Which speakers are the Classical Music Pros using?

[rdenney]

Gould discusses experimentation with technology and recording techniques in this interview:

Dammit! Now I have to go find that recording.

Rick “this forum is expensive!” Denney

 

[krabapple]

As can be seen in discussions in this and other audio forums, many people have moved on from pure stereo, enjoying the benefits of processed (upmixed) or real multichannel movies, music videos and music - always with the opportunity to revert to stereo. Those who by choice or circumstances remain in the stereo domain, still seek something that the two-channel format cannot deliver, and endless debates ensue.

Bravo!

 

[krabapple]

Pardon my ignorance about upmixing - is the high frequency energy in the signal sent to the surround speakers rolled off a bit to simulate the effects of air attenuation over a much longer reflection path?

That would make sense when surround channels are used only to provide the illusion of space/ambience. That's the case for most 'classical' and purist recordings. But for popular music multichannel mixes (and movie soundtracks), the surround channels are often for musical content....like actual instrumental or vocal parts... or for sound effects

FWIW I think, early Dolby Pro Logic upmixer implementations for consumer audio initially shelved the surround channel high frequency output...but I *think* it was made full range for later iterations? Or perhaps the difference was between modes (Movie, Music, Game) of DPLII? I forget!

 

[krabapple]

This is great for multichannel source material, but my experience with upmixed stereo (I haven't yet tried Auro-3D) is that a centre channel makes the centre image collapse to something much smaller than in stereo, and not in a very satisfying way. I suppose better up mixers don't do that... but it is almost as if the stereo mixing is done in the knowledge that they could make the centre panned material bigger than just the width of a centre speaker...but then the up mixer signs it all to just the centre speaker and it diminishes.

I have had to adjust the upmixing to spread some of the centre front 'assignment' back to FL and FR. That helps.

(These are just impressions from me and a few visitors over time.)

It really depends on the upmixer and what options pertain to it. For the Dolby Pro Logic II series, there was a Center Width setting that was user-adjustable across a range, with shared center content placed fully L/R at one end of the range, to fully center channel at the other. Its replacement , Dolby Surround Upmixer, has just two settings for what it dubs 'Center Spread' -- yet another downgrade from the previous algorithm :< IME default (off) is 'narrow', suitable for movie dialogue, wide (on) is for actual listening ;>

 

[krabapple]

Dolby upmixer definitely does what you describe, and I also hate it. You can mostly fix it by turning "Center Spread" off on the Dolby configuration, but not all AVRs allow that. Auro3D doesn't do that. At the strength I use(6-7) it more or less just heightens and widens the soundstage, and makes the center image a bit more stable.

Actually you fix it by turning the spread *on*. By default the center steering of DSU is narrow...I presume that's because Dolby expects the consumer to be using their AVRs mostly for home theater (movies/TV)

 

[krabapple]

JJ's system is surprisingly simple. What little has been said of it was high praise for how real it could sound. Playback is to assigned channels and recording microphones are specified as well. It was a 7.1 system using 9 recording microphones in a pattern slightly larger than someone's head. Two mics were for encoding up and down into the other channels thru a simple calculation that worked as it should. Or today i suppose you could add those two height channels in a 9.1 system. In which case no calculations or signal mixing would be needed, it would track straight channel to channel the way stereo does.

IIRC from his posts, JJ also has an infinite baffle in-wall subwoofer system .

 

[tuga]

That would make sense when surround channels are used only to provide the illusion of space/ambience. That's the case for most 'classical' and purist recordings. But for popular music multichannel mixes (and movie soundtracks), the surround channels are often for musical content....like actual instrumental or vocal parts... or for sound effects

How can an upmixer select and extract the "right" information for the surround channels from a 2-channel recording? (rhetorical question)

 

[Longshan]

More horns

View attachment 157859
View attachment 157860

I recognize the lower one, but what is that nickel-silver horn? And is it a triple horn?

 

[krabapple]

How can an upmixer select and extract the "right" information for the surround channels from a 2-channel recording? (rhetorical question)

Magnets, how do they work?

How can consumers set their systems up so that the reproduction is 'right'?

 

[Duke]

How can an upmixer select and extract the "right" information for the surround channels from a 2-channel recording? (rhetorical question)

I'm not sure "extracting" is the entirety of what's going on - I'm under the impression that what is extracted (de-correlated energy) is then processed rather extensively.

Sound On Sound interviewed David Griesinger of Lexicon about twenty years ago, and apparently there was a lot of applied psychoacoustics going into the algorithms he created. Maybe I'm just remembering the past through rose-colored hearing aids, but imo the Lexicon worked as advertised.

 

[Kvalsvoll]

How can an upmixer select and extract the "right" information for the surround channels from a 2-channel recording? (rhetorical question)

De-correlated, or sound with no direction - which will also in a 2-ch playback appear like sound that kind of fills the room, backwards. By extracting this information and reproduce it on speakers located in the back of the room, the effect can be much stronger.

How well sound at 180 degrees and further back is reproduced in a 2-ch system depends on room acoustics and speakers, and can be very different between systems. Some systems do this very well, but everything is kind of diffuse and floating. A good systems manages to maintain focus up front, while filling the room with reverb and reflection cues from the recording. This is easier to achieve with multichannel.

 

[tuga]

De-correlated, or sound with no direction - which will also in a 2-ch playback appear like sound that kind of fills the room, backwards. By extracting this information and reproduce it on speakers located in the back of the room, the effect can be much stronger.

How well sound at 180 degrees and further back is reproduced in a 2-ch system depends on room acoustics and speakers, and can be very different between systems. Some systems do this very well, but everything is kind of diffuse and floating. A good systems manages to maintain focus up front, while filling the room with reverb and reflection cues from the recording. This is easier to achieve with multichannel.

Could you elaborate on what you mean by "de-correlated, or sound with no direction"?
Does the upmixer system "recognise" what is direct sound information and ambience information and extract the latter?

I can use the fader in the car to increase envelopment, but the system is of poor quality and it's good only for background music, not for serious listening.

 

[Duke]

Could you elaborate on what you mean by "de-correlated, or sound with no direction"?

In this context, the de-correlated energy would be signal present in one channel but not identically present in the other.

By inverting the phase of one channel, and then summing the two channels, everything that is identical to both channels is cancelled (because it's exactly out-of-phase and at exactly the same level), so what's left over is the "difference signal", which includes (but is not limited to) desirable ambience information. At least that's my understanding.

One simple incarnation of this idea is sometimes called the "Hafler hookup" after Dynaquad inventor David Hafler, and apparently the idea appealed to Brian Eno. See this thread.

And then, from way out at the tail end of the bell curve, along came David Griesinger.

 

[tuga]

In this context, the de-correlated energy would be signal present in one channel but not identically present 1n the other.

By inverting the phase of one channel, and then summing the two channels, everything that is identical to both channels is cancelled (because it's exactly out-of-phase and at exactly the same level), so what's left over is the "difference signal", which includes (but is not limited to) desirable ambience information. At least that's my understanding.

One simple incarnation of this idea is sometimes called the "Hafler hookup" after Dynaquad inventor David Hafler, and apparently the idea appealed to Brian Eno. See this thread.

Thanks, that makes it a lot clearer. So the attack would probably be missing from the signal being fed to the ambience channels.
Griesinger mentions delay and band passing in the link you posted above, which makes sense too.

 

[Duke]

So the attack would probably be missing from the signal being fed to the ambience channels.

I don't know, unfortunately! I would guess that filtering signals which would tend to give false localization cues was probably part of what Griesinger did at Lexicon, and presumably part of what others have done since.

What Griesinger brought to the game was an amazing and perhaps unique understanding of and insight into the relevant psychoacoustics, AND the math & engineering skills to figure out and implement "HOW TO".

For instance, intuition might tell us that the lower the direct-to-reverberant sound ratio, the more "envelopment" we get. But not necessarily!! Griesinger again:

"Where the background stream [reverberation] is easily separated from the foreground stream [the direct sound], envelopment is about 6 dB stronger for a given direct-to-reverberant ratio."

Imo this implies that early reflections can actually degrade envelopment, and therefore we want a time gap before the strong onset of reflections.

 

[Floyd Toole]

Indeed David Griesinger brought a useful combination of DSP expertise, psychoacoustic knowledge, a sensitivity to the priorities of good music reproduction and intelligence. Somewhere I'm sure I have commented on LARES - Lexicon Acoustical Reverberation and Enhancement System. https://en.wikipedia.org/wiki/LARES. This was for the manipulation of large venue acoustics, making overly dead spaces more suitable for many types of music. It had many parametric adjustments. It used many microphones, connected through signal processing to many loudspeakers all distributed stategically throughout the venue. Numerous new venues were deliberately constructed to be on the "dead" side - good for pop concerts, dramas, etc, knowing that the reverberation attribute could be "dialed in" to whatever might be desired for operas and classical music of different genres.
At a point David decided to create a domestic version, which at the time was called LARES Light. We set up a system in the Harman demo room, in Northridge, CA basically a home theater, and proceeded to drop more than a few jaws. Perhaps the best demo I created for it was to begin with some well reproduced stereo music, then move to various levels of Logic 7 upmix enhancement, generating an improved sense of envelopment. Along the way, the music would be paused so I could add descriptive commentary and answer questions. Of course, everything, music and reverb ceased instantly. During one classical music segment I gradually faded in the LARES Light system. It had been set up to mimic the reverb in the recording, so its addition was not noticed by the listeners. However, this time when I hit the "pause" control, the reverb did not cease, but decayed in a relatively normal way. When anyone spoke they were speaking in the concert hall. Hand claps reverberated in an uncannily real way. It was impressive, and something most people thought would be impossible.
Other times I would have the system running when people entered the theater, and they were amazed at the large venue acoustics they experienced in this small deadish room. It was absolute fun! It was not location sensitive so the "hall effect" was appreciated by all in the room.
Employing only four small loudspeakers and four microphones it was not as good as it could have been, but it was enough to be very engaging. Frankly I wanted one at home. It didn't happen because the powers that be decided that there was not enough of a market for such a niche product. Pity.
So, the technology exist(ed) - maybe still exists somewhere; it isn't magic - to superimpose large venue acoustics on small room acoustics. It is now known that in such a contest humans gravitate to the larger perceptual space, fortunately. It was great fun to play with the parameters and change the apparent size and reverberation time of the space we were in, and then play stereo in it - it was remarkably transparent because all of the mics and loudspeakers were very neutral.
FYI

 

[GXAlan]

It was great fun to play with the parameters and change the apparent size and reverberation time of the space we were in, and then play stereo in it - it was remarkably transparent because all of the mics and loudspeakers were very neutral.

https://www.lexicon.com/on/demandware.static/-/Sites-masterCatalog_Harman/default/dw662f9201/pdfs/Lexicon_SL-1_System_Brochure.pdf

Were you at all involved with the Lexicon SL-1? It would seem like that sort of setup would let you have a lot of control over ambient experience by throwing more or less audio off axis?

Unfortunately, it looks like it's already been discontinued. (Hopefully that means there will be a 90% off promo at Harman at some point!)

 

[Floyd Toole]

I had no involvement with it. Once it is appreciated that resonances are to be avoided, the essence of a loudspeaker is in bandwidth, directivity and playing loud enough to be engaging. This was really a massive DSP exercise, involving Dr. Ulrich Horbach. It was a success in that it performed as intended and sounded very good. From the engineering perspective it was an interesting concept, well executed, and quite handsome.

However, it does not take a great mind to realize that it would be very expensive, addressing a tiny population of potential customers - dedicated two-channel stereo listeners with deep pockets and listening room real estate to spare. It was easy to imagine the owner on the floor, in the lotus position, deliberately located away from the boringly conventional symmetrical stereo seat, pressing the remote control button and suddenly being in the modified stereo seat. Cool! What price convenience? The price would equip a pretty decent multichannel audio system that would entertain several listeners. I could not get the thought out of my mind: "what happens if it malfunctions?", especially after a few years.

I have no knowledge of what happened to it as a product, but things have been very quiet, so perhaps you are right.

 

[richard12511]

Have you ever tried upmixing and then monoing the center (or other channels individually)? Can you hear artefacts?

Never tested turning them all except one off, but I have tested turning the front 3 off and putting my ear next to one of the surrounds for some time. All it really plays is echo noise, with the spl being tied to the strength you use. What would I be listening for?

 

[Duke]

Perhaps the best demo I created for it was to begin with some well reproduced stereo music, then move to various levels of Logic 7 upmix enhancement, generating an improved sense of envelopment. Along the way, the music would be paused so I could add descriptive commentary and answer questions. Of course, everything, music and reverb ceased instantly. During one classical music segment I gradually faded in the LARES Light system. It had been set up to mimic the reverb in the recording, so its addition was not noticed by the listeners. However, this time when I hit the "pause" control, the reverb did not cease, but decayed in a relatively normal way. When anyone spoke they were speaking in the concert hall. Hand claps reverberated in an uncannily real way. It was impressive, and something most people thought would be impossible.

Brilliant demo!!

I have a question about this statement: "[The Lares Light system] had been set up to mimic the reverb in the recording."

Could you get thoroughly enjoyable results with "set it and forget it", or did the system call for dialing in the settings for each recording?

How about for Auros 3D - set & forget, or trial & dial?

So, the technology exist(ed) - maybe still exists somewhere; it isn't magic - to superimpose large venue acoustics on small room acoustics. It is now known that in such a contest humans gravitate to the larger perceptual space, fortunately. [emphasis Duke's]

Is there a perceptual "point of inflection", past which the ear accepts the large venue acoustics "package" of the recording as being more credible than the small room acoustics "package" of the playback room? If so, what are some of the characteristics that tip the scale in favor of the former? (I have my amateur ideas about such, but would rather learn from you.)