Thanks for this reply. So, let's imagine a place outside where one can stand and walk, but the surface is somehow 100% acoustically absorbative - just to get rid of any ground-bounce effects.
You're standing there and some distance in front of you, a person is walking from your far right, following a line that passes in front of you and goes to your far left. S/he's talking / singing / playing a sax continuously as this path is walked...
So the tone of their sound you hear will change due to comb filtering due to the distance between your ears? I would have figured that part would be essentially the same, just time and SPL differences at ears arrival cuing the brain "S/he's over to the right" - "S/he's in front of you" - S/he's over to the left".
I guess in my dull-mindedness, I've never really paid attention to the timbre of some sound source being different, depending on where it is about you. My perception's concern seems to stop at "point to where the sound is coming from". Not gee, Fred_the_bear sounds a little bit more like Ted_the_bear, when he's right in front of me.
Did i ever tell the story of when I was hanging out with a bunch of audiophiles, the guy made some small change to his system and everyone (but me...) jumped out of their seat in response? I'm sure one could be trained to hear this. On the other hand, perhaps blissful ignorance... :')
No, because this is ONE person, one source, from one direction. Not the same as two speakers playing the exact same thing. That's the problem, two speakers playing the SAME signal and the timing of how that arrives at your ears. If your head were 6mm wide, then there would be no comb filtering in the audible range. But the ears spaced about 15cm apart, it becomes a problem when two sources are identical.
Oh no! it won't change that easily. The tonal balance gets disrupted because we try to emulate a sound at the (phantom) center using 2 loudspeakers, while the proper solution to mimic your "reality" would have been one moving speaker
. We play 2 speakers together to mimic a sound source that isn't located there. We can play with level and timing to place it anywhere between the two speakers, and actually, even beyond the speaker using both speakers. But the combing will be there because of those two separate speakers.Like I said, if it doesn't concern you...
A few posts ago I showed an IR as measured at my listening spot...
Although there are (enough) reflections, as it is just a relatively normal room, the level of reflections is quite low for the first 20 ms.
It makes all the difference under which circumstances one hears the speakers. If you move close to a pair of Studio monitors, the
cross talk effects and disruption of tonal balance will be way less noticeable too. If there were a couple of peaks in that graph of a
higher level, I wouldn't have to resort to all this trickery. But I actually wanted this
. I wanted to get rid of the queues of my relativelysmall living room. I substitute it with something more fitting to the music. Oh well, if it works for me
. My-Fi.But yes, we can learn to listen and it isn't always the best thing to do for yourself. I've recorded the sound at my listening spot, and
played it back on the same speakers... Doubling up the room effect that was present. It was quite hard to unlearn what I had heard,
while most people aren't really 'that aware' of what their specific room sounds like.
OK, I'm beginning to understand. Thanks for taking the time with me, guys. Still, makes me wonder who first noticed? Was someone playing with the balance control back in the 60's and noticed their systems voice doesnt sound the same with it centered, playing a mono record? Back then, I was elated just to hear something different emanating from each speaker!
What's a good way to try a mono center speaker - without replicating entirely one of the L / R units? Given someone's IAD of ~14cm, is there a frequency below which a mono center need not be concerned with
I can understand the motivation for that.
As a college student, I attended a demo where the guy did that with two tape recorders, initially recording his voice - repeated until all that was left were the resonances of the room (auditorium). One would think these days with PC based gear "room data" from such a method would be easily obtainable. Maybe REW does it in one shot - but that's not the same as hearing it exaggerated.
I'm really bewildered about this whole issue. How does this change in tone even matter when you're in the sweet spot? You would have to move in either direction out of the sweet spot to hear it but then other factors swamp the effect off axis so how does this matter?
When I've placed my speakers I have never preferred an equilateral triangle. Mind you I have never had a square room, always rectangular. So I start with an equilateral position and incrementally toe them out until it sounds correct. Initially it's always a dark sounding narrow sound stage and as mentioned it changes as I move my head from side to side. The op's issue here for me is just incidental. As I toe them out more and more it brightens up and expands becoming as the sound engineer mixed the recording,(I presume). It now has the cues I assume were intended, ie: the artificial 3D imaging the engineer mixed in. Once I've gone too far and tweaked them back to the optimal position for my room, that dark center is gone and the sound stage is equalized right across with no change in tone. There's no more left and right speaker. It now sounds like one huge speaker with equal source positioning from right to left, obviously depending on the quality of the recording. However the distance from the speakers to the front wall behind them must be exactly the same as the distance from my ears to the back wall behind me.
This is how a stereo system is designed/engineered to function, is it not? When I put on my glasses, I don't see two separate frames. Both sides meld into one and look like a large single perimeter until I pull them far enough away from my face and the separation becomes apparent. I think that's called stereo vision. In the same way, speaker positioning accomplishes the same in terms of stereo sound, no? In any case, I fail to see why tonal imbalance even matters when using the system optimally sitting in the sweet spot.
When I've placed my speakers I have never preferred an equilateral triangle. Mind you I have never had a square room, always rectangular. So I start with an equilateral position and incrementally toe them out until it sounds correct. Initially it's always a dark sounding narrow sound stage and as mentioned it changes as I move my head from side to side. The op's issue here for me is just incidental. As I toe them out more and more it brightens up and expands becoming as the sound engineer mixed the recording,(I presume). It now has the cues I assume were intended, ie: the artificial 3D imaging the engineer mixed in. Once I've gone too far and tweaked them back to the optimal position for my room, that dark center is gone and the sound stage is equalized right across with no change in tone. There's no more left and right speaker. It now sounds like one huge speaker with equal source positioning from right to left, obviously depending on the quality of the recording. However the distance from the speakers to the front wall behind them must be exactly the same as the distance from my ears to the back wall behind me.
This is how a stereo system is designed/engineered to function, is it not? When I put on my glasses, I don't see two separate frames. Both sides meld into one and look like a large single perimeter until I pull them far enough away from my face and the separation becomes apparent. I think that's called stereo vision. In the same way, speaker positioning accomplishes the same in terms of stereo sound, no? In any case, I fail to see why tonal imbalance even matters when using the system optimally sitting in the sweet spot.
This image is the best explanation of the effect that I have found. It is also straight from the horse's mouth (Toole). See what happens with a real center and a phantom center. The 1.6 kHz dip is filled in by shuffling the phase of the two channels, so that they don't cancel because of being in phase in a regular config.
In my setup, the dip is mild and in most rooms, it is filled in somewhat by other reflections (per Toole). But switching the shuffler on and off will give you a clear idea. Once you hear the right tonal balance, you won't go back.
In my setup, the dip is mild and in most rooms, it is filled in somewhat by other reflections (per Toole). But switching the shuffler on and off will give you a clear idea. Once you hear the right tonal balance, you won't go back.
The effect is most pronounced when a substantial amount of early reflections have been removed, either through directivity or absorption. If this is not the case you most likely will not hear the effect. When you have removed those reflections the change in timbre by moving your head can be quite obvious. There is a defined position right in the acoustic centre between the speakers where the tonality is correct and a slight movement away from this imparts a midrange timbre change which is quite obvious in vocals. It is really quite annoying to hear the sound change just by moving your head a few inches. When early reflections are present they fill in this region creating a timbre change of their own but it's not then so position dependent and obvious to hear.
Yes, as stated above, you need a well matched system (not too hard) and early reflections that are low in level. The shift in tonality isn't from you moving around, it's from the sound moving around. The same sound played in the center will sound darker than if it is panned hard left or right. That darkening by comb filtering notches can make dialog difficult to understand in a phantom center image (as opposed to a real center speaker) and that what the paper linked to in the top of the thread is about., The effect can also be heard when a sound in the music mix is panned hard left and hard right making it seem brighter than the center. The graphic posted by RA7 is a good illustration of the major frequency response notch.
I must say that it's a bit strange for a thread to go on over 1000 posts on the topic, as it's not something that most people ever notice or care about at all - at least in stereo music playback. In film/TV dialog it can be a problem, but in music you might not notice. Over the years with various systems and rooms I've heard it and it bothered me, so I set out to find a fix and to understand the origin of the problem. Originally it seemed that a Mid/Side EQ would be the fix, but then I stumbled on the shuffler and it fixes the tonal shift in another way.
I must say that it's a bit strange for a thread to go on over 1000 posts on the topic, as it's not something that most people ever notice or care about at all - at least in stereo music playback. In film/TV dialog it can be a problem, but in music you might not notice. Over the years with various systems and rooms I've heard it and it bothered me, so I set out to find a fix and to understand the origin of the problem. Originally it seemed that a Mid/Side EQ would be the fix, but then I stumbled on the shuffler and it fixes the tonal shift in another way.
Quick thought on the image and related comb filter. If moving along center normal between the speakers (see #1076) this comb filter would also shift in frequency and vary in amplitude as path length difference to each ear from each speaker changes, more or less head on the way. Path length difference being greatest when speakers are 90 deg each side of head, and least when closer to where the nose points at.
This would mean roughly, that the comb filter might get pushed high enough not to bother if speakers are closer together than listening distance. Also, if the speakers are further apart and even though onset of combfilter gets lower in frequency higher up there is now shadow of head redusing sound to other ear some. Anyway, something to think about. I have to try the phase suffling one day and see if it moves the proximity stuff further out in the room as well, to practical listening position with practical speaker position.
Last edited:
As I had EQ-ed my system for phantom center, I perceived brighter sides compared to the (right) phantom balance and decided to combat it with mid/side EQ.
When the subject came up in a different thread, Pano opened up this one to have that discussion central. The original shuffler worked perfectly for me to cure the tonal balance problem, but it made the position of the speakers as sound source obvious. Without the shuffler pointing to the source of the sound for even hard panned sounds was much harder. The phase only shuffler cured that. In the long run I preferred to use EQ only and later lateral reflections (still within the Haas limit) for music and a variation of the original shuffler (from the first post) for movies.
For music the shuffler made me suffer listening fatigue for longer listening sessions. Not noticed that for movies where it just works as advertised.
When the subject came up in a different thread, Pano opened up this one to have that discussion central. The original shuffler worked perfectly for me to cure the tonal balance problem, but it made the position of the speakers as sound source obvious. Without the shuffler pointing to the source of the sound for even hard panned sounds was much harder. The phase only shuffler cured that. In the long run I preferred to use EQ only and later lateral reflections (still within the Haas limit) for music and a variation of the original shuffler (from the first post) for movies.
For music the shuffler made me suffer listening fatigue for longer listening sessions. Not noticed that for movies where it just works as advertised.
- Why didn't I reposition the speakers (toe in/out etc.)to adjust the tonal balance? Because I absorbed early reflections, this wasn't an option.
- Why do I absorb early reflections? To increase intelligibility and clear up the sound and have better imaging and details.
It is why Ambiophonic have the speakers next to each other while using (one of) the race cross talk algorithm(s).
Last edited:
Yeah, or helmet with acoustic panel mohawk for better separation
Hmm, on a stereo speaker setup, if early side reflections are absorbed or reduced by directivity, then it would perhaps be better to put the speakers spaced further out and use shorter listening distance to get more in between the speakers, more shadowing with head. This works at least for me when I just move closer, but have to try again by keeping the listening spot still and space out speakers more. Just thinking out loud, need to test when there is day worth of free time, kids away.
Hmm, on a stereo speaker setup, if early side reflections are absorbed or reduced by directivity, then it would perhaps be better to put the speakers spaced further out and use shorter listening distance to get more in between the speakers, more shadowing with head. This works at least for me when I just move closer, but have to try again by keeping the listening spot still and space out speakers more. Just thinking out loud, need to test when there is day worth of free time, kids away.
Last edited:
I don't hear it. It's puzzling since I'm aware of the dark center which I equalize out and simultaneously bring out the full sound stage. My focus however has never been getting rid of the narrow dark center but rather realizing the full stereo image. Can you give examples of early reflections? Do you mean floor and ceiling? Furniture? Side walls?
Don't get us wrong, you don't need to hear it to lead a happy life. Just enjoy what you have. It need not be a puzzle as the inner workings are there and are clear.
Earlier we had a member post measurements at the ear position in a highly damped room, see: https://www.diyaudio.com/community/...eo-phantom-center.277519/page-31#post-4746283 (think: studio like and beyond). I have some measurements of the exact same thing (somewhere) in my thread as well. It's just physics. I used a foam ball as a head dummy with the microphone pulled trough. I think I remember Jim saying he used a pillow.
Most recordings are made to do just fine, but if you (still) notice higher pitched background singers it's there if the balance in the phantom center is right. A little lack of body for those side panned singers I'd call it, once phantom center balance is taken care of. Fix that tonal difference between center and sides and the background singers become "real people". It was a remarkable difference for me in my room. (huge panels behind each curtain and that poster also is a panel, pillows fill the space between couch and poster)
The effect is there if (the most strong early) reflections are diminished. Floor, ceiling, side wall, back wall and furniture. Especially the first 10 ms but every reflection within the Haas limit will still help work to hide it, but as the Haas limit often is said to be about 15 to 25 ms, it varies quite a bit in my own experience and even 15 ms can be perceived as echo once it's higher than 4.5 KHz and coming from a single side direction.
See my IR a few posts back for an example of "as measured at the listening spot" results, taken at 2.7m distance at my couch. I cleaned up the first ~20 ms and use a Haas kicker to get lateral reflections to keep a spatial effect. (see Griesinger's work on the perception of rooms and halls)
Lack of floor and ceiling reflections are free with arrays like fluid and I use. The rest (side wall, back wall) is done with absorbing panels. Furniture etc. hardly plays a role with a row of drivers line in an array. Only parallel planes need extra attention with speakers like these.
Just enjoy what you have. It need not be a puzzle as the inner workings are there and are clear.Earlier we had a member post measurements at the ear position in a highly damped room, see: https://www.diyaudio.com/community/...eo-phantom-center.277519/page-31#post-4746283 (think: studio like and beyond). I have some measurements of the exact same thing (somewhere) in my thread as well. It's just physics. I used a foam ball as a head dummy with the microphone pulled trough. I think I remember Jim saying he used a pillow.
Most recordings are made to do just fine, but if you (still) notice higher pitched background singers it's there if the balance in the phantom center is right. A little lack of body for those side panned singers I'd call it, once phantom center balance is taken care of. Fix that tonal difference between center and sides and the background singers become "real people". It was a remarkable difference for me in my room. (huge panels behind each curtain and that poster also is a panel, pillows fill the space between couch and poster)
The effect is there if (the most strong early) reflections are diminished. Floor, ceiling, side wall, back wall and furniture. Especially the first 10 ms but every reflection within the Haas limit will still help work to hide it, but as the Haas limit often is said to be about 15 to 25 ms, it varies quite a bit in my own experience and even 15 ms can be perceived as echo once it's higher than 4.5 KHz and coming from a single side direction.
See my IR a few posts back for an example of "as measured at the listening spot" results, taken at 2.7m distance at my couch. I cleaned up the first ~20 ms and use a Haas kicker to get lateral reflections to keep a spatial effect. (see Griesinger's work on the perception of rooms and halls)
Lack of floor and ceiling reflections are free with arrays like fluid and I use. The rest (side wall, back wall) is done with absorbing panels. Furniture etc. hardly plays a role with a row of drivers line in an array. Only parallel planes need extra attention with speakers like these.
Perhaps it’s due to hearing loss on my part. Anything above 10k diminishes rapidly. I also have a 4k hole and tinnitus.Don't get us wrong, you don't need to hear it to lead a happy life.
Earlier we had a member post measurements at the ear position in a highly damped room, see: https://www.diyaudio.com/community/...eo-phantom-center.277519/page-31#post-4746283 (think: studio like and beyond). I have some measurements of the exact same thing (somewhere) in my thread as well. It's just physics. I used a foam ball as a head dummy with the microphone pulled trough. I think I remember Jim saying he used a pillow.
Most recordings are made to do just fine, but if you (still) notice higher pitched background singers it's there if the balance in the phantom center is right. A little lack of body for those side panned singers I'd call it, once phantom center balance is taken care of. Fix that tonal difference between center and sides and the background singers become "real people". It was a remarkable difference for me in my room. (huge panels behind each curtain and that poster also is a panel, pillows fill the space between couch and poster)
The effect is there if (the most strong early) reflections are diminished. Floor, ceiling, side wall, back wall and furniture. Especially the first 10 ms but every reflection within the Haas limit will still help work to hide it, but as the Haas limit often is said to be about 15 to 25 ms, it varies quite a bit in my own experience and even 15 ms can be perceived as echo once it's higher than 4.5 KHz and coming from a single side direction.
See my IR a few posts back for an example of "as measured at the listening spot" results, taken at 2.7m distance at my couch. I cleaned up the first ~20 ms and use a Haas kicker to get lateral reflections to keep a spatial effect. (see Griesinger's work on the perception of rooms and halls)
Lack of floor and ceiling reflections are free with arrays like fluid and I use. The rest (side wall, back wall) is done with absorbing panels. Furniture etc. hardly plays a role with a row of drivers line in an array. Only parallel planes need extra attention with speakers like these.
Interesting; for at least 35 years I thought the "center" speaker used with surround systems was just a joke; "must be because their L / R speakers are so incoherant" that such is needed. Now, (thanks to the careful sonic observers in this thread and elsewhere) it appears to be an actual thing, right in the middle of the speech bandwidth!!
I'm currently listening to a Polk SDA setup - of my own design of course. In mono, the SDA speaker in each L / R is projecting an out of phase signal to the opposite ear, where one would think the stereo phantom center diagram shown in post 1110 would be affected. My question is, would this be at least theoretically in a positive light, regarding the dip caused by the in phase arrivals as depicted?
Thanks,