Most likely arrived at through experimentation. What helps and stays under a $20 price point etc. Ted did have an MLS measurement system but we can't ask about the design process because sadly he passed away two years ago.
Indeed. I've more than once observed that speakers with emphasis in the high end treble (> ~8Khz) of a specific contour can shift the apparent image source well above the highest driver in the speakers, by as much as 1-2 feet...so that the phantom centre channel image appears to be hovering between and significantly above the top of both speakers.
If the errors in treble response just so happen to conform to a valid HRTF that corresponds to an elevated angle of incidence I can't see why the brain wouldn't interpret it that way, since messing with the HRTF is the basis of many 3D audio technologies which (as well as other factors) convolve the frequency response to fool the brain into interpreting the sound as coming from somewhere other than the speaker.
Of course everyone's HRTF is a bit different, so one person might hear an elevated image centre and someone else may not or to a different extent. Many frequency response errors don't correspond to any valid HRTF direction, and in this case I suspect the brain may just become "confused" due to conflicting cues, and perceive an ill-defined image that doesn't localize well.
Certainly it's been my experience that the frequency response from about 2Khz up is very critical to imaging - some kinds of errors will shift the apparent image source location but still provide a strong (if inaccurate) image, while others will result in an ill-defined image, so getting the frequency response just right >=2Khz is just as important as other things like diffraction when seeking good imaging.
oh yeah, very valuable, it tells the poor audiophile how would the speaker sound in an anechoic chamber (= poor audiophile's dream listening room) 😉
certainly most valuable thing, just next to reading Blauert... 🙄
I wonder will You guys (You know who You are) ever understand that there is no such thing as direct sound at all, it is just anechoic/gated measurement artifact
the sound we can here IS a highly interpreted picture reconstructed by the brain from all the sound data it receives - the direct and reflected - up to around 50<100 ms (according to various researchers) from the first wavefront arrival combined
Last edited:
I have no doubt that you're hearing something, I just don't think you've taken sufficient steps to isolate with any degree of certainty either the frequency range where you're noticing the effect, or the cause of the effect. (By systematically eliminating various possible causes)
I think you're right there, the largest peak before the first notch is probably what is being perceived as a tone, much like sweeping a narrow peaking bandpass filter over a noise signal gives a sense of pitch.
As you point out yourself, comb filtering causes a whole series of peaks and dips, and whilst we probably perceive the frequency of the first peak we can't be certain of that. Pitch perception can be fooled by various means, for example a missing fundamental on bass notes with strong 2nd and 3rd harmonics can sometimes still be perceived as the fundamental note.
If you're sure the effect is above 1Khz and want to prove it, why not try applying a steep high pass filter (at least 24dB/oct) at 1Khz to your pink noise and see whether you still notice anything ? We've agreed that there will be an effect in the lower midrange, so this would help separate that out. If you can no longer hear it you'll know it's an artefact of our perception of the lower frequency comb filtering, but if it really is above 1Khz it should become even more apparent.
You did say that you were walking back and forth towards the speakers when you noticed it though. Most floor standing speakers would have a design axis whose height was between about 0.7m and 1m, while your ear height while walking would be somewhere between 1.6 and 1.8 metres (just guessing here 🙂 ) and if you're walking between 1-3 metres that's a very steep angular range, especially at the near end, and surely well outside the "usable" listening angle as far as crossover summing is concerned, even if the drivers themselves have sufficient dispersion. It might sound ok for general listening but there will be response ripples noticeable on pink noise at such a steep angle on all but the tiniest bookshelf speaker.
Hence my suggestion to try the same test with your ears at design axis height to eliminate this as a possible variable.
In this scenario were you listening to a single mono speaker or a stereo pair ?
Yes it is inconsistent with that notion, which is why I'm still skeptical when I believe insufficient steps have been taken to narrow down the cause. There are a number of possible factors involved that need to be looked at individually.
I've done tests of my own which don't give the same results as you though, some of them in the past, some of them now in response to this thread, just to try and see what's going on and whether practice and theory are conflicting or not.
I've taken steps to isolate the various causes, including:
1) Only testing one speaker at a time, to avoid comb filtering between left and right speakers. (very important)
2) Keeping my ears at the same height as the design axis of the speaker and moving directly towards and away from it, to eliminate variations due to driver off axis response, inter driver time/phase shifts in the crossover overlap region, and to minimize changes in baffle diffraction pattern. (Rather than raising the speakers on stands I opted to just crawl 😛 )
3) Applying a steep high pass filter at various frequencies to determine the highest frequency where I can notice comb filtering with movement. (As well as trying it unfiltered)
The unequivocal result from this was that with no high pass filter I could hear a definite comb filter related shift in pitch in the lower midrange (most obviously from about 1 metre to 2 metres) but with the high pass at 2Khz I could not hear any comb filtering or periodic effect over the same 1-2 metre distance range, and in fact there was almost no shift in the character of the treble at all.
Although I've already suggested some of the steps above to narrow down the specific cause, so far you've not said whether you've tried any of these things. With too many unknown variables I don't see how you can draw definite conclusions about the cause of what you're hearing.
Last edited:
The following paper seems to do a pretty thorough analysis on the perception of pitch in "Iterated rippled noise" (of which comb filtering is an example):
http://apl.sphs.indiana.edu/papers/Shofner%202008.pdf
Although much of the article is over my head and I've only skim read it, the summary in part C of page 12 is of particular interest.
Last edited:
on precedence effects and timbre:
http://murphylibrary.uwlax.edu/digital/journals/JASA/JASA1999/pdfs/vol_106/iss_5/2933_1.pdf
results and discussion at pages 10-12
http://murphylibrary.uwlax.edu/digital/journals/JASA/JASA1999/pdfs/vol_106/iss_5/2933_1.pdf
results and discussion at pages 10-12
I don't follow you here. What is binaural decolorization?
20ms means a surface 10 feet away. This isn't all that difficult and certainly double bounces can give good lateral energy in this time region. Are you saying an echogram would have nothing around 20 ms?
http://www.kenkantor.com/publications/magic_speaker/magic_speaker.pdf
Read Kates and Salmi.
I'm not talking about listening to music in an anechoic chamber. None of my threads have advocated that. Since this was a topic on omni speakers and we've converted it to a discussion of room effects, how about this as a stated goal: The loudspeaker and room should be designed as a pair with the following characteristics: Early reflections in the time region to 20 ms should be reduced as much as possible. Absorption, diffusion or speaker directivity should be used to reduce early reflected energy. Reflections in line with the speaker or in a vertical plane including the speaker and listener should be especially reduced. Later reflections (after 20ms) and especially later lateral reflections giving low IAACC should be encouraged.
David S.
true, with one exception of omnidirectional speakers (omnidirectional as opposed to polydirectional), I don't mean ideal 720° spherical omni but don't ask what I mean because in this case I will come and kill You because I have posted this hundred times already, quite ad nauseam
yes, this is true
true but imprecise - after 20 ms but before echoes start to appear, which is quite soon🙂
Listen to white noise from a single speaker preferably near a rigid wall and walk around the speaker. You should hear a changing pitch. Now plug one of your ears. Is the perceived pitch stronger or weaker?
Of course it is not difficult to build a large room where early reflections reach the listening position after 20ms but I was talking about 99% of all existing rooms. I'm too lazy to do the math but this is what Devantier found for a "typical" room - all the dots are under 10ms:
An externally hosted image should be here but it was not working when we last tested it.
This is a recent survey of typical US home theater room sizes (AES Convention Paper 8310 by Tomlinson Holman and Ryan Green):
An externally hosted image should be here but it was not working when we last tested it.
Looked at this before but why only emphasize lateral reflections and not contralateral and ceiling reflections?
I do have my speakers setup like this and the sound is dynamic, transparent with pinpoint imaging but it seriously lacks spaciousness. It probably would be even more dynamic, transparent and pinpoint within an anechoic chamber but that's not the point. The point is that the this approach isn't capable of creating ASW and LEV unless additional speakers are added. Don't get me wrong, multiple (more than 2) high directivity speakers (and multichannel recordings) might be the right approach for highest quality sound reproduction but I'm not yet done exploring different approaches as they have shown to create a strong early spatial impression. I'd like to know how the soundfield has to look like in order to create such an impression and how it compares to the high directivity approach.
This is an interesting point and it might well be true that spatial impression due to early reflections is the result of HRTF-like distortions. On the other hand this would be a very unfortunate truth because such a speaker-room-model would only work for a single person due to individual HRTFs.
http://www.google.ca/url?q=http://c...UQFjAA&usg=AFQjCNG0I9rA1zXW3j6h7y3ciAKuNgTXdQ
Nice paper on the subject by Griesinger (Lexicon).
"We have used this simulator many times to experiment with both reflections and reverberation....Lateral reflections with delays greater than about 10ms produce a sense of surround which is distinct from the central image. The threshold for the surround impression is about 20dB below the direct level, and when multiple reflections are used it is the total energy which determines both threshold and loudness. Thus, when you combine several reflections each of which is below audibility the result may be audible if the total energy is greater than -20dB. With noise there is very little change in the spatical properties of the sound as the reflection delay is increased beyond 10ms. The impression is always of a sharp central source in the presence of a surround field. The impression is constant until a single reflection is within 3dB of the direct energy, at which point the apparent width increases and the position of the direct sound shifts toward the reflection. ASW is a poor description of the effects of reflected energy on a noise signal above 300 Hz. The impression is of a sharp sound source in the presence of a surround."
And Later: " we noticed that there was a dramatic dhange in the surround impression when one or more of the delays exceeded 50ms. As the delay time is increased an abrupt change in the impression occurs at about 50ms. Single reflectins with a greater delay begin to become seperable from the direct sound." (Become echos)
So multiple lateral reflections adding to a level of -3 to -20 db and in the time interval from 10 to 50ms are best for adding spaciousness. A Zenith "circle of sound" system (or even a Bose 901) won't exactly give that.
David S.
The shape of the pinnea is remarkably similar between different individuals. Mainly the size differs, and I don't want to advocate that size doesn't matter, but individual HRTFs should by and large be identical.
vac
But unfortunately they are not. Have you ever heard through somebody else's ears, i.e. binaural recordings made with in-ear mics? I do own a Headzone Pro but it doesn't work because of it's non-individual HRTFs. However the Smyth Realiser works perfectly because it measures and uses my own HRTFs. So maybe it's just my pinna that is different from the norm (although it looks pretty normal)?
I make binaural recordings from both my real dummy head and a homemade dummy head. My real head wins easily. The dummy head still works to a degree, it's just less precise. I imagine if the mic/earbuds were better matched my real head recordings would be perfect. I've also heard engineered 3D recordings--they have about the same 3D effect as my dummy head only the original sounds are of higher quality and there's less noise.
Dan
Dan
yes, but not only HRTF but ATF - ATF for anatomic tranfser function
individual not necessarily means entirely different from all others - after all we are all humans, aren't we? 😉
so not that dissimilar - just a bit different as DBMandrake rightly put it
indentical - certainly not, basically similar - definitely yes therefore:
yes, so still there is some hope 🙂
yes, exactly my point, thank You for the source and numbers 🙂
Bose or Zenith were not omnidirectional, only polydirectional in their respective way
and so they sucked in their respective way 😉
c'mon Markus! something must be wrong with Your implementation of the setup 😉
according to Dr Geddes You just have to have spaciousness in such setup 😀
His setup is most certainly capable of everything anz audiophile can ever imagine, amen
please do not disrespectfully question this, God forbid!
amen amen
Just found a picture of how the current Audiorama 9000 speaker by Grundig looks like without its round housing:
An externally hosted image should be here but it was not working when we last tested it.
- Status
- Not open for further replies.