Indeed. What it says is that you are not listening in an anechoic chamber with your head in a vice...
If you attempt to EQ a high frequency diffraction effect it will only be right at one point in space, which means at other angles where the signal is heading towards something like a wall the response will be wrong.
As I said, you can't EQ it because it would only be right at one point in space. This means it will probably be wrong for most of the reflected sound in the room.
I have no quarrel with that.
I remember discussing in another thread a long time ago that anecdotally it seemed to me that a speaker with minimal diffraction at high frequencies (say >2Khz) tended to "disappear" and not be an obvious source of sound when playing a stereo recording with a depth image in it - the sound could all sound like it was coming from behind the wall the speakers were sitting in front of and not be identifiable as coming from the actual speakers locations.
But a speaker with a lot of baffle diffraction at high frequencies could be identified as a source of sound on the same recordings thus flattening the apparent image depth to be in line with the depth of the speakers from the listener.
I theorised that the brain might be able to distinguish the angular offset between the left and right diffraction edges of the cabinet and thus estimate a width of the sound source and from that conclude that the source was "near", while a true point source with no diffraction and no "width" would seem infinitely far away and thus localise to where the apparent sound source is in the recording itself.
Agreed again.
Diffraction control at higher frequencies is critical. This is why I wouldn't give two pennies for a speaker with a flat face dome tweeter mounted on a flat panel with normal cabinet edges - the diffraction effects are horrific. (just look at the polar response of the peerless tweeter I posted earlier, which is the fault of the baffle diffraction, not the tweeter per se)
To do it right, you either need to have a baffleless design - like the tweeter and midrage on something like a B&W Nautlius 802, where they are both basically mounted in a spherical teardrop, or if you are sticking to a traditional flat baffle, a waveguide.
For those of us whose construction skills are limited to fairly ordinary, boring boxes, the waveguide is the obvious way to go, and has the added bonus of correcting the driver acoustic centre offsets if the drivers are chosen well - which lets you get good phase tracking without needing to use unequal slopes or all pass filters.
I consider the fact that my ribbon tweeter has a waveguide with directivity control one of the key things that makes the speaker as a whole work so well despite the boring traditional shoebox design. Despite appearances the speaker as a whole has very little high frequency diffraction.
We are on the same page regarding diffraction, I think you have just extrapolated a bit too far from my comment that diffraction is minimum phase and can be EQ'ed for a single point in space - yes it can, but that doesn't give a desirable end result in a room.
To me, when you say complexity, I think "speed" which is inherent in the frequency response. If the frequency response extends far enough and flat enough for there to be no discernible phase shift in the audio band.... job done
I think a sine sweep is a really easy signal in the sense that the change is gradual. Even at the high frequencies the driver is oscillating at a fairly regular rate. A more dynamic signal test might reveal more, and make things like horns lmeasure better.
Don't get me wrong, linear distortion, particularly at the listening position, is still king.
I know the feeling. I don't bother too much with high end usb devices, amps, and soundcards.
Yes, and it says to me that reflections and diffraction were the issues I had been trying to identify all along.What it says is that you are not listening in an anechoic chamber with your head in a vice...
Yes, I can put my head in the mouth of one waveguide and the sound still comes from between the speakers. Each speaker itself is silent, regardless of the recorded material.
The way I have found to hear diffraction is to remove it. It can sound like a silhouette of the object.
A flat baffle is merely a point in a continuum between these two (notwithstanding the geometric mis-match of the room WRT what walls are illuminated). The former tends to offer more reflections while the waveguide requires greater amelioration of diffraction than the flat baffle due to the difference in radiation space when it ends.
I wouldn't say that. Once sound has left the driver it can be beyond equalisation.
I was impressed by the objective approach of you until I read this! Seriously, you could pick out 1db difference between 10-20kHz????? Are you batman
I totally agree.
Another explanation: It took you that long to convince yourself of what you wanted to hear. It can be a difficult thing to convince oneself of something false.
I think that the question was answered very early on an then it went in other directions based on that answer.
PS. You seem particularly negative tonight! Bad day?
No, not so, if I know the transient response of a linear system then I know how it will respond to any signal. And yes, if you didn't know that then do you need to study up.
Actually it is well know, check Brian Moore, who describes how group delay, which diffraction creates, is more audible at higher SPL. This research is decades old. My own research is a decade old now.
I don't know if this is him, looks interesting nonetheless
Professor Brian C. J. Moore, FMedSci, FRS — Department of Psychology
Professor Brian C. J. Moore, FMedSci, FRS — Department of Psychology
I had in mind the observation of Rod Elliott:
"The majority of studies (on websites or elsewhere) that have shown that Doppler distortion does exist, have used a spectrum analyser (or FFT - Fast Fourier Transform) to show the sidebands so generated. The spectrum analyser is completely the wrong tool to use for this, as it only shows frequency information - the effects we want to capture are in the time domain, and are best seen with an oscilloscope."
I don't enough to know if Rod Elliott's criticism of FFT tools apropos Doppler is correct here or whether your dismissal is correct.
But I do know that using the wrong tool results in the wrong results. Easy to mistake the "birds's eye view" from a textbook for the "worm's eye view" of the listener. Even wonderful software is still just a "black box" to many users, whatever the abstract truth of FFT.
B.
Last edited:
If course it should be audible at higher levels as with some other effects as well. The strength would be proportional to the source sound so that as the SPL increases, the effects become more audible.
er uh, I'm not sure that a driver is a linear system. I believe that what I am saying is the basis for the Klippel tests. IOW some signals are more distortion inducing than others.
For example if a low frequency pushes a driver to its xmax, then that will cause distortion in other frequencies because the magnetic field changes. So then a subwoofer can make a two way have less distortion by removing the bass frequencies.
I'm pretty sure that's what I am hearing. The less a single driver has to work the more "effortless" it sounds. Horns have to move less, and are very light so sound quicker. Of course I could be wrong, but it does seem like a sine sweep isn't all that a driver has to do.
Last edited: