Random Stuff About Surround Sound

[MidiMagic]

I don't exactly agree, while side images are not nearly as precise as front images they still exist even without turning your head, The effect is much better if the front and rear speakers are closer together. In a long deep room you will of course get the hole in the middle effect and hear the image from the closest speaker, or perhaps hear two distinct sources. Ambisonics might be able to remedy that but how is Dolby encoding or decoding going to change anything? I prefer having the rear speakers placed off to the sides rather than behind me, the rear left and right signals are then clearly heard to the sides and panned signals move more clearly front to back across the sides.

It's not in the encoding. The Dolby encoding is the same as QS as far as the recorded matrix signal is concerned. The same pan position in the room produces the same signal in both systems.

The Dolby surround and Dolby PL-I and PL-II decoders place a small delay in the surround channel(s). This makes the precedence effect work as long as your head is facing forward, and also works somewhat in other head orientation.

Look at a sound panned straight to the left. The left channel speaker puts out the major sound. But the surround channel also produces a delayed sound that steers the hearing system to locate the sound to the left rather than the left front.

Using the same recording I made, the panned sound moves slowly from left back to left front in Dolby Surround. With QS, it seems to suddenly jump from left back to left front at approximately the point where it is panned straight left in the recording.

 

[MidiMagic]

I've seen people turn their heads when they hear a sound that they believe is coming from a side location. I've done it myself. Why? Perhaps it's that a sound from the side is a distraction for which one wants to turn their head?

No. The imaging is not clear to human hearing unless you turn your head.

 

[chucky3042]

My how this thread has wandered!!

 

[Sonik Wiz]

My how this thread has wandered!!

Yes. It used to be straight ahead. But now it only stays on track if you have 17 speakers to brag about.

 

[chucky3042]

Yes. It used to be straight ahead. But now it only stays on track if you have 17 speakers to brag about.

Nah, let's do Atmos and bust the 20 speaker mark

 

[MidiMagic]

Why would you turn your head? Phantom imaging works perfectly fine in all directions and axis here including vertical. I've got demos and songs that go around in complete circles or rectangles around the room at different height levels even here with 17 speakers.

If phantom imaging didn't work in every direction, sounds wouldn't move smoothly around the room. With PS4 gaming (Standard multi-channel 5.1/7.1 output, not Atmos or anything), I can rotate my character 360 and a sound like a fire crackling moves around me in a smooth circle.

I almost jumped out my chair once when one of the characters in Dragon Age Inquisition suddenly yelled something right behind my left shoulder (that's where he was standing relative to my character) because it sounded so real (phantom image; there is no speaker where his voice came from).

In other words, turning your head is not a good indication of phantom imaging, but hard imaging.

The level-difference-panning used for 2-channel stereo works only if the pair of speakers are in front of your head or behind your head.

It does NOT work if the speakers are both on your left or both on your right. The sound "cogs" (suddenly jumps) to one speaker or the other. Your rotating demos probably rotate so fast that you do not notice the cogging. ...Unless you turn your head.

If you stick a speaker at LS, it works a lot better, but not smoothly.

I was quite disappointed when I heard 5.1 discrete and the first time I heard 7.1. In both cases, I heard bad cogging when a sound moved from back to front on one side. In 5.1, it sounded the same as QS. On 7.1, the sound cogged from the LB speaker to the LS speaker and then to the LF speaker.

This was a step backward from Dolby Surround, which eliminates the cogging.

 

[chucky3042]

Interesting about cogging. Really not trying to flog our stuff or a point but our SST really has zero cogging. Walk around, sway side to side, rotate, the image is always consistent

 

[MidiMagic]

Too much coughee.

 

[Sonik Wiz]

Interesting about coughing. Really not trying to flog our stuff or a point but our SST really has zero coughing. Walk around, sway side to side, rotate, the image is always consistent

Haha coughing vs cogging.

It's not in the encoding. The Dolby encoding is the same as QS as far as the recorded matrix signal is concerned.

You know very well that QS vs Dolby matrix encoding is not the same. Look at your own quad matrix pages. I know you hate modern surround audio but really, much improvement has gone on since your record changer & DIY decoder. Do try to encounter some more modern changes and you will be rewarded.

 

[chucky3042]

Coughing / cogging..... Fat fingers on phone with spell correct. Fixed it up! !

 

[MagnumX]

No. The imaging is not clear to human hearing unless you turn your head.

That is absolutely not normal sounding. Phantom imaging works in all directions so long as the total angular distance between any given speaker pair is less than about 90 degrees (60 or even less is preferable).

Sitting in the center of a traditional quad speaker arrangement puts each speaker at about 90 degrees from the next one. It should be able to place a phantom sound image anywhere within a 360 degree arc around the listener without moving your head.

I can't imagine turning my head constantly playing a PS4 game. I'd never be able to sit still.

 

[Soundfield]

That is absolutely not normal sounding. Phantom imaging works in all directions so long as the total angular distance between any given speaker pair is less than about 90 degrees (60 or even less is preferable).
Sitting in the center of a traditional quad speaker arrangement puts each speaker at about 90 degrees from the next one. It should be able to place a phantom sound image anywhere within a 360 degree arc around the listener without moving your head.

Being able to place signals electrically in those positions is easy. It does not mean that they are perceived that way. Human hearing is not symmetrical. It is optimised for location of sounds coming from the front. It is particularly poor at angular discrimination of sounds coming from the sides where there is little angular separation between the ears to establish phase differences and amplitude differences are skewed by the shadowing by the head of the opposite ear. You only have to look at how your ears are positioned on your head and their design to understand that human hearing is forward biased. All it can do is roughly tell you there is a sound coming somewhere from the side; if you want more angular / distance precision you have to reposition both sensors to obtain the most phase / amplitude information – i.e. turn your head to the source. To a certain degree the brain does this automatically by creating small movements of the head automatically that you are not aware of, but it is of limited effectiveness.

 

[MagnumX]

Being able to place signals electrically in those positions is easy. It does not mean that they are perceived that way. Human hearing is not symmetrical. It is optimised for location of sounds coming from the front. It is particularly poor at angular discrimination of sounds coming from the sides where there is little angular separation between the ears to establish phase differences and amplitude differences are skewed by the shadowing by the head of the opposite ear. You only have to look at how your ears are positioned on your head and their design to understand that human hearing is forward biased. All it can do is roughly tell you there is a sound coming somewhere from the side; if you want more angular / distance precision you have to reposition both sensors to obtain the most phase / amplitude information – i.e. turn your head to the source. To a certain degree the brain does this automatically by creating small movements of the head automatically that you are not aware of, but it is of limited effectiveness.

You can say or believe whatever you want, but what you two are saying about turning one's head or blurry imaging is absolutely not what I experience at all in the slightest. Imaging to my sides is sharp as a razor blade. It's BEHIND that is a bit harder to gauge the distance to something, but even then it's not even slightly blurry sounding or whatever you think you're describing. I've got a helicopter demo in Atmos that flies around the room, which I can have it play overhead or at ear level and it doesn't get blurry or ill defined as it goes around the room, but distance behind is harder to gauge compared to in front of me or to the sides. But the idea that the sides are somehow not sharp or distinct sounding is absolute nonsense. Perhaps you need to listen to some modern discrete surround recordings instead of QS/SQ records that have poorly defined surround channels?

 

[jaybird100]

You can say or believe whatever you want, but what you two are saying about turning one's head or blurry imaging is absolutely not what I experience at all in the slightest. Imaging to my sides is sharp as a razor blade. It's BEHIND that is a bit harder to gauge the distance to something, but even then it's not even slightly blurry sounding or whatever you think you're describing. I've got a helicopter demo in Atmos that flies around the room, which I can have it play overhead or at ear level and it doesn't get blurry or ill defined as it goes around the room, but distance behind is harder to gauge compared to in front of me or to the sides. But the idea that the sides are somehow not sharp or distinct sounding is absolute nonsense. Perhaps you need to listen to some modern discrete surround recordings instead of QS/SQ records that have poorly defined surround channels?

"Poorly defined surround channels" is more a function of the decoder being used, than inherent in the matrix system. The Surround Master is far better at rendering discrete-like rear channels than most other decoders out there, and while not everyone on the forum has one, it should be a goal. Yes, discrete is clearly the best, but had a decoder, with the specs of the SM, been out there in the heyday of quad (at a reasonable price), there might have been no need for a system like CD-4. The decision, on the part of the Warner Music Group back then, to adopt QS, instead of the CD-4 system that they ultimately went with, would have boosted the matrix camp considerably. The only reason they decided to go with CD-4 was pressure from Brad Miller, whose Mystic Moods albums were slated for release on WB. Miller threatened to pull his catalog from WB if they went with a matrix. He ended up pulling them anyway, to start Soundbird Records.

 

[par4ken]

The level-difference-panning used for 2-channel stereo works only if the pair of speakers are in front of your head or behind your head.

It does NOT work if the speakers are both on your left or both on your right. The sound "cogs" (suddenly jumps) to one speaker or the other. Your rotating demos probably rotate so fast that you do not notice the cogging.

If you stick a speaker at LS, it works a lot better, but not smoothly.

I was quite disappointed when I heard 5.1 discrete and the first time I heard 7.1. In both cases, I heard bad cogging when a sound moved from back to front on one side. In 5.1, it sounded the same as QS. On 7.1, the sound cogged from the LB speaker to the LS speaker and then to the LF speaker.

This was a step backward from Dolby Surround, which eliminates the cogging.

I partially agree with you about cogging but it can be greatly reduced by proper speaker placement. the attached is from the Lafayette SQ-W owners manual, and has long been my prefered speaker arrangement (long before seeing it in print). There are several causes for the effect that you describe. The first is "hole in the middle stereo" you get cogging even with stereo if the speakers are too far apart! Side imaging is not as good as front imaging so the effect is even worse to the sides. With the speaker arrangement shown I have no trouble hearing smooth pans. I admit that I may be unconsciously turning my head slightly with the pan, but so what it's easier than remaining in a rigid fixed position. Yes most Quad pans were done quickly so that the cogging is less noticeable.

 

[Sonik Wiz]

Being able to place signals electrically in those positions is easy. It does not mean that they are perceived that way. Human hearing is not symmetrical. It is optimised for location of sounds coming from the front. It is particularly poor at angular discrimination of sounds coming from the sides where there is little angular separation between the ears to establish phase differences and amplitude differences are skewed by the shadowing by the head of the opposite ear. You only have to look at how your ears are positioned on your head and their design to understand that human hearing is forward biased. All it can do is roughly tell you there is a sound coming somewhere from the side; if you want more angular / distance precision you have to reposition both sensors to obtain the most phase / amplitude information – i.e. turn your head to the source. To a certain degree the brain does this automatically by creating small movements of the head automatically that you are not aware of, but it is of limited effectiveness.

Ah, thank you for making my point better than I did!

 

[par4ken]

"Poorly defined surround channels" is more a function of the decoder being used, than inherent in the matrix system. The Surround Master is far better at rendering discrete-like rear channels than most other decoders out there, and while not everyone on the forum has one, it should be a goal. Yes, discrete is clearly the best, but had a decoder, with the specs of the SM, been out there in the heyday of quad (at a reasonable price), there might have been no need for a system like CD-4. The decision, on the part of the Warner Music Group back then, to adopt CD-4 instead of the QS system, as they had previously announced, would have boosted the matrix camp considerably.

The same goes for the Tate system, had it been available (and at reasonable cost) in the early days there would of been absolutely no need for CD-4. Same if the CD would of came out in the mid seventies and with discrete four channel ability right from the start (not just proposed) things might of been different.

 

[Soundfield]

You can say or believe whatever you want, but what you two are saying about turning one's head or blurry imaging is absolutely not what I experience at all in the slightest. Imaging to my sides is sharp as a razor blade. It's BEHIND that is a bit harder to gauge the distance to something, but even then it's not even slightly blurry sounding or whatever you think you're describing. I've got a helicopter demo in Atmos that flies around the room, which I can have it play overhead or at ear level and it doesn't get blurry or ill defined as it goes around the room, but distance behind is harder to gauge compared to in front of me or to the sides. But the idea that the sides are somehow not sharp or distinct sounding is absolute nonsense. Perhaps you need to listen to some modern discrete surround recordings instead of QS/SQ records that have poorly defined surround channels?

You are lucky to have super-human hearing. For the rest of us, a simple Google search on the acuity of human hearing will reveal vast amounts of research into the subject, all of which clearly identify the non-omnidirectional nature of our hearing.

The key points to note from all the published material are (1) that all adults are better at localising broadband noises (like your helicopter) than musical tones. (2) They are also better at localising sounds in the horizontal plane than in the vertical plane and better at localising sounds placed in front rather than behind or to the side of them (see Middlebrooks and Green, 1991).

The accuracy of localisation depends on the azimuth of the sound source and of its spectral characteristics. For a 100mS white noise burst typical accuracy figures often quoted are 3–4° for front sources (i.e. azimuth 0°), 5–6° for rear sources (i.e. azimuth 180°), and around 10°-15° for side sources (i.e. azimuths of 90° or 270°). These figures vary quite widely from individual to individual (some people may only achieve half these levels of accuracy).

The problems associated with the especially poor localisation of sounds from the sides arise from what is known as the cone of confusion. This is a surface extending laterally from each ear on which multiple sources will exhibit the same inter-aural time and level differences resulting in ambiguities which cannot be resolved

 

[MagnumX]

That diagram looks almost like a modern 5.1 side placement. You might get stronger side imaging, but likely at the cost of sound behind you.

Imaging gets weak to the sides or behind because in a traditional quad layout there's 90 degrees angular separation which is pushing it for some speakers and listeners as it's right at the limit for phantom imaging, which falls apart in the middle first. Reduce angular separation to 60 degrees (or even better yet 30 degrees) and you won't have those issues any longer.

You essentially need at least 6 speakers for strong 360 imaging (either surrounds behind with front like I use since my rear surrounds are way behind me or surrounds to the sides with rear surrounds somewhat behind. I think it's pretty much why Dolby went to PLIIz, which accommodates rear surround support.

You can achieve it without direct decoding support, however. Active mixing can add a +3dB set of channels to strengthen the phantom imaging, which is how I'm using front wides at the moment.

No, I don't have super human hearing. I just use a layout that actually works.

 

[par4ken]

No, I don't have super human hearing. I just use a layout that actually works.

So do I, (use a layout that actually works)!

To further illustrate that PWM side imaging can work just listen to stereo in a vehicle then adjust the fadder from front to rear. To my ear the sound moves down the sides and then to the back smoothly no cogging at all. Sadly many/most modern vehicles place the rear speakers low in the back doors to that when you fade to the rear the sound is a bit muffled. The effect was excellent in vintage vehicles with the rear speakers mounted on the rear deck. Mono sources image in the middle of the car at a bit higher elevation than the extreme left and right signals. Those Q8's with the full centre vocals sounded excellent. The big difference is that the speakers are closer together then they are in the home.

Modifying the speaker layout as shown in the Lafayette manual is one way to at least partially compensate, taking advantage of the ears greater sensitivity to frontal images. I still hear full 360° sound with that speaker placement. That placement is also the easiest to implement in the typical listening room.