I think part of the problem is just in definitions. What many are calling nearfield is not the same as in a studio environment where it is truly nearfield, less than a meter away from the speakers sitting on the top of the console and the walls many meters away. In our own homes it is nothing like that unless you have the speakers out in the middle of the room and you are again sitting within a meter of the speakers in a triangle. So semantics do come into play in the words that we use. I don't know anybody who really listens in the same way that is done inside the studio on nearfield monitors.
At home, listenning pleasure is what really matters. I don't consider myself a critical/serious listener, except when i am tweaking a loudspeaker, which is a different kind of pleasure...and a real pain at the same time, like critically listening to a mix,btw, a job, a task, etc, a pain...
Last edited:
Ok, maybe, sorry... i seldom read a file like a book from page 1...
Though this one is only about 40, while others...hundreds...
No worries, you could try the shuffler, or the pre-shuffled tracks put up by Mike. Though it would only make sense trying this with the absence (or suppression) of early reflections.
I noticed 2 simple EQ cuts cleared up coherency on some material I listened to.
In hindsight they are an exact fit with this whole combing problem inherent in Stereo listening.
While putting up more speakers would help the center, you're actually creating more combing problems between the dead left, center and right positions. With this trick, as that's how I view it, it can be seamless from far left to far right without any cue to exact speaker positions. If everything works as planned of coarse.
I disagree. The left-center combing effect is less of an issue compared to the symmetrical left-right effect.
Plus, movies have a discrete center mix. Music that is mixed for surround either uses just the L/R for stereo pairs and puts spot mics in the center, or uses an actual surround mic that again, has a discrete center (or can output one, ie. from an ambisonic mic). I guess it varies by material, but in the context of movies, the center channel is typically quite isolated from the other speakers.
I've had enough of the shufflers and such for now, but will revisit once I have a listening setup that might benefit more.
This is actually what I meant when I referred to nearfield. At a desk, etc.
You've got a point there, that will be a big difference in favour of the multichannel setup. I'm still not making a third tower though
.I hope you do revisit at one point, I'm going to live with my (max 25 degree) shuffler setup for a while. I might (have to re-)adjust tonality over a longer listening period and make up my mind after that. I'll also include it to my pseudo HT setup. View a couple of movies that way (with my flawed 4.0 setup, can't call it proper surround but it works beyond expectation)
What didn't work for me was real 4 channel quad mixes. It becomes pretty clear that the ambient surrounds I run have no bottom end. It has not bothered me with movies yet, the effects seem to work quite well. Way better than I ever imagined.
Yes i have been playing with M/S similar corrections and find them usefull when a less/more than ideal 60º triangle is imposed as the result of atypical ( and not so atypical...) room conditions, but i believe that adequate speaker design and placement, avoiding any weird room acoustic treatment to absorb mid and high frequencies, is enough to avoid the side effects commented in this file.
Regarding the room, i totally agree with Linkwitz when he says that a room is good enough for playing music in it, when it is good enough to make normal life in it...
Also believe that Toole and followers would make better sounding speakers, if designed in and for normal rooms, without wall carpeting disasters like in cinemas and similar horrors...
Last edited:
Neither do I. It's the sides that bother me.
Since so much is panned to dead center, we usually adjust the system so that the center image has correct tonal balance. Once that's done, anything panned hard to one side will sound brighter, and often a bit drier, than sounds in the center. Not something you might notice until an instrument or voice is panned across. Once you hear it, it's hard to ignore.
Of course if you are trying to do movie/TV dialog that was mixed with a dedicated center channel, then the phantom center will be a bit dull. A dip at around 2K and 5K won't do wonders for speech intelligibility.
Neither do I. It's the sides that bother me.
Ok, i now understand. I have the same problem and it is the main issue i try to fix when setting up a pair of speakers.
What works fine for me is to avoid listening the speakers on axis, directed to the listening point, but insted, listening clearly off axis, crossing the axes aprox half way between the speakers and listening pos.
Off course this also requires that the speakers have been eq to give flat response off axis and have a directional enough pattern to keep power response under control too. The idea is to reduce the tonal difference between direct and reverberated sound, the worst case being when speakers are listened to straight on axis...
Imho, we usually pay too much attention to direct sound and give at best a secondary role to the reverberated sound. We think as purists, and this is what is wrong for me...
Last edited:
Also believe that Toole and followers would make better sounding speakers, if designed in and for normal rooms, without wall carpeting disasters like in cinemas and similar horrors...
I think you misunderstand Toole. He is definitely designing for normal rooms, and is probably the most vocal advocate of a smooth off-axis response that is tonally similar to the direct sound.
He has some recommended treatments for dedicated listening rooms, but not the sidewalls... not everyone agrees with this, but in my experience it substantially ameliorates the "dull center" effect.
It is not an intelectual question. I like his papers, but don't like the speakers his company sells. None of them. So there must be something there...
Btw, what i do has little to see with what i have been reading, since what i do is deliberately creating a rising on axis response and then listen off axis, when the response has become flat.
I believe that the question is not that off axis must be smooth, this is not enough. What i try to achieve is that direct sound has also the same color as reverberated sound, or as close as possible. Off axis /On axis does not really matter for me, direct/reverberated is what i try to control.
In other words, i try to mimic what i find important in an omni point source (which for me is the only acceptable one), with a directional one ( which are also the only available ones in real life).
Last edited:
What if you extract an L-XR and R-XL signal (stereo matrix using opamps), and an L+R signal, from the stereo pair (not after the matrix) (so now you have 3 signals instead of 2), adjust X to at least .5 (.7 might be better), decorrelate the L+R signal with multiple phase delays (so comb cancellations might be somewhat filled in) (you may only need 2 or 3 delays to get a good result), then mix the decorrelated L+R back in with the L-XR/R-XL, such that you're back to a stereo feed with approximately equal amplitude balance between L, R and center.
There will be some crosstalk which will to some degree creates comb filter cancellations, but that's happening anyway, both from the phantom center image coming out of two speakers, room acoustic reflections, and sometimes issues in the recording process. In a typical room, it's the multiple reflection delays that fill in each others cancellations, and here we're adding a few more much shorter delays, but same general theory.
It seems like this would make the phantom center image have more of a sense of separateness and clarity, maybe widen the sweet spot area, and the bottom line comb filter situation may not necessarily be any worse than it already was, in a normal room.
As far as the in between images (between L and center, or R and center), and with L to R pans, I'm skeptical as to how much this approach would actually damage that. Maybe some, but clarifying the center phantom image is probably going to involve some sort of tradeoff one way or another.
Maybe a better way would be to put a 4 pole active crossover ahead of any processing, only process freqs above 1kHZ as described above, then recombine (mix) everything back to a wideband stereo pair. It seems that the main reason the phantom center image is "phasey" or dull, is because of the size of the wavelengths of the frequencies above about 1kHZ... so maybe we should leave freqs below 1kHZ alone.
If two delays have a 1:1.4 or 1:1.62 time ratio, they may do a better job of filling in each others cancellations over frequency. Cancellations may be less likely to ever double up in the harmonics.
There will be some crosstalk which will to some degree creates comb filter cancellations, but that's happening anyway, both from the phantom center image coming out of two speakers, room acoustic reflections, and sometimes issues in the recording process. In a typical room, it's the multiple reflection delays that fill in each others cancellations, and here we're adding a few more much shorter delays, but same general theory.
It seems like this would make the phantom center image have more of a sense of separateness and clarity, maybe widen the sweet spot area, and the bottom line comb filter situation may not necessarily be any worse than it already was, in a normal room.
As far as the in between images (between L and center, or R and center), and with L to R pans, I'm skeptical as to how much this approach would actually damage that. Maybe some, but clarifying the center phantom image is probably going to involve some sort of tradeoff one way or another.
Maybe a better way would be to put a 4 pole active crossover ahead of any processing, only process freqs above 1kHZ as described above, then recombine (mix) everything back to a wideband stereo pair. It seems that the main reason the phantom center image is "phasey" or dull, is because of the size of the wavelengths of the frequencies above about 1kHZ... so maybe we should leave freqs below 1kHZ alone.
If two delays have a 1:1.4 or 1:1.62 time ratio, they may do a better job of filling in each others cancellations over frequency. Cancellations may be less likely to ever double up in the harmonics.
I have been wondering about this myself. Jim Bongiorno created his 'Trinaural Processor' ...
Trinaural Processor' SST
I have vivid memories setting across a table at a diner attempting to get him to explain the details of his invention, but he refused. Be clear tho in the case of the Trinaural Process the Left and Right channels are processed to create a Center channel. A physical Center channel loudspeaker is required but it seems to me some of what Jim was doing should apply to our discussion here.
It is not an intelectual question. I like his papers, but don't like the speakers his company sells. None of them. So there must be something there...
Btw, what i do has little to see with what i have been reading, since what i do is deliberately creating a rising on axis response and then listen off axis, when the response has become flat.
I believe that the question is not that off axis must be smooth, this is not enough. What i try to achieve is that direct sound has also the same color as reverberated sound, or as close as possible. Off axis /On axis does not really matter for me, direct/reverberated is what i try to control.
In other words, i try to mimic what i find important in an omni point source (which for me is the only acceptable one), with a directional one ( which are also the only available ones in real life).
I see you don't like JBL's. Well, I haven't really liked any I've heard either. But, I quite like the recent Infinity speakers with waveguide tweeters, I have a pair and their off-axis response is outstanding. Would these meet your criteria?
Would the JBL M2?
An externally hosted image should be here but it was not working when we last tested it.
I don't know what you mean by this: "Off axis /On axis does not really matter for me, direct/reverberated is what i try to control." Those are intrinsically related! Actually I may know what you mean. The direct sound could be the off-axis sound, if you are toeing in.
And I agree with toeing in the speakers to aim the brightest sound off-axis, though "it depends". Yet, this is exactly the setup of a typical waveguide speaker, ie Geddes. I like this idea generally for several reasons - widening the sweet spot (farther speaker is more on-axis), balancing the reflected energy better (walls tend to absorb treble, so the added brightness is less problematic as a reflection), and mitigating diffraction effects (response is usually smoother slightly off-axis).
I continue to be a Toole apologist because he has arrived at these criteria not by guesswork but by extensive blind testing and not just in that Harman room you pictured. At Canada's NRC is a IEC standard listening room similar to this one: Listening room I don't think you disagree with him at all really! Many important Canadian speaker companies were born out of the research Toole headed up at the NRC including PSB, Axiom, and Paradigm. Don't judge harshly based on the JBL's you've heard. (If you have heard all of the speakers I've mentioned above and don't like any of them, I would be surprised).
Anyway back to phantom center business.
Cool.. yes and I'm sure you've all seen Elias's 3 speaker matrix pages: http://elias.altervista.org/html/2_vs_3_stereo_high_freq.html
As for the SST processor, it would be pretty cool to do some measurements on one and find out what's going on!
The version I'm building right now will include an actual center speaker, but I thought much of the same circuitry could be used for clarifying a phantom center image with just two spkrs as well.
.....it would be pretty cool to do some measurements on one and find out what's going on!
A beer on its something close to some kind of +/- 20º Pano Rephase shuffler