What is Stereo Recording Angle?

[Tomás Mulcahy]

Had an interesting interaction with a person from a well known developer of spatial audio plugins. He claims that because speakers are set at -30° and +30° there is only 60° "of liberty in stereo sound" (sic). He disagrees that the image is more like 180° despite sending him this link:
http://www.sengpielaudio.com/HejiaE.htm

The fact that XY90 gives an SRA of 195° really threw him.

But it is generally 180° (depending on mic array etc.). Right? Since loudspeakers are basically cardioid, and we're assuming an acoustically treated room. Right? I'm having a moment of doubt!

[Hugh Robjohns]

There is only 60 degrees angle of possible image movement/positioning when listening over stereo loudspeakers. The whole audio environment has to be fitted into that +/-30 degree angle. He's right.

The SRA only applies to stereo mic arrays and only denotes how the space in front of a stereo microphone array is mapped into that 60 degree angle in front of our ears.

So if he's developing processing plugins, he does only have +/-30 degrees of area to play with for a loudspeaker listener.

[Tomás Mulcahy]

Eek. I've totally misunderstood the concept. So what is the point of, say, XY90 which gives 195° if the speakers cannot reproduce the whole image? How is it mapped in to the narrower space?

Are headphones better/ wider?

[Drew Stephenson]

How does that 'wider than the speakers' effect work then Hugh? Not challenging you understand, just curious!

[Hugh Robjohns]

Eek. I've totally misunderstood the concept.

When I first started as a lecturer I was told that to teach someone 10% of a subject, you need to know 90% of it... And I soon found out how true that is!

So what is the point of, say, XY90 which gives 195° if the speakers cannot reproduce the whole image?

They can reproduce it... Just not with identical relative positions. It's about creating a representative impression of the real acoustic world for a pleasing effect... If you want real world imaging accuracy you need to step up to a 'surround sound' setup.

How is it mapped in to the narrower space?

It's much the same as taking a photo. You could use a 16mm wide angle lens on the camera to capture an image of the whole space, or a 200mm lens to capture an image of a small specific area... But the printed output will end up on the same 5x4 paper whichever lens you use.

Are headphones better/ wider?

Not really... Different angular/spatial distortions, but still distortions...

[John Stafford]

It strikes me as odd when I hear recordings that sound like nobody considered any of this stuff, e.g. some well-known concerto recordings that sound like the piano is wider than the orchestra.

[Hugh Robjohns]

How does that 'wider than the speakers' effect work then Hugh? Not challenging you understand, just curious!

The 'outside the speakers' effect uses intentional opposite-polarity signals in the two channels -- a situation which can't ever happen in real life.

The phase anomalies from the direct and reglected sound from both speakers create peaks and dips in the frequency response which can be similar to those introduced naturally by the curves of the pinnae.

The brain is a bit confused with the conflicting information and so tries its best to make sense of it by creating an image location somewhere outside the speakers... Could be left or right, or it could be up or down... And to make matters worse, different people hear the 'phantom sound' in different places because the dips/peaks correspond to different locations depending on the shapes of different ears -- and we all have different pinnae!

[Drew Stephenson]

Thank you!

[Tomás Mulcahy]

It's much the same as taking a photo.

OK that I understand. I'm struggling to find any diagrams of it. Or at least, ones that I understand. I've been wading through Michael Williams's books. Maybe I don't understand the diagrams there!

I didn't think the angular distortion would be more than 5 degrees on a good system?

I hear you on the lecturing

[Hugh Robjohns]

I'll try and dig out some pics (or draw some) tomorrow.

But another way to think about it is this: if you have an array like XY fig8s, say, which has an SRA of 90 degrees or so, a single sound source in front of the stereo mic array has to move just 45 degrees to one side or other to move fully across the sound stage from the speakers to reach one speaker or the other.

But if you had XY cardioids, with an SRA or 180 degrees or more, the same source would have to move around 90degrees to reach the speaker!

In a recording situation you have a sound source (orchestra, ensemble, soloist, etc) that occupies a certain physical width, and you'll want that to occupy a certain proportion of the sound stage between the speakers, possibly with some room acoustic extending to the sides. So that's the end goal to keep in mind.

You'll also want a certain direct/ambience/reverb balance which obviously depends on the distance between mic and source, combined with the polar pattern.

So once you know where the mic needs to be, you'll also know the SRA you need to create the sound stage you desire. There are a number of different stereo mic arrays that can create the desired SRA, with different polar patterns, mutual angles, and capsule spacings... So you pick one with the characteristics you want that works with the right angles.

It's really no different, conceptually, from standing a set distance from a subject with a camera, and choosing a lens (or adjusting a zoom lens) to frame the subject in the way you want.

[Hugh Robjohns]

I didn't think the angular distortion would be more than 5 degrees on a good system?

It can get up to about 8-10 degrees in some cases, but with most stereo arrays it is under 5 degrees most of the time.

But the term 'angular distortion' in this context is about a departure from linearity -- the equivalent of pin-cushion or barrel distortion in a camera lens.

So consider five sources equally spaced in an arc in front of an XY cardioid stereo array. So one is at -90, one at -45, one at 0, one at +45, and one at +90 degrees.

They should be portrayed from the speakers at -30, -15, 0, +15, +30 degrees so that their relative positions are in the same proportions.

But all stereo mic arrays have some small non-linearities -- angular distortions -- so in practice you might find the perceived angles are -30, -20, 0, +20, +30 degrees. In most cases, sources near the edges are pushed wider, so tend to bunch near the edges, while those nearer the center tend to pull wider apart than they should be. It's a pretty subtle effect, though, and generally goes unnoticed.

[Tomás Mulcahy]

I'll try and dig out some pics (or draw some) tomorrow.

Fantastic, thank you. However, your explanation there makes it a lot clearer thank you!

[Hugh Robjohns]

[Tomás Mulcahy]

Hmm maybe I do need a diagram . I can kind of see how headphones give a genuine sides signal but not so much an impression of things being in front. Whereas speakers can do in front, but not sides. I think I can imagine the shape of the angular distortion in each. But why can't I find visualisations of stereo or surround speaker array dispersion patterns?

5.1 cannot do sides. 7.1 can. But do both have a "hole" in the middle of the rear?

And why do a lot of vectorscopes default to a semicircle display? Obviously a kind of compormise but I never really thought about it before. It makes more sense (to me) than, say, the diamond shape.

[Guest]

Very interesting. Would love to see the diagrams.

[Hugh Robjohns]

I can kind of see how headphones give a genuine sides signal but not so much an impression of things being in front. Whereas speakers can do in front, but not sides.

I think perhaps you're overthinking this and confusing yourself with possibly misunderstood terminology and unhelpful format comparisons!

Conventional stereo was invented for loudspeakers. Full stop. Forget headphones -- that's a completely different experience. (And, while you can play two-channel audio material over both formats, the experiences are inherently very different. To get the best from each format, you need audio material recorded specifically for each format.)

So, back to speakers: the stereo speaker format creates a stereo image stretching between the speakers. That's your 'window' into the recorded world, in the same way that the TV screen is a 'window' into telly-land.

Stereo speakers are not designed to provide an immersive experience. A two-dimensional sound stage is laid out in front of you (width and depth). If you want an immersive experience of sound all around you have to go to one of the surround formats (each with their own compromises and benefits).

The Stereo Recording Angle of a stereo mic array simply relates how far to the left and right of the mic array you need to go to create the impression of moving from the centre to the edge of your stereo speaker image. The SRA is chosen to suit the spread of sound sources around the mic array, and to create the stereo image width wanted from the speakers.

...why can't I find visualisations of stereo or surround speaker array dispersion patterns?

If I knew what you were talking about I might be able to help... but I don't.

Most manufacturers normally mention 'dispersion' in the context o a speaker's polar pattern -- showing how the different frequencies are radiated in different directions. Most monitor speakers try and provide a uniform spread of sound at all frequencies over about 120 degrees or so... but most also tend towards being omni-directional at low frequencies just because of the physical size o the bass driver and cabinet.

But the idea is to disperse sound into the room as spectrally uniform as possible. This has nothing to do with the stereo imaging, per se -- that's predominately down to the on-axis sound being thrown at your ears.

I think what you might be looking for is a map of how the sound image locations can be varied depending on placement and number of speakers. I've seen such things for quad formats and 5.1 I think... I'll try and dig some out next week.

5.1 cannot do sides

.

It can to a degree, but the imaging accuracy is poor (just as it was with quad). But to be fair, the consumer 5.1 format is a ridiculous bodge of misunderstanding and should never have been entertained!

7.1 can.

Yes, it is less bodged...

But do both have a "hole" in the middle of the rear?

Not really... stereo speakers don't have a "hole" in the middle (unless the recording technique is crap)... but for large spaces (ie cinemas) an extra centre rear speaker is available in some formats for better defined imaging for off-axis listeners.

And why do a lot of vectorscopes default to a semicircle display?

It's just an alternative way of displaying the stereo information that some people find easier to comprehend.

It makes more sense (to me) than, say, the diamond shape.

Fair enough...

Some 'vectorscopes' (eg, The Box) show only the top quadrant of the 'diamond' display, and that portion is what is effectively being stretched out into the semi-circle display.

The 'diamond' display is derived from the Lissajous display shown on scopes back in the day, and it effectively shows the positive and negative portions of the waveform in separate areas, so it actually conveys more information once you know how to interpret it -- such as waveform asymmetry, DC offsets and so forth.

[ManFromGlass]

Great stuff! So theoretically in order to get an immersive audio experience that is as close to real life as possible one would have to be suspended in the centre of a large ball made of speakers?

[Hugh Robjohns]

Yes.

But it's not a very popular format with interior decorators or domestic managers...

Speaker stereo is an illusion. It relies on the brain making things up, and it's not dissimilar in that regard to the trickery of showing 24 still images each second while we perceive natural movement (mostly) across the cinema or tv screen.

The 60 degree angle of the speakers generates false time of arrival differences at the ears based on interchannel level differences at the speakers, but the fakery is enough to create believable, stable imaging information across that limited angle in front of us.

The same imaging trickery doesn't work for speaker's at the sides, and it's not as accurate for speakers behind either. So for stable and believable surround imaging you really need to create more natural time of arrival signals at the ears, and that typically requires a lot more speakers in a lot more places. Hence Dolby Atmos and all the other large multichannel formats.

5.1, 7.1 etc were designed for the cinema where the rear channel signals were duplicated across a lot of speaker's with the idea of creating a sense of envelopment rather than accurate rear imaging. Entirely different intention and outcome.

[ManFromGlass]

I’m not even allowed to connect a stereo to the tv so I get that!
But what an ideal man cave a speaker ball would make! but then the tv screen would probably cause weird reflections. The physics of things can be so inconvenient at times!

[Drew Stephenson]

Physics, ruining things since the 4th century BC.

[Hugh Robjohns]

[James Perrett]

Great stuff! So theoretically in order to get an immersive audio experience that is as close to real life as possible one would have to be suspended in the centre of a large ball made of speakers?

Which is the idea behind higher order Ambisonics as I understand it. A couple of years ago I went to a demo of it at the ISVR in Southampton. They have created a 32 channel ball of speakers with a microphone array to go with it and the results, from a spatial audio point of view, are impressive.

[The Elf]

I’m not even allowed to connect a stereo to the tv so I get that!

£$%^& that!

Once Coronation street had been heard in (at first) stereo, then (later) in surround there were no complaints!

[ManFromGlass]

It’s almost like being there . . . . . . . . . .
Aaaaaaaaaaaah, run away, run away.
Our neighbour does Cory St. binge watching on a Saturday or Sunday morning. I think she records a wack of them. One is not allowed to talk to or disturb her. Do so at your peril!

[Dynamic Mike]

I sort of get how the brain evaluates sounds heard in the left & right ears and calculates where the source is in a lateral plane.

Stereoscopic vision is similar, although it can be thought of as though one dominant eye looks directly at an object & the brain calculates the distance from the angle other eye would have to deviate to see it (it only actually deviates at very short distances but the brain knows whether it's off-centre on the retina.) But the brain also knows when we're looking up or down so can estimate a 3D point in space where the object is likely to be. Biologically speaking that's grossly over-simplifying steropsis, but the principle is fairly sound.

But back to sound, do we have height perception of a sound source & if so how given we don't have a top ear & a bottom ear? (Although I do seem to recall Captain Kirk having a third ear. )