It doesn't have to be a dead end i.m.h.o. In fact, you might be closer to something great.
It's true that the more clear and reflection free the first wave front at the sweet spot becomes, the more obvious the comb filtering due to cross talk becomes too. But if you add some diffused later reflections to fill in the dips created by the combing...
Jim1961 has shown his (passive) Haas kicker does just that. It's in this thread. The usual description of such a Haas kicker: a more 3D sound and imaging without being aware of the sound coming from behind. You don't hear it as extra sources, just a better defined front.
I still use mid/side EQ, just not the standard S-curve mentioned in this thread. I use it to adjust tonal balance (between phantom centre and sides) making sure it worked for a wide variety of songs and genres. But the cross talk dips get filled in by the virtual Haas kicker I use.
When playing 5.1 material over my 2 loudspeakers I do notice tonal balance differences. As that true centre channel mix wasn't adjusted to compensate for cross talk effects. So my processing there is (only slightly) different.
I've sat with the barrier to test, no way I would want to do that on a regular basis
. I've also played with cross talk cancelation software, more than one in fact including my own variants. Those gave me more problems than they solved and eventually I gave up on that.I've experimented a lot with the shuffler, in the end I had to conclude it didn't really solve the problem without creating some new side effects. After hearing the side effects I could find it in the measurements and wrote about it here.
Using the Haas kicker? Definitely a keeper for me (as long as it doesn't draw any attention to itself). It's band passed and attenuated. Mine isn't passive, meaning I have full control on what they play. Key is to have its output diffused.
Mid-side EQ? also a keeper for me, gradually adjusted it to my liking over time. This showed me I don't alter the phantom centre (like I thought I should when I started with it) but adjusted the side tonal balance for a better overall balance. Even though the differences are pretty small between mid and side, it still makes enough of a difference in listening.
I may at one point revisit some of the possibilities. I try to do that every once in a while. With a fresh approach and armed with new info, who knows.
For me the Haas kicker and mid/side EQ win for this particular phantom problem (in my specific setup). No fatigue problems and I've listened to this for months now, still leaving me impressed with what it does.
It sounds natural, in the end that's what I was looking for. No tizz and boom though.
The best spot? Definitely the sweet spot. Though moving to the sides isn't a punishment. The sweetspot is just that, the best spot to listen and dream away listening to your favourite tunes. You'll know when it's good if time flies and you can't stop listening. Not for a day, but for weeks on end.
Last edited:
Hi wesayso, thank you for the response to my cul-de-sac comment. Jim's active HAAS kicker is very interesting, thank you for the input. I checked GS posts.
If back speaker can be used, there is another way to cancel out comb filter effect connecting rear speakers like this:
Amp L+ out to Back Speaker L+
AMP R- to Back Speaker R-
Back Speaker L+ to Back Speaker R+
Back Speaker L plays L-R, which should cancel out the Front R signal at the left ear, vice versa, at least theoretically. I read a report saying this method worked very well, so it could be a good solution as well. Please someone try it. Note: This type of connection could be harmful to the amplifier of certain output circuit.
Jim's active HAAS kicker is very interesting, thank you for the input. I checked GS posts.If back speaker can be used, there is another way to cancel out comb filter effect connecting rear speakers like this:
Amp L+ out to Back Speaker L+
AMP R- to Back Speaker R-
Back Speaker L+ to Back Speaker R+
Back Speaker L plays L-R, which should cancel out the Front R signal at the left ear, vice versa, at least theoretically. I read a report saying this method worked very well, so it could be a good solution as well. Please someone try it. Note: This type of connection could be harmful to the amplifier of certain output circuit.
Last edited:
That's the old surround trick, popular in the 70's was it? Don't do it with just any amp though, you might fry it. Even though I do use a similar signal (L+R) and (L-R) in my ambient setup, that alone does nothing for the phantom centre or cross talk cancelation. It's just a variation on Hafler Stereo.
You can do miles better than that by using a band passed, attenuated and properly delayed (L-R) and (R-L) signal. My specific interest went out to mixing in some centre, as a passive diffusion panel would reflect centre information too.
Jim has a combo of passive panels and later added active helpers with reverb possibilities.
His room and projects are amazing. It shows what can be done with dedication. Not a single sound hits that listening spot without being carefully planned.
I've played with pretty much the same ingredients, though completely active and infinitely adjustable. One of the more succesfull projects/experiments I did that I can highly recommend.
Even Linkwitz played with this idea: Surround stereo system
I guess his "Watson-SEL" experiment started life as an anti cross talk solution: WATSON-Stereo_Expansion_Loudspeakers
No idea if any of his experiments passed the test of time. I first tried it in September 2015 and haven't removed it since except to run the occasional tests etc. I revisited it later to play with reverb, both reverb trough algorithms and IR's. The algorithm's won that test.
With home theatre they double as "normal" surround speakers.
Another interesting read on that Hafler wiring: http://www.diymobileaudio.com/forum/system-design-help-me-choose-equipment-my-car/9806-rear-fill-do-you-use.html
Look for the posts by "werewolf". That's who sparkled my specific plans. His later handle was Lycan. Always interesting to read what he wrote about.
You can do miles better than that by using a band passed, attenuated and properly delayed (L-R) and (R-L) signal. My specific interest went out to mixing in some centre, as a passive diffusion panel would reflect centre information too.
Jim has a combo of passive panels and later added active helpers with reverb possibilities.
His room and projects are amazing. It shows what can be done with dedication. Not a single sound hits that listening spot without being carefully planned.
I've played with pretty much the same ingredients, though completely active and infinitely adjustable. One of the more succesfull projects/experiments I did that I can highly recommend.
Even Linkwitz played with this idea: Surround stereo system
I guess his "Watson-SEL" experiment started life as an anti cross talk solution: WATSON-Stereo_Expansion_Loudspeakers
No idea if any of his experiments passed the test of time. I first tried it in September 2015 and haven't removed it since except to run the occasional tests etc. I revisited it later to play with reverb, both reverb trough algorithms and IR's. The algorithm's won that test.
With home theatre they double as "normal" surround speakers.
Another interesting read on that Hafler wiring: http://www.diymobileaudio.com/forum/system-design-help-me-choose-equipment-my-car/9806-rear-fill-do-you-use.html
Look for the posts by "werewolf". That's who sparkled my specific plans. His later handle was Lycan. Always interesting to read what he wrote about.
Attachments
Last edited:
I have tried the barrier with the expectable result, that it sounds like headphones plus some reverb of one's listening chamber.hahaha, have you tried the barrier by yourself? I love it.
Anyway, those Japanese invention are probably not what you call Beam, they are Reverse-Olson. I like to hear more detailed explanation about your Beam idea.
Records should contain intensity stereophonics only. This is a premise of mono compatibility. Runtime stereophonics should be caused either by our head in case of loudspeakers playback or a (rather simple, certainly possibly analogue) electric processor in case of headphones playback. (Headphones playback circumvents the acoustic function of our head, so it must become simulated.) Common array stereo loudspeakers introduce runtime differences within the space well before our head, what is wrong. Only our head or a substitute for it should cause runtime differences.
plasnu, I had a topic here, too, describing the concept of my GA13 loudspeaker and showing the beam stereo, which I had as second idea equipped it with after having been stumbled upon it here. We demolished the topic, because some things wrap around number space. Stereo playback theory
Beam stereo usually causes runtime differences, too, since the mono beam is not very narrow but reaches reflecting side walls, too, and the stereo beams directed to side walls are not very narrow but reach the listener directly, too.
We build loudspeakers usually as pairs and run them simultaniously. Yet we do not show response plots of that situation at a typical ear position 20cm to one side off axis but always only the response of a single loudspeaker. This is shizophrenic or dishonest and why, I search for a remedy.
Yes, quite.
When I started this thread it was the tonal balance problems I was hoping to solve. Sounds panned hard right or hard left sounded brighter on my system than sounds in the center. It's an artifact of a clean system, tho I didn't know that at the time.
For me, the overall system tonal balance is dominated by the the phantom center, probably because there is so much important content there. It wasn't that I wanted to brighten the center, so much as to dull the sides a little to match the center. That's when I began reading about the problem and found the shuffler. EQ was the path I thought I'd follow, until the cause of the tonal shift was made clear - crosstalk causing cancellations.
Using M/S EQ may still give the best results, or least offensive side effects. Perhaps just a notch filter or 2 on the sides to match the sound of the center.
All the imaging results and opinions were beside the point to me, that was not that problem. If getting a smooth tonal balance across the front means a better image, all the better.
Nice writeup Pano
Thing that is annoying is some mastering have taken care of subject some have not and then there is panned mono recordings, so even there is solutions its a mess of decisions when to correct or not and with what filter that said this is really great thread digging into subject.
Thing that is annoying is some mastering have taken care of subject some have not and then there is panned mono recordings, so even there is solutions its a mess of decisions when to correct or not and with what filter
that said this is really great thread digging into subject.
wesayso, quality was superior and got so impressed as to never forget when i sat there in sweet spot sucked into recording room and great musicality performance i couldn't know if you tweaked anything on computer down the road, so good to know you had ambient channels on plus mild mid/side EQ all the way thru the various tracks we listened .
when i sat there in sweet spot sucked into recording room and great musicality performance i couldn't know if you tweaked anything on computer down the road, so good to know you had ambient channels on plus mild mid/side EQ all the way thru the various tracks we listened .
Yep, this would be the side effect of the barrier. I also have noticed that the barrier alters natural room ambience somehow. I thought it was in a good way, but further listening test would be needed for sure. IME, when the speakers and room get better (objectively better spec), it sounds somewhat more similar to headphones, ironically...
Last edited:
You're welcome! It arose out of a minor irritation, a minor flaw that wanted fixing. It's been interesting to read about other people's experiences and problems.
I always used to joke about how as time passed and my speaker projects got better technically, what really happened was that the remaining distortions became more and more clear... so overall enjoyability didn't necessarily change much. I think there is some truth to that. In the case of trying to get a phantom center image to be more clear and stable (when listening off axis somewhat), it seems that the problem is primarily high frequency comb filter effects which vary significantly with even small head movements (due to the size of the wavelengths involved).
The thing that often minimizes room acoustics comb filter effects is oddly more reflections. Each reflection creates a comb filter effect, often with rather deep cancellations. Other reflections with different time delays create comb filter effects, but the cancellations are at different frequencies because of the different time delays involved. You add enough of these together, and most or all of the cancellations get mostly filled in. An FR at the listening position that used to have 12dB variations may now only have 3 dB variations.
So here's my theory: Take an L+R signal that's BW limited to above about 800HZ, delay it a few times with very short delays (less than a millisecond or so), add those in with this L+R signal, then remix that "blurred" complex comb filtered L+R signal back into the stereo mix, but down maybe 3-6dB. Some EQ may be needed because of the addition being only above 800HZ.
So then you've effectively created additional reflections with emphasis on the L+R images, above 800HZ, that will add to the unprocessed signals, thereby reducing or minimizing any major comb filter effects in the phantom center image at the high frequencies wherever you sit. If the delays times are chosen strategically, none of their comb filter effects would cause cancellations at the same frequencies on axis, and may be an improvement when added to the L and R signals off axis.
So you're adding a "reverb" to the L+R signal that is way too short to perceive as reverb, to minimize the depth of comb filter effects created by just two speakers trying to create a phantom center image, when all the signals add together acoustically. It would be interesting to try anyway. Maybe some of you already have (?).
The thing that often minimizes room acoustics comb filter effects is oddly more reflections. Each reflection creates a comb filter effect, often with rather deep cancellations. Other reflections with different time delays create comb filter effects, but the cancellations are at different frequencies because of the different time delays involved. You add enough of these together, and most or all of the cancellations get mostly filled in. An FR at the listening position that used to have 12dB variations may now only have 3 dB variations.
So here's my theory: Take an L+R signal that's BW limited to above about 800HZ, delay it a few times with very short delays (less than a millisecond or so), add those in with this L+R signal, then remix that "blurred" complex comb filtered L+R signal back into the stereo mix, but down maybe 3-6dB. Some EQ may be needed because of the addition being only above 800HZ.
So then you've effectively created additional reflections with emphasis on the L+R images, above 800HZ, that will add to the unprocessed signals, thereby reducing or minimizing any major comb filter effects in the phantom center image at the high frequencies wherever you sit. If the delays times are chosen strategically, none of their comb filter effects would cause cancellations at the same frequencies on axis, and may be an improvement when added to the L and R signals off axis.
So you're adding a "reverb" to the L+R signal that is way too short to perceive as reverb, to minimize the depth of comb filter effects created by just two speakers trying to create a phantom center image, when all the signals add together acoustically. It would be interesting to try anyway. Maybe some of you already have (?).
Yep Bob, that's basically what the Shuffler does, at least the original one. It adds some clean echo to both channels, in alternate phase. It isn't limited to the L+R signal, but applied to Left and Right separately to kill the comb filtering. See post #13.
I made another version using the rePhase software that doesn't use the echo effect, but relies simply on phase shift. It seems to work as well, with fewer side effects. Neither is perfectly transparent, tho.
I made another version using the rePhase software that doesn't use the echo effect, but relies simply on phase shift. It seems to work as well, with fewer side effects. Neither is perfectly transparent, tho.
One of the most popular pop vocal effect in the professional music studio has been Eventide H3000 Micro Pitch Shift. In 90's, before internet age, it was a secret weapon among professional engineers.
What is does is adding 2 different slightly delayed and slightly pitch shifted vocal signal to L and R separately. All the pro studio had them, and this very expensive pitch shift effect unit is used for this preset ONLY almost all the time.
What is does is adding 2 different slightly delayed and slightly pitch shifted vocal signal to L and R separately. All the pro studio had them, and this very expensive pitch shift effect unit is used for this preset ONLY almost all the time.