At the end we have to consider the ear signals, everything going in is summed. Allways signal and its delayed version are summed together the total signal have different phase than the original signal. This is what happends when a reflection reaches at the listening position. Think two sinusoids summed together at different phases, it is the simplest example.
- Elias
Sorry Elias,
this is plain old summing location. It has been shown that this doesn't work in each and every listening situation. If the localisation cues become too vague, summing will not happen in that stage.
Rudolf
Exactly, the ITD will be erroneous due to reflections. That's why we don't want room reflections in the ITD freq range, below 1kHz.
- Elias
Having some downtime at work today, I did a quick read of Smyth's thesis. Very interesting, but I do have one concern.
Smyth divided his frequency bands oddly, based on transition frequency points from various other researchers, rather than by equal octaves. In music, at least, hearing tends to divide the frequency spectrum by octaves. For instance, his lowest band "A" (20 to 800Hz) is about 5.5 octaves wide while his highest band (12kHz to 20kHz) is less than one octave!
Changes were made by moving blocks of the spectrum in whole over to the offset speaker and recording the listeners reports on how the sound image appeared to move. I think, all other things equal, I'd hear a change from relocating 5 octaves over more than from relocating just one. Particularly if the one is about where most people's hearing normally goes weak (and is probably not given much attention because of that).
Smyth did some analysis to show that in his white noise tests there was more energy and more calculated loudness in his chosen higher band blocks, but I don't know that is relevant. It is a matter of detecting changes that is being investigated. The loudness of the bands were the same for both moved and not-moved cases, having bands seem louder doesn't imply (to me at least) that they should be more detectable as having moved. But having the part moved span a lot of octaves might swing the bias.
?
Last edited:
Left speaker from listening position microphone horizontal:
www.dipolplus.de/assets2/Hoerp_links_waag.wav
Left speaker from listening position microphone vertical:
www.dipolplus.de/assets2/Hoerpl_links_senk.wav
Have fun
Rudolf
Thanks. Did you also measure the right speaker?
Is the response corrected with Acourate?
It's possible, but under our context - not likely. (..there are some *extremely* directional techniques available, so directional that any reflection off-of your body and on to any room surface would be so attenuated that it would be inaudible.)
I think a less extreme version however could be achieved.
That's not Smyth.. Stephen Smyth (..of Smyth Realizer) was mentioned from a conference paper he gave in relation to DTS surround that prompted the thesis paper by *Hartman*.
Near field comes to mind.
Anyway, removing the room equals making it anechoic. Not practical. But the problem really is distortion of the interchannel stereo signal as it is filtered by room interaction. If we can undo that filter, improvements can be made.
Last edited:
Thanks for posting that. I'll try to catch up with the thread and read the paper.
It seems like a breath of fresh air to me - because for years people have told me that it's the high frequencies and the tweeters that aid location clues. I just don't hear it that way. Tweeters add texture and tone for me, not space or location. I hear space and location clues down low, mostly in the octave around 400Hz.
Since that's where I hear the location clues I have long been puzzled why tweeters don't do much for me, except change tone. The study looks worthwhile to pursue.
Pano,
A simple clicker test somewhat answers about the higher frequencies being used for localization. Have someone wall around a room with a clicker and I don't think for a second you won't be able to localize that sound no matter how reflective the room may be. Now move a band pass subwoofer that has a fairly low frequency output and I bet that is much harder to localize in the same room. But I would say that our hearing is fairly good at localizing anything above 200hz without much trouble in that same reflective room.
A simple clicker test somewhat answers about the higher frequencies being used for localization. Have someone wall around a room with a clicker and I don't think for a second you won't be able to localize that sound no matter how reflective the room may be. Now move a band pass subwoofer that has a fairly low frequency output and I bet that is much harder to localize in the same room. But I would say that our hearing is fairly good at localizing anything above 200hz without much trouble in that same reflective room.
All I can say is there was a certain cognitive ease when listening to the VOTT. It made sense. The illusion that there is a real instrument in space that I can point to and say, "there it is" was unshakable. And while I have heard great imaging from econowaves crossed at 1 kHz, that cognitive ease is not there.
Not yet. I'm applying small changes to the frequency response or move the speakers from time to time. So measurements were as of today. Can do the right side easily tomorrow.
Tells me that you haven't seen the response curves yet
No, Acourate is for boys
. I still try to understand what's happening in my room and how to correct flaws at their origin before I let the machine bend my space.Rudolf
They do image well, but not as well as the VOTT, in my opinion. To what's that due, I don't know. Probably a combination of factors. The VOTT has a sense of scale that smaller speakers lack - it should! - considering its size. The comparison is basically unfair, taking into account the size and complexity of each.