Ambience extraction from stereo recordings

Robert Harley -- Mon, 06/09/2008 - 09:50

I received this e-mail from a reader, and thought I'd post it along with my reply.

Mr. Harley,

In the past I have written to you a couple of times with questions in regards to Stereo systems in which you were kind enough to answer. For some reason i have lost your other e-mail address but have located this one.

As you know back in the 1970s during the Quadraphonic rage Hafler had a simple 4 channel box called the Quadadaptor marketed by Dynaco. Around the same time Realistic/Radio Shack had brought out a similar item known as Quatravox. I believe it was the same principle as the Hafler Model.

Anyway back then I had one of the Realistic Quatravox units and had lots of fun with it as it extracted the ambience hidden in most 2 channel recordings. I have recently acquired one of these again brand new. Back then I had a low powered system consisting of a Receiver, turntable, and tape deck.

The neat thing about this Quatravox is there was no extra amplification needed for the rear channels and you only needed 4 speakers. It had 3 switches and 2 knobs. It had no power cord as it required no wall power.

Switch one was for straight 2 channel or adding the Quatravox.
Switch two was for Surround A or B
Switch three was Rear or Out
Knob one was for Blend
Knob two was for Rear Level

In essence I guess it really didn't alter the signal at all only extracting what was already there. Please correct me if I am wrong on this point.

My questions to you before hooking it up are

1. Would it be safe to use this unit with the 2 channel system I own today consisting of a Sony XA7ES cd player running direct through a Bryston 4BST 250 w/ch Amplifier. (I use no pre amp.) Speakers are ATC 19s

2. This Quatravox unit has RCA jacks for all connections including the Speakers. I use 12 gauge speaker wire now and would purchase quality RCA jacks and solder the speaker wire to them using the bare ends which would be connected as usual to the Amp and speakers and adding the RCA jacks to the opposite end for the connection to the Quatravox. Is there any danger using this setup with bare wire leads on one end and RCA on the other as long as everything is wired correctly and in phase.

An answer would be most appreciated before I go through all the trouble of purchasing and soldering the higher quality RCA jacks and most of all I do not want to damage my amp and or speakers.

I have read about surround systems in your Guide To High End Audio Book that i own but understandably there is not to much on Quadraphonics as it is pretty much a thing of the past. I am not interested in Home Theatre but i did enjoy the effects of this simple gadget years ago and for that reason and its simplicity it would be fun once again and good enough for me. I may even go as far as selling the ATC19s and purchasing a set of four smaller speakers such as the Totems or Proac models.

Thanks for your time.

Robert Harley -- Mon, 06/09/2008 - 09:57

The connection you describe is commonly called the "Hafler Hookup." It is a passive device that extracts the out-of-phase information in the stereo signal and routes it to a rear-channel amplifier and loudspeakers. The device has no effect on the front-channel signals.

I tried a home-made version in the early 80s briefly and went back to two-channel.

You can perform the extraction at line-level and then connect the rear-channel outputs to a power amplifier and speakers. The unit you have operates on speaker-level signals, but it's odd that it would use RCA jacks for speaker-level signals. There should be no danger in using the RCA jacks on one end and bare wire on the other.

This is a very simple and inexpensive method of adding multi-channel sound, although the effect varies considerably with the recording. It has the advantage, however, of leaving the front signals unaltered.

NAD developed a more sophisticated implementation of this idea in its AVRs that reportedly worked better than the straight "Hafler Hookup."

FJC54 -- Mon, 06/09/2008 - 10:52

Thanks Mr. Harley, for your answers to the questions I sent to you.

Oddly enough all of the Realistic Quatravox connections are made with RCA plugs. Two for the input from the amplifier and four for the output to the front and rear speakers. I believe this is possibly because a lot of the Realistic speakers especially the smaller ones back when the Quatravox was on the market had RCA jacks instead of screw connections which were the other most common connection back then. Some Realistic speaker models had both options.

As you stated this unit can be used as is without extra amplification. I was unaware however that it could be used with an extra amplifier driving the rear channels of the Quatravox as some of the more sophisticated Four Channel systems available back then could do . Maybe some reader will describe how to do that.

Also what I wanted to hear most from your reply is that the Quatravox does not alter the original front channels. Great News! I guess it gives you the best of both worlds. Pure 2 channel sound when wanted and added four channel ambience recovery by flicking the Quatravox in or out via the switch.

Maybe there will be some more reader discussion on this subject.
Thanks Again

Steven Stone -- Mon, 06/09/2008 - 11:05

As far as I know EVERY derived rear channel system is based on some variation on the Hafler matrix.

Some, like Meridian's Trifield, add a center channel to the mix. Many also add digitally-controlled delay to the rear channels.

If you eliminate front steering from the Logic-7 mode on a Lexicon, it too functions as a digitally controlled Hafler matrix.

The shortcomings of the orginal Hafler circuit are its inability to offer delay and roll-off of high frequencies for the rear channels. Both of these are needed to prevent excessive localization.

Steven Stone
Contributor to The Absolute Sound,, Vintage Guitar Magazine, and other fine publications

Tom Martin -- Mon, 06/09/2008 - 11:12

I had a Hafler setup in the '70s. I didn't like it, perhaps because the the delay issues that Steven mentions.

Meridian TriField is another matter entirely. For years I used the Meridian 861 controller and my default setting was TriField. The center channel is a major feature in my view. And, as usual, Meridian's implementation was sonically very clean.

CEO and Editorial Director, Nextscreen LLC

Robert Harley -- Mon, 06/09/2008 - 13:17

From Anthony H. Cordesman:

The Hafler "quad" circuit is actually a derived rear channel that drives the rear speakers with out of phase data. There are some old Stereo Review and Audio articles on how to make your own unit, but they are hard to find. The Wikepedia suggests making a connection from the right-positive post on the power-amp to the positive post on the right rear speaker. Then running a single wire from the negative post of that speaker to the negative post on the left rear speaker. And finally, running a single wire from the left rear's postive post back to the left-positive post on the power-amp.

The resulting sound creates an out of phase signal in the rear, without a clear sense of direction wi. The circuit does not somehow extract ambience, alter the timbre, or have any delay effect. All it does is add a sense of "space" with some recordings -- although the effect can sometimes be weird. It is a good idea to listen to a given recording with the front channels off to know exactly what is coming out in any given case. Performance is also very much a matter of set up, and room effects.

The user should be aware that this is a passive circuit, and effectively adds a new form of load to the amplifier and stereo speakers. This interaction rarely has a critical impact on sound, but is scarcely what high end gear is designed for, and can sometimes present problems.

It is not a good idea to use an old Halfer box designed for much lower wattages and far less sophisticated speakers. Anyone who experiments with this should research the various options, and then make their own box as a do it yourself project with connectors and wire gauges suited to a modern stereo system. The user should also know that speaker efficiency match will be a problem and it may be necessary to add a L-pad(s) into the circuit to match speaker levels to get the best effect.

Quite frankly, the various ambience circuits in a good AVR modern preamp or receiver will sound far better, although any synthetic rear channel effect will always be a "midfi" cludge. The Lexicon AV preamps contain the best options for creating such rear channel data that I have yet heard, but even these are not my thing. It takes a surround recording to get anything approaching high end surround sound. Barring that, I'd stick with stereo.

Anthony H. Cordesman

Robert Harley -- Mon, 06/09/2008 - 13:21

Meridian's Trifield, which creates a center channel signal from stereo, is highly beneficial. When I had a Meridian system, I used Trifield for reproducing stereo signals. It widens the sweet spot and solidifies imaging.

A visiting manufacturer---a two-channel purist who would dismiss anything other than two speakers---was listening to my system in Trifield without knowing it. He really liked the sound, and then I switched back to stereo without telling him what I was doing. His face fell, and said in disappointment "What happened?"

Steven Stone -- Mon, 06/09/2008 - 19:08

Gordon Holt used to regularly do the multi-channel to two-channel trick.

He'd begin the listening session with a two-channel devotee listening in derived multi-channel mode and then switch to two-channel after a couple of minutes of listening. The "What happened? " response was common. Most listeners preferred the derived multi-channel presentation to two channel mode on phase-coherent sources such as our own live concert recordings.

Steven Stone
Contributor to The Absolute Sound,, Vintage Guitar Magazine, and other fine publications

FJC54 -- Tue, 06/10/2008 - 09:05

Thanks, to all who posted replies to my questions in regards to using the Realistic Quatravox.

Upon reading the replies, especially the one posted by Mr. Cordesman stating that it is probably not a good idea to use this unit with a much more powerful amplifier and sophisticated speakers as I now have.

I certainly don't want to damage my current system and am not interested in re-wiring this unit or in building a new one.

When I originally had one of these units back in the 1970s it was used with a low powered receiver and four small low cost speakers. Rather than taking a chance of damaging my current amplifier and or speakers I have decided not to use it with my current system.

Although as I mentioned earlier it was a lot of fun but I doubt if I would go to the expense of adding one of todays processors which need extra amplification and more than four speakers.

I am not to familiar with todays units as I am not interested in Home Theatre in which I believe these newer setups are geared for. In addiition todays units are all designed electronically. The Quatravox didn't add or subtract anything to the original source, or setup.

Sometimes less is more!

Thanks Frank

Andy Simpson -- Mon, 09/22/2008 - 05:33

robert_harley6 wrote:
A visiting manufacturer---a two-channel purist who would dismiss anything other than two speakers---was listening to my system in Trifield without knowing it. He really liked the sound, and then I switched back to stereo without telling him what I was doing. His face fell, and said in disappointment "What happened?"

While I don't intend to question the merit of such a design, since both you & Steven mention these 'what happened' experiences I thought I would address that instead.

It is these 'what happened' / 'a blanket was removed from the speakers' experiences that give the audiophile a bad name.

The reason this experience is frowned upon is that it is so very easy to unconsciously or accidentally bias such a test with simple gain (SPL).

In both cases described above, was SPL maintained exactly?

For example, when people 'compare' stereo to surround, if stereo is biased by being 6dB louder than surround, I would expect it to win despite the expectation that it would lose in the face of the extra speakers at the rear.

In fact, I'd say that with a 10dB loudness advantage, even mono would have a good chance of beating stereo.


Steven Stone -- Mon, 09/22/2008 - 09:06

In my comparisons the levels were closely matched, but not claibrated. A 6 dB difference would have been more than a little bit noticeable. :wink:

Steven Stone
Contributor to The Absolute Sound,, Vintage Guitar Magazine, and other fine publications

Andy Simpson -- Wed, 09/24/2008 - 05:09

StevenStone1 wrote:In my comparisons the levels were closely matched, but not claibrated. A 6 dB difference would have been more than a little bit noticeable. :wink:

Indeed - but it gets a little more complicated when trying to match the SPL as, if I understand the device correctly, we have effectively only increased the SPL of the 'S' (of Mid/Side) without increasing the 'M'.

If the rear speakers simply added extra 'S', the 'S' gain would have been around 6dB?

A better comparison might be to use a M/S processor to simply increase the 'S' gain by the correct amount and then compare with that.

This is not to say that the idea does not have some merit.

If we are to increase the indirect-sound levels in a recording, without problems of 'loss of clarity' due to the extra indirect-sound masking the direct-sound, putting the extra sound in rear speakers would certainly make it easier to maintain the original masking relationship.


Mike J (not verified) -- Thu, 12/24/2009 - 21:40

I am nowhere near so sophisticted as the posters herein, so this query will likely engender great hilarity, but here goes: can a Hafler hook-up work with CDs, or is good-old-vinyl & a turntable needed to extract out-of-phase musical info??

Steven Stone -- Fri, 12/25/2009 - 01:55

 Actually the Hafler matrix works MUCH better with CD's.
LPs groove noise, and azimuth variations cause the Hafler Matrix (which works with the out-of-phase part of the audio signal) to go nuts sometimes. Especially records that aren't centered perfectly (many are not) will cause the circuit to "pump" and disrupt image stability.

Steven Stone
Contributor to The Absolute Sound,, Vintage Guitar Magazine, and other fine publications

SundayNiagara -- Fri, 12/25/2009 - 15:24

Get a Koetsu!

Steven Stone -- Fri, 12/25/2009 - 16:40

 And why, pray tell, would one brand of cartridge be more immune to the vagaries of miscut or mis-centered records than another?

Steven Stone
Contributor to The Absolute Sound,, Vintage Guitar Magazine, and other fine publications

SundayNiagara -- Fri, 12/25/2009 - 18:23

What does that have to do with the price of tea in Yugoslavia?

Steven Stone -- Fri, 12/25/2009 - 18:39

You Said, "Get a Koetsu".
I responded to your comment.
Now we are up to date.
Shall we proceed?

Steven Stone
Contributor to The Absolute Sound,, Vintage Guitar Magazine, and other fine publications

SundayNiagara -- Fri, 12/25/2009 - 18:53

Get a Koetsu, you won't need a Hafler-type device!

cryocynique -- Mon, 03/18/2013 - 04:52

Mr. Harley,

I appreciate the information you and others have posted here. I see that you personally have not been on the forum since last summer. I am hoping you are still active as I have a question specific to the use of the Realistic Quatravox, but with a twist.

Let me first say that I am a novice to many of these concepts, but at least understand them conceptually. Although my question involves audio signals from either a 2-channel system augmented with a Quatravox or similar device, or 4-channel system regardless of how the inner workings produce the "rear channel" signal, the ultimate output of those signals will not be used for my purposes as sound, but as light.

Ever since I was teenager in the 70s, I have always been fascinated with the concept of "seeing" sound through a color or light organ. That fascination has finally found a practical application as I have a side business as a karaoke host and have added that visual to my system. I acquired a matching set of 30" vintage 4-channel (color channels, not sound channels) color organs. Through my ignorance, I over drove them with too hot of an amplified signal. Fortunately, I have found and purchased 4-channel circuits to replace the damaged ones.

Since these new circuit boards will only handle a maximum audio signal up to 100 watts, I decided to purchase a couple of vintage low wattage stereo amplifiers to experiment with. I purchased a Realistic SA-10 and an RCA SA-155. My schedule has not permitted the repair of the color organs or experimenting with the amplifiers.

However, in the meantime, I have also acquired pairs of smaller 3-channel color organs that I am considering converting to 4-channel. That's when the idea struck me that if I used the 30" color organs as "front speakers" and the smaller 3-channel color organs next to the 30s as "rear speakers", would the frequency output of the "rear speaker" signals from a 4-channel audio system be sufficiently "different" from the "front speaker" signals to produce a different lighting effect even if I converted the 3-channel organs to 4-channel?

And as I shopped for vintage 4-channel receivers, I ran across the Quatravox, along with your discussion on this forum.

So my question is, since my ultimate goal is to essentially have 4 audio channels that each produce a unique or nearly unique set of frequencies to drive 4 color organs, am I better off purchasing something like the Panasonic RE-7740/7750, or would either the Realistic SA-10 or RCA SA-155 augmented with a Quatravox serve just as well for this purpose?

I look forward to your or anyone’s knowledgeable assistance in this matter.

All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.