Colin said:
This idea has some merit and is the one preferred by Floyd Toole. To me it seems impractical and not necessary. What I have done is a little different. Put the speakers out away from the corners and hang some draps across the corners in front of the speakers. This hides the speakers and reduces the appearance issues, and then I put a lot of absorbing material to fill the corners (behind and arround the speakers but all behind the drape) to help damp out the LF modes. We both use a toe in of about 45°. We both taper off the high end a bit, he more than I. With flat Summas bhind a drape I get a nice gradual rolloff at the high end of just under -3 dB/oct.
I would use more subs (none are evident).
The room response seems questionable since there don't appear to be any subs, the room is very reverberant, but no LF modes are apparent. That goes against my experince.
el`Ol said:
Hello Oliver,
by various occasions I was able to hear the Fraunhofer wave field synthesis loudspeaker rows around the listener. It is a huge step in the direction of realistic sound reproduction because it isn’t reliant upon the uncertain phantom source perception, like the conventionally procedures. But unfortunately the procedure isn’t able to bearing down the most audible disadvantage of all known loudspeaker reproduction procedures; the limitation onto the horizontal plain of the listener remain very disturbing.
graaf said:
You shouldn’t overrate the impact of the effort. No procedure in history was inhibited by that reason as far as the advantage for the user was apparent. Some multimillionaires would be purchase such gadget only because it would be too expensive for a normal millionaire . Unfortunately i am not among.
graaf said:
…o.k. I want not to enter the tread by my idea. Possibly we can discus the topic in this thread: http://www.diyaudio.com/forums/showthread.php?s=&threadid=119931
But the topic is loudspeaker and room as a system. The most of the errors in the frequency domain, that should concern the topic surely, caused by wrong superposition of reflected wave fronts in the playback room. Equalizing such comb filter effects in the frequency domain is the wrong way. Those errors are time related. It must become correct in time domain congruously as described on the website, not the frequency domain!
graaf said:
fortunately 1296 speakers with special electronics are not neeeded for that - in fact stategically positioned just two fullrange drivers without special electronics can do the job
best,
graaf
No, it doesn’t. We are far away from the goal of realistic reproduction by all phantom source based procedures. Only reason would be Ambisonics or wave field synthesis.
syntheticwave said:
by saying "do the job" I meant "do the job" "of being much better than what we have now"
I certainly didn't mean "do the job" "of wavefield synthesis"
in that sense surely it doesn't
but in a sense "much better" it does
and what about ambiophonics? "transaural stereo"?
http://www.ambiophonics.org/
what about binaural approach?
in words of Floyd Toole:
http://www.sonicdesign.se/tooleinw.htm
what do You think?
ambiophonics or binaural playack are perfectly technically feasible. And would be much more commercially feasible. Well, this doesn't change anything in reality. From the perspective of the industry these are and most probably will remain only curiosities
And "1296 speaker device" is even something more extreme, a real monstrosity
best,
graaf
syntheticwave said:
well, psychoacoustically the loudspeakers are perceived in stereo as early reflections of a sound whose direct version we missed
what is wrong with that? seems to me that nothing per se
what is fundamentally wrong is rather the positioning of these early reflections in a room
it is having these early reflections in a middle of the room what is higly unnatural and leading to lack of realism
try damping of the wall in front of the listener and Beveridge positioning where the loudspeakers are acoustically integrated into the side walls
what do You think?
best,
graaf
Hi Graaf,
Very interesting topic!
Jean-Michel Le Cléac'h wrote a document about phase distortion where he gives some numbers about the ability of the human been to localise point source in an horizontal field (the ability to localise in a vertical field is much less accurate). Between 100Hz and 10kHz, the human accuracy is between 1º and 3º, which is very precise! You can find the figures on page 23 of that document: Distortion de phase.pdf (7Mo). But he also says that parameters such as timing and intensity are not sufficient to pin point accurately a source. He doesn't say what else is needed to achieve the accuracy quoted earlier (1 to 3 degrees between 100Hz and 10kHz)...
He also says that the mechanisms of localisation are phase difference for the lower frequencies and intensity difference for the upper frequencies. Phase difference localisation is only possible up to 1500Hz due to the size of the head. See Stevens & Newman for the complete info. 3kHz tone are somewhat harder to localise. Above that, the intensity differences take over.
The room acoustic has always been one of my favorite subject. I myself consider the room as the weakest link in too many hifi set ups... Therefore your type of speaker combined with the appropriate positioning is quite interesting! Nevertheless, something catches my intention. Some of your requirements are high directivity for the high frequencies and early reflections coming after 5ms, 10 ms being much better. So far, so good! Then you place the driver firing upward. Since the later is directive in the highs you will have very little direct sound and mainly delayed reflected sound...
I don't get how this can lead to "absolutely more realistic than the standard stereo - occupying a real space, 3D, very palpable images" as you said yourself. I'm not questioning your feelings, I just try to understand the mechanism(s) behind!
I will try your set up myself and read the Carlsson white paper you linked to. This might help me understand better.
Regards,
Etienne
Very interesting topic!
graaf said:
well, maybe it all boils down to a question how "pin point" is "pin point"?
to make myself clear - I don't know the experience of other contributors to this thread but loudspeaker and set ups I write about here give me precise imaging, not uncomfortably "diffuse" nor "vague" at all
not at all "un-pinpoint" in any negative sense
and absolutely more realistic the standard stereo - occupying a real space, 3D, very palpable images
best,
graaf
Jean-Michel Le Cléac'h wrote a document about phase distortion where he gives some numbers about the ability of the human been to localise point source in an horizontal field (the ability to localise in a vertical field is much less accurate). Between 100Hz and 10kHz, the human accuracy is between 1º and 3º, which is very precise! You can find the figures on page 23 of that document: Distortion de phase.pdf (7Mo). But he also says that parameters such as timing and intensity are not sufficient to pin point accurately a source. He doesn't say what else is needed to achieve the accuracy quoted earlier (1 to 3 degrees between 100Hz and 10kHz)...
He also says that the mechanisms of localisation are phase difference for the lower frequencies and intensity difference for the upper frequencies. Phase difference localisation is only possible up to 1500Hz due to the size of the head. See Stevens & Newman for the complete info. 3kHz tone are somewhat harder to localise. Above that, the intensity differences take over.
The room acoustic has always been one of my favorite subject. I myself consider the room as the weakest link in too many hifi set ups... Therefore your type of speaker combined with the appropriate positioning is quite interesting! Nevertheless, something catches my intention. Some of your requirements are high directivity for the high frequencies and early reflections coming after 5ms, 10 ms being much better. So far, so good!
Then you place the driver firing upward. Since the later is directive in the highs you will have very little direct sound and mainly delayed reflected sound...I don't get how this can lead to "absolutely more realistic than the standard stereo - occupying a real space, 3D, very palpable images" as you said yourself. I'm not questioning your feelings, I just try to understand the mechanism(s) behind!
I will try your set up myself and read the Carlsson white paper you linked to. This might help me understand better.
Regards,
Etienne
Thank You!Etienne88 said:
probably what He refers to is head related transfer functions (HRTF) and especially the pinnae functionsEtienne88 said:
"high directivity for the high frequencies" and omnidirectional dispersion pattern are needed to make the individual loudspeaker in a two speaker stereo array harder to be distinguished from the side wall and harder to be localized as a distinct sound source and a real (and separate from the room acoustics) cause of virtual sound imagesEtienne88 said:
the aim is to acoustically integrate the speakers into the side walls just to make the room acoustics to be a carrier of left and right signals
by making the loudspeakers very short and positioning them next to the side walls I want to emulate (as close as is possible with dynamic cone driver) the Beveridge line sources
the additional advantage of high directivity in such a positioning can be less amount of detrimental high frequency "very early reflections" off the adjacent side wall
yes, in such a set up what we perceive in the highs is mainly sound vertically reflected off the ceiling arriving at the ears with about 10 ms of delay
but this is not a problem because the hearing mechanism is processing what the ears receive in "stages" or "phases"
the first stage is sound source localization and it lasts <1 ms counting from first wave front arrival at the ear, the second is sound source size assessment and it lasts for the next 1<30 ms, the third is sound source identification which begins with pitch recognition occurring between 10<30 ms
thus all sound source localization is done completely before the delayed reflected sound reaches the ears, no problem at all
moreover, in case of sound localization in reverberant space like a room hearing relies mostly on time difference mechanism - the law of first wave front and the precedent effect (Haas effect) and additionally on phase difference mechanism and in that regard most important are the spatial cues around 1 kHz at which my speakers are not yet that much directional
on the other hand the spatial cues encoded in a stereo recording are in most cases of the intensity type or at least they should be according to the original Blumlein stereo patent and a coincident stereo microphone set up invented by him for that application
this is simply how "stereo" should work
unfortunately this is often ignored diminishing the standard stereo ability of realistic spatial reproduction even more
Etienne88 said:
I am also nor completely sure about that
I am only hypothesizingand as I have already pointed to above in the thread the main problem with realistic spatial presentation of sound in audio reproduction at home is with the fact that we are listening to the speakers perceiving them as distinct sound sources, we are hearing the speakers – something that has no counterpart in nature
in spite we should listen to what the speaker do while staying unaware that this is done by the speakers
ambiophonics is trying to achieve that by the means of stereo crosstalk cancelling and making the listening room a little anechoic chamber
in that way the speaker themselves are made "invisible" to the hearing
but I believe that this can be done as well by integrating the speakers into the side walls instead of projecting the sound in the direction of the listener
in that way we can perceive that sound just occurs in the room like it would be if the room was attached through the wall in front of the listener to an original recording space (see image attached)
Etienne88 said:
I recommend to You the Beveridge white paper in the first place
http://www.bevaudio.com/White_Paper.htm
my approach is a sort of combination of the two ideas - of Carlsson and of Beveridge, and more precisely - while trying to emulate Beveridge loudspeakers’ radiation pattern I arrive at something resembling Carlsson speakers
it just happened that way - Carlsson was not my direct inspiration
best,
graaf
Attachments
graaf said:
Hello Graaf!
This was the setup (just destroyed to make place for an other experiment):
An externally hosted image should be here but it was not working when we last tested it.
Driver is the Ciare HX201.
The next experiment in this direction will be a compression tweeter with narrow directivity, directed towards the ceiling.
graaf said:
Graaf
Your numbers here are not correct. The first 10 ms is the critical time for localization then out to 20 ms coloration (Tembre) is affected. Beyond 20 ms there is no effect at all on localization, but coloration can still be a factor if the loudspeakers don't have matching power and direct response curves. If they are CD then beyond 20 ms reflections are only a good thing - they create spaciousness - especially if they are lateral.
The rest of your commenst are too long to comment on at the moment. This is all a very complex subject.
graaf said:by saying "do the job" I meant "do the job" "of being much better than what we have now"
I certainly didn't mean "do the job" "of wavefield synthesis" in that sense surely it doesn't but in a sense "much better" it does
and what about ambiophonics? "transaural stereo"?
http://www.ambiophonics.org/
Unfortunately I haven’t the occasion to listen such system. But I know well the website and can imagine the sonic impression: It should be the only realized loudspeaker approach until, which reach the stage to fault our brain for listen a genuine event.
Only disadvantage is the small working area. By my proposal that problem would be solved for the price of higher effort. The main ideas are very similar.
graaf said:
…. ambiophonics or binaural playack are perfectly technically feasible. And would be much more commercially feasible. Well, this doesn't change anything in reality. From the perspective of the industry these are and most probably will remain only curiosities And "1296 speaker device" is even something more extreme, a real monstrosity
It would work also by 144 speakers, but it wouldn’t be perfect
graaf said:
I am conform by Ralph Glasgal when he wrote on the site:
"Tonal accuracy is the best that can be hoped for in a traditional audio system; true spatial accuracy will never happen. Audio products should come bearing this disclaimer: WARNING: IMAGE PRESENTED IS LESS THAN LIFELIKE."
The fundamental problems of conventionally reproduction are indeed the early reflections. Its starting points are the mirror sources everywhere in the recording room. The attempt to reduce its spatial distribution upon a few channels must lead inevitably to utterly different perception.
The loudspeakers in the playback room generate an utterly different reflection pattern. That is shown in that small animation for the undirected radiation in main tone range:
http://www.syntheticwave.de/pictures/Stereo50.swf
It is revealing the utterly different impulse response. More adapted perception only is possible, if the recording room and the playback room vaguely similar. In all other cases only the suppression of the wrong playback room reflections remains. But by that way only the speakers remain as sound source. The reproduction becomes reduced only upon the few sources in the horizontal level of the speakers. Hardly IACC attractions arise, the sound becomes boring.
Only way for avoid that problem is changing the playback room acoustic dependent of the reproduced program. In that matter my idea is very less expensively as continually employee fife craftsmen’s.
The problem with any of these multi source systems is the audio quality of each single source. The source problem will dwarf the image problem unless high quality sources are used, and then you are talking mega-bucks in sources alone. I have heard many of these "image field" enhancers and each one of them lack sound quality because of the poor sources. The images were where they were supposed to be, that much worked, but the sound quality was very poor and ruined the whole thing IMO.
gedlee said:
OK!
I accept those corrections, thank You
anyway, the numbers are all approximations
my approximations are taken from Josef Manger’s presentation "Acoustical reality", they are after Eberhard Zwicker and Hugo Fastl ("Psychoacoustics – facts and models" Springer Verlag Berlin 1999) I believe (though I am not sure)
anyway, the numbers You gave and the numbers I gave after Manger are not so much different, not significantly different
You say "first 10 ms are critical for imaging" - OK! I concur!
but the sense of the direction of sound is established sooner, isn’t it? the precedence effect, it's <1 ms I believe
from "Introduction to Sound Recording"
see: http://www.tonmeister.ca/main/textbook/node339.html
even more scientifically put:
from "The precedence effect" ASA Paper
see: http://www.waisman.wisc.edu/~litovsky/papers/1999-3.pdf
You say "first 10<20 ms are critical for timbre" - OK! I second that as well
10<20 ms or 10<30 ms - what’s the practical difference? I believe none
best,
graaf
graaf said:
OK!
I accept those corrections, thank You
anyway, the numbers are all approximations
my approximations are taken from Josef Manger’s presentation "Acoustical reality", they are after Eberhard Zwicker and Hugo Fastl ("Psychoacoustics – facts and models" Springer Verlag Berlin 1999) I believe (though I am not sure)
anyway, the numbers You gave and the numbers I gave after Manger are not so much different, not significantly different
You say "first 10 ms are critical for imaging" - OK! I concur!
but the sense of the direction of sound is established sooner, isn’t it? the precedence effect, it's <1 ms I believe
from "Introduction to Sound Recording"
see: http://www.tonmeister.ca/main/textbook/node339.html
even more scientifically put:
from "The precedence effect" ASA Paper
see: http://www.waisman.wisc.edu/~litovsky/papers/1999-3.pdf
You say "first 10<20 ms are critical for timbre" - OK! I second that as well
10<20 ms or 10<30 ms - what’s the practical difference? I believe none
best,
graaf
In small rooms these small time differences make BIG perceptual differences because the time delays are always in this range.
The "Summing localization" and "precedence" effects are widely overstated in these discussion because they indicate - as you said - the principle direction of the sound perception. They do not indicate whether there is an increase in image blur or coloration, both of which exist for these small delays. SO yes the "principle" direction is set very early 1-2 ms, but the stability and coloration of that image is strongly influenced by the next 8-10 ms. I want a stable and uncolored image, not just a "good idea" of which direction its in. Blauert admiots such in his "Spatial Hearing" book.
gedlee said:
well, in small rooms, in most listening rooms, with standard loudspeaker positioning the time delays are typically <10 ms
this is why I say 10<20 ms or 10<30 ms is no difference in practice
only from that perspective
I’ve demonstrated in my first post in this thread that by means of special positioning of special loudspeakers in a room we can achieve a delay of around 10 ms for every reflection, floor and ceiling included, in bigger room perhaps even up to 20 ms but with very high ceiling, a really big room
and 30 ms of delay is not realistically achievable at all
I understand that with CD speakers like Summa or AI-Audio something different is achieved - the level of reflected sound is lowered, reflections are as if they were filtered with a low pass filter
and with the positioning I proposed the reflections are not filtered nor lowered in level but delayed by around 10 ms or more
are the effects comparable? can this two approaches be regarded as equivalent alternatives? like in time/intensity trading?
can we just choose a solution best suited to our rooms acoustics or décor?
who knows?
I don’t know
but why not? it makes sense
best,
graaf
Graaf
I tried to go back and read your first post. It is not clear enough to me to agree or disagree, only that with omni-directional loudspeakers in a small room like that I would expect a whole lot of VER (< 10 ms.) coming from all over the place. You seem to "assume" otherwise. In my rooms I measure the impulse response so I KNOW that my VERs are very low. I think if you measure the rooms impulse response for your setup you will find that its not what you think it is.
I tried to go back and read your first post. It is not clear enough to me to agree or disagree, only that with omni-directional loudspeakers in a small room like that I would expect a whole lot of VER (< 10 ms.) coming from all over the place. You seem to "assume" otherwise. In my rooms I measure the impulse response so I KNOW that my VERs are very low. I think if you measure the rooms impulse response for your setup you will find that its not what you think it is.
gedlee said:
I don’t "assume" otherwise
I just did some simple calculations, perhaps to simple, I don’t know
I drew the sound path lines, calculated them and then I calculated the delays. The only assumptions I made is that "the speed of sound = 344m/s"
and also that "angle of incidence = angle of reflection"
two assumptions
I have found similar way of reasoning and same assumptions in Harold Beveridge brochure (see image attached):
see: http://www.bevaudio.com/White_Paper.htm
is anything wrong with that reasonig? the law of reflection does not apply in this case? I don't know - I assume that it applies
Is this assumption wrong?
best,
graaf
Attachments
Thank you Graaf for your answer and thank you Earl for completing
I have been listening to some music with my LS lying along the side walls of my living room. There are things I do like and things I like less with this set up. (My LS are FE167E Fostexes in a double BR box)
1) the WAF is bigger! my girlfriend said it was much better looking since the LS almost disappeared instead of being in the "middle" of the room.
2) the bass output is much stronger. This is due to the fact that the LS is much closer to 2 surfaces: the floor and the side wall. Nevertheless I can now locate where the bass is coming from what I couldn't do before. It is not pin point localisation but still!
3) I am not sure if the room disappeared but I can say that there is a feeling of space that I did not have before. For example, with a recording made in a church, the church has become much bigger than before.
4) I will tend to say that I have a more 3D feeling but I lost some depth, that for sure! the 3D feeling might come from the LS positioning but as well from the fact that I don't see the drivers any more (they are hidden from my view by two couches that are facing each other along the side walls).
5) this one seems a bit contradictory to me and it is kind of hard to explain: I have deeper silences between the notes but it doesn't sound as clear as with the drivers facing me. The "deeper silences" thing comes from the fact that I can now hear some noises from the CD player while listening to music! About the clear sound: I have the feeling that I am much far away from the musicians than before. This is maybe linked to the lake of treble I feel with this set up...
6) The sweet spot has grown for a tiny little spot to an area!
Conclusion: I like it very much so far and I will keep it like that for a while to see what it does to me in the long run. It makes me realise how much the reverberations pollute, distort the music. Bass speaker in a corner is very good as well (as long as it is not dipole bass!) I will certainly take all these experiences into account for the building of my next pair of speakers.
It also makes me realise that I am missing quite some knowledge about psychoacoustics. Any good book you recommend there? I checked Earl 's book but I wish something covering the topic in some more depths.
I am very interested by the question you lifted up whether the early reflexions should be delayed over 10ms or if they should be lowered in intensity (by means of absorption and/or diffusion).
Regards,
Etienne
I have been listening to some music with my LS lying along the side walls of my living room. There are things I do like and things I like less with this set up. (My LS are FE167E Fostexes in a double BR box)
1) the WAF is bigger!
my girlfriend said it was much better looking since the LS almost disappeared instead of being in the "middle" of the room.2) the bass output is much stronger. This is due to the fact that the LS is much closer to 2 surfaces: the floor and the side wall. Nevertheless I can now locate where the bass is coming from what I couldn't do before. It is not pin point localisation but still!
3) I am not sure if the room disappeared but I can say that there is a feeling of space that I did not have before. For example, with a recording made in a church, the church has become much bigger than before.
4) I will tend to say that I have a more 3D feeling but I lost some depth, that for sure! the 3D feeling might come from the LS positioning but as well from the fact that I don't see the drivers any more (they are hidden from my view by two couches that are facing each other along the side walls).
5) this one seems a bit contradictory to me and it is kind of hard to explain: I have deeper silences between the notes but it doesn't sound as clear as with the drivers facing me. The "deeper silences" thing comes from the fact that I can now hear some noises from the CD player while listening to music! About the clear sound: I have the feeling that I am much far away from the musicians than before. This is maybe linked to the lake of treble I feel with this set up...
6) The sweet spot has grown for a tiny little spot to an area!
Conclusion: I like it very much so far and I will keep it like that for a while to see what it does to me in the long run. It makes me realise how much the reverberations pollute, distort the music. Bass speaker in a corner is very good as well (as long as it is not dipole bass!) I will certainly take all these experiences into account for the building of my next pair of speakers.
It also makes me realise that I am missing quite some knowledge about psychoacoustics. Any good book you recommend there? I checked Earl 's book but I wish something covering the topic in some more depths.
I am very interested by the question you lifted up whether the early reflexions should be delayed over 10ms or if they should be lowered in intensity (by means of absorption and/or diffusion).
Regards,
Etienne
graaf said:
344m/s is OK for the speed of sound in the air at 20ºC at see level!
click this link if you need to discuss that in details.
More seriously: your second assumption is not always true. It is not true for the lower frequencies were ray theory can not be applied any more, wave theory has to be applied.
Where does this transition take place is the obvious next question!
First this transition does not take place at one unique frequency below which wave theory applies and above which ray theory applies. The transition takes place over 2 octaves during which both ray and wave theories rule.
The frequency where it starts (= the frequency below which wave theory can be applied without doubt) depends on the volume of your room and the absorption properties of the room. It follows the following formula:
F2=11250*SQRT(RT60/V) RT60 being the reverberation time in seconds and V being the volume of the room in cubic feet.
This formula yields F2=180Hz for a room of 63 m3 (25 m2 by 2,5 m which is more or less you room, if i remember well...) with a RT60 of 0,5 seconds.
The other frequency (the one above which ray acoustics can be applied without doubt) is 4 times the previous one.
This yields F3=720Hz for the same room as before.
All this comes from the Master Handbook of Acoustics by Alton Everest, page 324.
From that I can say that the assumption "angle of incidence = angle of reflection" is true from the highest frequencies down to F3, is less and less true from F3 to F2 and is unappropriated below F2.
Regards,
Etienne