thread.
Just summing L+R isn't correct but none of the available solutions are. The real solution is more channels.
Never heard it myself but this matrix should work pretty well. Klaus (KSTR) has some experience with matrix system(s).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.