I do not know Markus, if you have followed the earlier discussion on Cepstral Analysis.
If not here is a primer :
Reflecting on Echoes and the Cepstrum
If not here is a primer :
Reflecting on Echoes and the Cepstrum
Well, we may again have to define a term : early reflection. I am sure you mean reflections on the walls of the listening room including the floor and ceiling. What i try to minimise in the ZDL is "very early reflections" on the drive units and cabinet. Bill Waslo talkes about that in the paper. We discussed how that reflections influence the perception of what we called "disappearing loudspeaker" and that is what i try to build.
Well if we are back to where I was before and eqing diffraction/reflections from the cabinet and drivers. That is generally minimum phase and can be corrected 100% with EQ for a specific design point and the correction, while not perfect, will extend with reasonable improvement over a wider window that some would expect, as I showed on my plots. That does not mean we should not strive to eliminate such artifacts at the source.
How wide is that window? What is the time limit until the equalization becomes unstable for different listening positions?
Anything <1.5ms is of importance for localization but in my opinion this has nothing to do with speakers "dissapearing". Speakers dissapear when spaciousness is increased. Place the left and right surround speakers at 60° from the listening position and switch on PLII. The sound stage changes dramatically.
Agreed. The sub 1 millisecond regime is important for localization of the sound image. And if the baffle's effects on temporal and amplitude response are essentially symmetrical between the left and right speaker - impact of even a fairly wide baffle should be minimal. I think there are some common misconceptions here about baffle width, image localization, "localization" of the speaker itself, and directivity.
A great deal of effort has been expended on developing speakers with improved directivity to control early reflections using waveguides. Much of this improved directivity is a lot more easily obtained by simply using a larger baffle, maintaining sufficient toe-in, minimizing distance to side walls, and listening at fairly close distances (less than 10 feet). Room treatments can also be a major factor - nothing new there.
The problem with "low diffraction" designs is the variation in very early amplitude response as we go off axis where one speaker could be producing a noticeable amplitude boost and the other a deficit. This is particularly a concern at frequencies higher than 1.5khz where amplitude differences begin to contribute more to localization than phase differences.
As for locating the speakers in space, it's fairly widely acknowledged that later reflections have more impact than early. And as Joachim hinted earlier, listening distance has a huge impact. The attached plot positions a 1 meter measurement against a 4 meter measurement in the same room where the longer distance was measured fairly close to a side wall. Clearly, the ratio of direct to reflected energy is vastly different at 3ms on out. These are the cues that are telling us where the speakers are in the listening space. In contrast, the sub millisecond timing and amplitude differences are the ones telling us for example that the bass player is positioned a few feet to the right of the drummer. Apples and pomegranates.
Attachments
John, I don't follow this or I don't agree - not sure which is which. I don't see diffraction, as I understand it, in the impulse that you show. Diffraction, to me, would be a delayed impulse following the inital one and I don't see that. You also keep claiming that diffraction is MP and I have trouble with that as well. If I have an impulse followed by a second impulse and you correct this with a minimum phase filter, is your claim that the second impulse will disappear? And if so, could you actually do that on an impulse that I created?
This is completely wrong. The baffle size has little to no effect on the directivity, and none above the ka= 2.0 frequency of the source - the region where a waveguide is the most effective.
"Low diffraction" designs, at least mine, do not do this.
How much baffle size has to do with directivity is a function of wavelength and baffle width - which obviously varies. The original statement was never intended to suggest that baffle width alone could produce desired directivity. A combination of factors were mentioned that could reduce early sidewall reflections. If effective baffle width had nothing to do with directivity, then in theory, you should be able to build waveguides that do exactly what your waveguides do that are no larger in diameter than the driver itself. Width of the structure that transitions to free space is very much a factor in the end result - directivity being one particular aspect.
No one was mentioning your designs specifically. Your designs are not the only "constant directivity" designs in the world today. Frankly, like any other design, they have their problems as well as advantages and none of that was introduced into the discussion by me.
Not completely certain, but I do think that very early reflections have something to do with it. And my tendency would be to say that spaciousness only acts to mask the non-disaperance of many speakers.
Usually what happens when the "sweet spot isn't up against the wall or 25 feet away from the speakers with back or side walls being a lot closer. Anyone who thinks the speakers are going to disappear whilst walking around the room is totally delusional. AFAIC, being hyper concerned about constant directivity and minimal early reflections in a system that's plagued by inter aural crosstalk and phase issues is like blowing your Hummer's tires up to save gas. If you're not in or very close to the sweet spot, the rest is fairly irrelevant.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.