5.1: gimmick, or closer to reality?

[The Byre]

Many 5.1 productions allow a constant stream of reverb, crowd noise, whatever, to irritate from the side-rear and that tends to let the side down. The Floyd remix of Pulse was like that.

Just as the best lighting effect is sometimes darkness - i.e. an absence of light - so a good stereo mix has mono in it and a good 5.1 has just stereo some of the time.

There seems to be a great temptation to put too much into a 5.1 mix and I find (just my opinion BTW!) that I get better results by keeping things simple and using those side channels only now and then.

[John Willett]

A well recorded 5.1 is wonderful.

Going from 5.1 to stereo the sound sounds horrible and flat - worse than the difference between stereo and mono.

Now - my tests were done on excellent classical recordings done by Mike Skeet, listened in his home with ATC SC 20 Pro fro front L & R and the rest ATC Pro 10s (the metal active ones).

He has even been doing experiments with side height information - *very* interesting.

But the 5.1 has to be properly recorded and not gimmicky - and having read lots of Rick and The Byre's comments on other forums think they will agree with me.

[The Byre]

Hi John! And yes, of course I agree.

I find what you say about height and positioning interesting , because we have been experimenting with reproduction in which every instrument was given its own speaker. This was for a six-piece string ensemble and the results were quite exciting. The speakers had to match the instrument, however, to get the best results, so the bass fiddle had a sub added, the rest were just little Genelec nearfields.

I would love to play more with some of the issues raised in this field, but the daily grind of recording anything and everything prevents this! Fully booked since July - damn!

(Anyway, I have to build a new workshop, so that I can fiddle with old cars and tinker with valve equipment, so 'needs must!')

[StudioTinPanAll]

well i love surround,
but i hate that center speaker,
so lets go for height and go 2.2.2
left, right, left high, right high, left rear, right rear
i heard it at galaxy studio about 2 years ago and i liked it

the next thing is mp3surround, and even surround on cellphones
so there will be some interest,
but the interest in audio quality is fading fast,
so if you don't put a lot in the rears, people think there's something wrong
with the dvd or their setup

grtz,

[matyas]

I'm really beginning to fall in love with good 5.1. I'm not set up to record in 5.1 right now, but I love listening to great surround releases, even on my modest home setup.

[Russ (Al) Prat]

Originally Posted by David Rick

In an unrelated thread, one of our esteemed Remotesters wrote

First, I want to know what Al means by "genuine", and what kind of 5.1 stuff he's been listening that he doesn't like.

"Genuine" for me means, genuine I guess. And it's not that I don't like 5.1, the fact is I'm trying to learn as much about it as I can, because I'm having to bring my ranching up to date, yippee yi yo ky yay, which is why this esteemed remotester is poking around these forums. I think I could honestly say that I haven't been overwhelmed by the 5.1 productions I've heard, like I was the first time I heard a good audiophile system. I haven't had the luxury of engineering 5.1 in a optimum listening area either, which I hope to do shortly. And I am also curious about the height, both front and rear, does that mean that 7.1 is more kickass (kickassier?) than 5.1? Is that a valid consideration or is 7.1 redundant? Would you say the audiophile crowd supports surround since the market has had time to mature?

[John Willett]

Originally Posted by StudioTinPanAll

well i love surround,
.....
so if you don't put a lot in the rears, people think there's something wrong
with the dvd or their setup

You don't need to put a lot in the rear - too much in the rear and I switch back to stereo.

But 5.1 with the right amount in the rear is wonderful even if you are not too aware of it - switching back to stereo soon shows it up.

[mohthom]

I still don't understand when people talk about problems with surround in classical music why they don't talk about Ambisonics . . . especially with height considerations!

MohThoM

[Roland]

Originally Posted by Russ (Al) Prat

First, I want to know what Al means by "genuine", and what kind of 5.1 stuff he's been listening that he doesn't like.

"Genuine" for me means, genuine I guess. And it's not that I don't like 5.1, the fact is I'm trying to learn as much about it as I can, because I'm having to bring my ranching up to date, yippee yi yo ky yay, which is why this esteemed remotester is poking around these forums. I think I could honestly say that I haven't been overwhelmed by the 5.1 productions I've heard, like I was the first time I heard a good audiophile system. I haven't had the luxury of engineering 5.1 in a optimum listening area either, which I hope to do shortly. And I am also curious about the height, both front and rear, does that mean that 7.1 is more kickass (kickassier?) than 5.1? Is that a valid consideration or is 7.1 redundant? Would you say the audiophile crowd supports surround since the market has had time to mature?

I think that it is fair to say 7.1 is generally more about coverage than the experience. My personal experience with 5.1 is that it gives much more of the "immersion experience" than a stereo recording. As for the production of "height" from a 5.1 or 7.1 recording, personally I'm skeptical. As John Willets was mentioning above, we have a mutual friend, Mike Skeet, who at the moment has been experimenting with periophony (in his case eight speakers). He reckons that the results are wonderful, however the implimentation of such systems is tricky and there are many practical problems.


Regards



Roland

[andy_simpson]

As I have said elsewhere, my problem with surround sound is that it is fundamentally flawed.

It is sad that so many people use their valuable research work to explore this impossible illusion. 10,000 speakers will not do much better, nor 1,000,000.

The major flaw in this concept - which permeates further than it might at first seem - is that in 'panning' from front-to-back, there is no sensation of an interpolated phantom image.

This is not the same as panning left-to-right.

Aside from this fundamental and completely fatal problem....

In my experience, those in pursuit of the surround concept have often never had the epiphany of hearing a good high-efficiency horn-loaded set of speakers. It's all in the time-domain response - spatial imaging, presence & 'life' not least.

5.1 of ANY direct-radiator monitor is far less realistic than a stereo pair of great horn-loaded monitors (K&H, Strauss, etc) - or even those horn-loaded monitors found in a THX cinema theatre.

....but, having said that, a 5.1 of K&H sounds great and I would agree that the overall effect is marginally greater immersion.

Andy

[brucie]

Originally Posted by mohthom

I still don't understand when people talk about problems with surround in classical music why they don't talk about Ambisonics . . . especially with height considerations!

MohThoM

Exactly!

I have some amazing classical recordings (choirs and string quartets) made with a Soundfield ST250, played back with 8 Genelecs (pantophonic configuration). It is a PITA to set up sometimes, but it does sound great. Mic placement can lead to some interesting immersive effects, e.g. in the middle of the choir or orchestra.

All the best

Neil

[Roland]

Originally Posted by andy_simpson

As I have said elsewhere, my problem with surround sound is that it is fundamentally flawed.

It is sad that so many people use their valuable research work to explore this impossible illusion. 10,000 speakers will not do much better, nor 1,000,000.

The major flaw in this concept - which permeates further than it might at first seem - is that in 'panning' from front-to-back, there is no sensation of an interpolated phantom image.

This is not the same as panning left-to-right.

Aside from this fundamental and completely fatal problem....

In my experience, those in pursuit of the surround concept have often never had the epiphany of hearing a good high-efficiency horn-loaded set of speakers. It's all in the time-domain response - spatial imaging, presence & 'life' not least.

5.1 of ANY direct-radiator monitor is far less realistic than a stereo pair of great horn-loaded monitors (K&H, Strauss, etc) - or even those horn-loaded monitors found in a THX cinema theatre.

....but, having said that, a 5.1 of K&H sounds great and I would agree that the overall effect is marginally greater immersion.

Andy

With the greatest respect Andy, I think thats a load of cobblers! Horn loudspeakers are far from accurate and there is no science that I can think of to suggest that they produce better time domain response than a direct radiating system. That you may enjoy listening too them is a personal view that I know not everyone will share. My personal experience with 5.1 is if done well switching to stereo removes the level of "immersion". Perhaps you would enjoy a 5.1 horn system?

Regards


Roland

[XLR]

A niche area, but for me Dabringhaus' demonstration of his 2+2+2(+2) at Vienna AES was really impressing. The organ recording was very realistic.
(PA used: Large Geithain Monitors :-)
But the listener has to build up another special setup...
http://http://www.2plus2plus2.info/

[d_fu]

Originally Posted by brucie

Mic placement can lead to some interesting immersive effects, e.g. in the middle of the choir or orchestra.

I really don't think that's where I'd want to be during e.g. Mahler's 8th, gettin' poked in the nose by a violin bow or being shoved around by choir singers...


I must admit I have not yet developed a particular interest in surround recording, but that may be due to lack of opportunity and demand...

[David Rick]

Though I don't share Andy's enthusiasm for horn-loaded speakers, I definitely agree with him on the following point:

Originally Posted by andy_simpson

As I have said elsewhere, my problem with surround sound is that it is fundamentally flawed.
<snip>
The major flaw in this concept - which permeates further than it might at first seem - is that in 'panning' from front-to-back, there is no sensation of an interpolated phantom image.

I agree. Consequently, I'm more than a little puzzled by a system like the Schoeps/Bruck KFM 360 in which, while the left/right cues are based on time and shadowing, the front/rear cues are generated by what's basically a pair of rotated M/S mics, i.e. the coincident microphone equivalent of a pan pot!

Similarly, though I've been quite an advocate of the Williams stereo array charts on this forum, to apply a time/intensity trading curve developed for the front listening quadrant to the sides and a widely spread rear seems completely unjustified. So I maybe it doesn't surprise me that my few experiments with Williams / Le Du five-microphone arrays have had pretty mixed results.

OTOH, do we even need accurate panning behind the listener? I don't really think so. If one's mix layout is "band front / hall rear" as is true of nearly all classical productions, who the heck cares exactly where various lateral sound reflections come from? The only important thing is that they're lateral.

Increasingly, I'm tending to think of location surround recording as two distinct problems: recording the musicians and recording the room. In recording the musicians, I care a lot about localization. In recording the room, I don't. What I do care about is keeping as much direct stage sound as possible out of the surround feeds, because it tends to mess up the front imaging.

Some pop mixers do the opposite. They'll put the lead vocal into Ls and Rs too. Comments?

David L. Rick
Seventh String Recording

[andy_simpson]

Originally Posted by Roland

With the greatest respect Andy, I think thats a load of cobblers! Horn loudspeakers are far from accurate and there is no science that I can think of to suggest that they produce better time domain response than a direct radiating system. That you may enjoy listening too them is a personal view that I know not everyone will share. My personal experience with 5.1 is if done well switching to stereo removes the level of "immersion". Perhaps you would enjoy a 5.1 horn system?

Regards

Roland

Cobblers?! It is nice to hear some native English after I've been in Poland for such a long time.

Yes, indeed, as I said above, the K&H 5.1 setup I heard was very nice, but as I also implied, a single pair of 0500s sounds a good deal more immersive than a 5.1 setup of direct-radiators. (Yes, the K&H range does count as horn-loaded, especially the 0500s - I spoke to the designer - 'waveguide' is simply a new marketing term for horn).

As to a scientific reason to suggest improved time-domain response of horn-loaded transducers, one need look no further than the concept of acoustic impedance.

(For now we can discount amplifiers as a source of time-domain error for the sake of simplification).

If we examine the sources of time-domain distortion in transducers we come to mechanical error relating to the physical operating principle.

If we consider this mechanical error in relation to acoustic impedance we see a relationship.

Error is proportional to impedance difference.

The horn raises acoustic impedance at the transducer, which reduces mechanical error.

Certainly, I would agree that design is critical but in simple terms the horn-loaded transducer has a very large advantage.

Andy

[andy_simpson]

Originally Posted by David Rick

.....

I agree. Consequently, I'm more than a little puzzled by a system like the Schoeps/Bruck KFM 360 in which, while the left/right cues are based on time and shadowing, the front/rear cues are generated by what's basically a pair of rotated M/S mics, i.e. the coincident microphone equivalent of a pan pot!

Exactly!

Similarly, though I've been quite an advocate of the Williams stereo array charts on this forum, to apply a time/intensity trading curve developed for the front listening quadrant to the sides and a widely spread rear seems completely unjustified. So I maybe it doesn't surprise me that my few experiments with Williams / Le Du five-microphone arrays have had pretty mixed results.

OTOH, do we even need accurate panning behind the listener? I don't really think so. If one's mix layout is "band front / hall rear" as is true of nearly all classical productions, who the heck cares exactly where various lateral sound reflections come from? The only important thing is that they're lateral.

Increasingly, I'm tending to think of location surround recording as two distinct problems: recording the musicians and recording the room. In recording the musicians, I care a lot about localization. In recording the room, I don't. What I do care about is keeping as much direct stage sound as possible out of the surround feeds, because it tends to mess up the front imaging.

Some pop mixers do the opposite. They'll put the lead vocal into Ls and Rs too. Comments?

David L. Rick
Seventh String Recording

I had an interesting experience in a cinema theatre recently. The lead actors were having a converstation and I was observing the phantom sound imaging with some confusion.

Usually (or perhaps often) the main dialogue in films is centered (or even mono) in the stereo mix and localises at almost the same distance as the screen.

On this occasion I was very puzzled by the localisation cues, since the sound didn't appear to be exactly localised at the screen distance as usual.

After a moment or two I noticed that the balance engineer for this particular film had sent the dialogue to the side/rear speakers almost equally, which gave a sense of a vaguely non-localised sound source .... very strange.... which presumably came from a confusion of differing cues from the many speakers in the auditorium.

Andy

[Yannick]

5.1 is definately no gimmick.

At the very least, switching on the rears gives:
a. more immersion
b. more information about the original hall
c. more precision in the L-R speakers, as if the resolution is heightened - the same stereo mix, with the rears switched on, sounds more transparant !
d. more front-back localisation in the front speakers.

The centre speaker HAS to be more realistic than just stereo :
a. we NEVER use the centre speaker, unless for soloists.
b. putting the soloist (eg. soprano) gives you 1 amp - 1 speaker to deal with this ordeal of reproducing this signal.
c. you eliminate the severe coloration of this important mono source, due to time delays and shadowing around your head in normal (phantom) mono on stereo replay.
d. I repeat : the biggest flaw of stereo is the mono component. Ever listened to a good mono recording ? On just one loudspeaker ? (not on two as all those audiophiles do ...)

I do agree that panning from front to back does not work (well). We tend to use a mix between coincident and delay based surround.

[just.sounds]

My experience is that a 5.1 setup can have some advantages but they are limited IMHO.
As i want to record with omnidirectional microphones because of the sound they give me i have problems using the center speaker. If i want to create a nice phantom image the center speaker does always more harm than good. Because the inability of a nice phantom image on the side with the front and rear speakers the rear speakers serve only one purpose for me. The diffuse part of the reverb. As this part of the sound has no "direction" in real life the ear seems to ignore the fact it comes from a little box instead of all sides.

[edit]But with pop it can be different I love the surround mix from The Beatles: love[/edit]

[andy_simpson]

Originally Posted by just.sounds

....
If i want to create a nice phantom image the center speaker does always more harm than good.

Can you elaborate on this a little? Then maybe we can discuss the possible reasons why?

Because the inability of a nice phantom image on the side with the front and rear speakers the rear speakers serve only one purpose for me. The diffuse part of the reverb. As this part of the sound has no "direction" in real life the ear seems to ignore the fact it comes from a little box instead of all sides.

Which I suppose brings us back to the immersion effect.

I look forward to the day when research catches up and we are able to place phantom source objects in a real 3D field but we are a long way from there.

Andy

[Westmalle]

Originally Posted by XLR

A niche area, but for me Dabringhaus' demonstration of his 2+2+2(+2) at Vienna AES was really impressing. The organ recording was very realistic.
(PA used: Large Geithain Monitors :-)
But the listener has to build up another special setup...
http://http://www.2plus2plus2.info/

I have CD-recordings of music of Vierne and Guilmant of Ben Van Oosten made on the Cavaillé Coll organ in Rouen, made by Dabringhaus und Grimm, and it sounds absolutely fantastic.

[just.sounds]

Originally Posted by andy_simpson

Can you elaborate on this a little? Then maybe we can discuss the possible reasons why?
Andy

When i use omni microphones in a A-B-C (three in a row) i hear three phantom images.
one between L-R one between L-C and one between C-R To avoid this i should make better channel separation. but when i put the microphones wider the hole in the middle of A-B setups hits me THREE times. The way to make better channel separation would include directional microphones witch is for acoustic recording a compromise IMHO.
So that is why i stay with stereo for the front for now. does someone have other experiences??

[just.sounds]

Originally Posted by andy_simpson

I look forward to the day when research catches up and we are able to place phantom source objects in a real 3D field but we are a long way from there.

Andy

Not as far as you think :-)
Wave field synthesis - Wikipedia, the free encyclopedia

I'm researching on an 192 speaker system. The main problem of the system is the coloration between speakers because of the size of the speakers itself. But progress is made

[Yannick]

You are right.
Even with coincident techniques.

The C speaker is only useful for soloist mics in classical recordings IMO

[d_fu]

Originally Posted by Yannick

The C speaker is only useful for soloist mics in classical recordings IMO

Would you send a soloist's spot mic to C exclusively, thus panning the voice (or whatever) dead center? Would this not collide with the soloist's perceived position in the main mic if he/she/it is actually off-center, next to the conductor?

Daniel

[just.sounds]

Originally Posted by Yannick

5.1 is definately no gimmick.
:
a. we NEVER use the centre speaker, unless for soloists.
b. putting the soloist (eg. soprano) gives you 1 amp - 1 speaker to deal with this ordeal of reproducing this signal.

Interesting technique! How do you deal with crosstalk?? Is it a problem? I would guess if you use delay to integrate the support with the mains it would be.

Originally Posted by Yannick

I do agree that panning from front to back does not work (well). We tend to use a mix between coincident and delay based surround.

I like to make a convolution of the hall and use that to process the back channels.
if you make a fade-in on the convolution sample you can cut the direct sound and early reflections off.

[Yannick]

Convolution is also the keyword, especially if the hall is great !

You can do so many tricks, it gets scary. Manipulating the early reflections can be a very powerful tool indeed !

We try to get the solo in the middle of the orchestra, to avoid imaging problems.
A great position for a voice can be right in front of the winds.

Anyway, if the are slightly off center, no problem, because most of the time the spot is SO loud ...

We use stereo (blumlein) spots, narrowed down, but this gives us the possibility to shift the voice slightly. Of course, the M signal goes to the C speaker.
We add convolution reverb to the LR and LsRs speakers, the spot gets treated separately (more predelays).

[seriousfun]

stereo - from the Greek sterios, meaning solid. Nothing about 2.

Every sound we (even those among us with hearing in only one ear) hear is from a 3-dimensional space (unless we are in an anechoic chamber). An acquisition/delivery format must be adequate for a 3-dimensional space.

2-channel stereo can do this with some material if it is recorded and played correctly. Playback is critical, and the phantom image can only be portrayed if the listener is positioned dead-center between the two speakers. Even then, the phantom image has time and frequency issues that a single speaker will not have (look up Head Related Transfer Function as a start).

Making stereo from LCR instead of LR gives a more-accurate representation of any front soundfield. And there are now three phantom images instead of one (L-C and C-R being less solid than LR, but they are there).

5-channel stereo gives us ten phantom images! The L-Ls, C-Rs, etc., are perceived differently by our ear/brain mechanism than the L-R, but they are there. 5-ch also adds the ability to have smooth dynamic pans from left to right, and diffuse or discrete elements from the rear. These are all great tools, and fewer channels can be used when appropriate (C-only is the ultimate way to present mono recordings, for example).

5.1 (the .1 LFE channel has no bearing on stereophonic sound, and probably should never be used in a music-only mix anyway) was developed as the minimum configuration to deliver a convincing 3-dimensional image. Our ears do not perceive the phantom image from the right or left (made by L-Ls and R-Rs), so 7.1 systems with discrete speakers to fill this gap can help; our ears do not perceive the Cs phantom made by Ls-Rs well, so 6.1 systems with a discrete Cs can be used. No current widely-used system, from 1.0 to 7.1, can deliver much of a height-image, so height channel(s) can be used. 10.2 as developed by Tomlinson Holman have all of these, so this might be the logical next step. Where we stop between 2.0 and ~.x is anyone's guess.

[mohthom]

I look forward to the day when research catches up and we are able to place phantom source objects in a real 3D field but we are a long way from there.

Ambisonics . . . . . .

Please see Computer Music Journal, Vol 19 no. IV, esp p. 64 to the bottom of 65.

MohThoM