Well that evidence would work as a positive indicator that the Blumlein patent works as mathematically derived.
The smoothness of the color gradient is a direct function of the coarseness of the mesh. I'll try and post one that is very fine. But as the mesh size goes up the real-time update degrades.
Isn't the model used in the Kates paper simple enough? Here's the ideal directivity for the 45° toe-in case (49° 3dB beamwidth):
Yes, that was kind of weird. When I overlayed the speaker directivity plots and the ideal curve there was "some" discrepancy
Attachments
Did the following test: Played correlated pink noise with the same level in L and R. There's a strong center phantom image when listening at the sweetspot.
I moved 50cm to the right. The phantom image snaps to the right speaker.
Now I did raise the level of the left speaker in order to move the phantom image back to the center.
This did not work. At one point the left speaker becomes localizable as a distinct additional source. Localization is somewhat diffuse with noise coming from in between both speakers.
At about 10dB difference noise is fully localized in the left speaker.
Anybody else could try this and report his findings?
I moved 50cm to the right. The phantom image snaps to the right speaker.
Now I did raise the level of the left speaker in order to move the phantom image back to the center.
This did not work. At one point the left speaker becomes localizable as a distinct additional source. Localization is somewhat diffuse with noise coming from in between both speakers.
At about 10dB difference noise is fully localized in the left speaker.
Anybody else could try this and report his findings?
I've read the Kates paper. The Blauert book post dates Kates. Why would I not use the experts curves unless there is more recent evidence that supersedes it? At any rate, I thought that I was clear that I am not interested in deriving the "optimal" directivity, that's easy to do, but pointless if one cannot achieve it. I want to look at how real speakers in real rooms work.
Where your speakers toed in or out?
I'm with Pooge here. The tradeoff works for me and most people who have auditioned my system. It's a common comment.
So how do they work? The CD speakers I know don't provide the necessary level difference required for trading. All other speakers are worse because they present a frequency-dependend trading effect.
A center speaker is probably a more reliable solution. If I were a speaker manufacturer I would immediately agree with that assessment
So how do they work? The CD speakers I know don't provide the necessary level difference required for trading. All other speakers are worse because they present a frequency-dependend trading effect.
A center speaker is probably a more reliable solution. If I were a speaker manufacturer I would immediately agree with that assessment
Well actually my speakers have more level difference than the theory (i.e. Kates) would suggest. But your second point is the key to my motivation here. How do the different design approaches stack up, in the real world, knowing full well that none of them is perfect. And more importantly is the a better way to set up a given pair of speakers. For example, from my studies I have found that my speakers would image better if they were closer together. I may try that - although changes in my room are not easy. But if I cannot move them closer together, then it appears that a greater toe-in is preferred.
Once I get this done I'll post the software and people can play with it - those people for whom it works at least. I still have no idea how to find out why it doesn't work in those occasions where it fails. Every one appears to be different. Developing web apps is not my occupation so there is a very limited amount of effort that I can expend in this regard.
That, of course, is going to cause a total collapse of the image once one moves off center. I am sure that you realize that the whole effect is quite dependent on how close together the two speakers are. Did you try changing that?
That's not what my test is about. I didn't move, the signal of the left speaker is varied. It shows if trading is working at all. It doesn't.
What happens when you play correlated pink noise through L/R and move 50cm to the left or the right? Do you still get a well defined, stable center image?
What are you trading if you stay in the same spot and vary the loudness of one speaker? You have to vary distances from the speakers in sync with level to get the tradeoff.
I'd get it the same place most people do - either measure it or just make it up.
I was testing if trading works at all when the listener isn't located on the center baseline of the stereo tringle. If the phantom image would keep its "locatedness" and simply move by varying the level of one speaker, I would have to conclude that trading works in principal. Unfortunately it doesn't work if the listener doesn't sit in the middle of a stereo triangle.
Did you try it?
Markus
I simply do not understand your point of view. What I am interested in is how to make a system such that the "image" does not collapse to the near speaker when one moves laterally. Mine does not.
When I heard a pair of Wilson speakers one time they had very good imaging, but it completely collapse when one moved laterally. If what you are saying were true then this situation could not have happened - both systems should collapse in the same way, there is no tradeoff effect, so they would have to collapse in the same way. Not my experience.
I simply do not understand your point of view. What I am interested in is how to make a system such that the "image" does not collapse to the near speaker when one moves laterally. Mine does not.
When I heard a pair of Wilson speakers one time they had very good imaging, but it completely collapse when one moved laterally. If what you are saying were true then this situation could not have happened - both systems should collapse in the same way, there is no tradeoff effect, so they would have to collapse in the same way. Not my experience.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.