Trinaural-type matrixing using transformers?

Design and use of the various types.
Post Reply
Brian Clark
Posts: 6
Joined: Sun Jan 30, 2005 8:35 am

Trinaural-type matrixing using transformers?

Post by Brian Clark »

Dave, do you think it a feasible idea to use 3 line-bridge type trannys, specially wound with appropriate dual primaries, to provide the 2-channels -to-3 blending along the lines of Jim Bongiorno's "Trinaural" processor which uses <place peg on nose> IC op-amps <remove peg from nose> :?:

The classic matrix is :

Left channel = 0.967L + 0.259R
Centre channel = 0.707L + 0.707R
Right channel = 0.967R = 0.259L

:)

Brian
Raj Gupta
Posts: 37
Joined: Fri Feb 11, 2005 2:19 am

Post by Raj Gupta »

Wouldn't you have to be listeninbg to stereo program material to do this? And isn't that pointless since all the really good music was recorded pre-stereo anyway?

:)

-j, who thinks that the Jamaican word "mon" is short for "monaural"
Attachments
ffrr.jpg
ffrr.jpg (2.46 KiB) Viewed 17697 times
vangogh.jpg
vangogh.jpg (3.77 KiB) Viewed 17697 times
Brian Clark
Posts: 6
Joined: Sun Jan 30, 2005 8:35 am

Post by Brian Clark »

Raj Gupta wrote:.... all the really good music was recorded pre-stereo anyway? :)
Good point Raj! :wink:

However I was thinking of early stereo recorded 3-track on Mercury, Command and so on. There was some decent stuff laid in vinyl back then :)


Brian (who is a mono mon at heart really)
[/img]
Raj Gupta
Posts: 37
Joined: Fri Feb 11, 2005 2:19 am

Post by Raj Gupta »

It all started to go to hell when they started mastering onto tape if you ask me . . . .

-j
Brian Clark
Posts: 6
Joined: Sun Jan 30, 2005 8:35 am

Post by Brian Clark »

You won't catch me using one of those pesky transistory hearing aid gizbobs. Give me mah HORN!

Okay, what were you saying j? :wink:
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: Trinaural-type matrixing using transformers?

Post by dave slagle »


Left channel = 0.967L + 0.259R
Centre channel = 0.707L + 0.707R
Right channel = 0.967R = 0.259L

Brian
do you have a better reference on this?

haow about scribbling up a schematic of what you propose so we are all on the same page.

dave
Raj Gupta
Posts: 37
Joined: Fri Feb 11, 2005 2:19 am

Post by Raj Gupta »

I think it would look like this.

Each pair of outputs would feed a summing bus which would have to be buffered somehow so as not to load or couple to the opposite channel.

-j
Attachments
trinaural.jpg
trinaural.jpg (25.87 KiB) Viewed 17672 times
Brian Clark
Posts: 6
Joined: Sun Jan 30, 2005 8:35 am

Post by Brian Clark »

Hmm. My proposal involves 3 transformers, each with two primary coils interconnected such that LHC had R out of phase with L, CC with L & R in phase and RHC with L out of phase with R. The number of turns of each primary coil set to produce the desired ratio.
All mounted in the same box as the system TVC.
Two jacks in, 3 jacks out.

Buffers are an abomination before the Lord! :wink:
If they are necessary it would invalidate the semper simplistica aspect of the proposal.

Curiosity sparked by reading about Jim Bongiorno's Spread Spectrum Technologies Inc Trinaural Procesor and enthusiastic reviews thereof.

Matrixing info gleaned from

http://mailman.io.com/pipermail/sound/2 ... 05213.html

Brian.
Lynn Olson
Posts: 16
Joined: Thu Jun 23, 2005 5:25 am
Location: Colorado, USA

Matrix Math

Post by Lynn Olson »

Let's get the math right first. The correct 3-channel decoder is:

Left channel = 0.967L - 0.259R
Centre channel = 0.707L + 0.707R
Right channel = 0.967R - 0.259L

The purpose of the anti-phase crossfeed into the L and R channels is to prevent excessive "monoing", or center-heaviness of the stereo image, which is what happens with the old Paul Klipsch matrix, dating back to the 1950's:

Left channel = 1L
Centre channel = 0.5L + 0.5R
Right channel = 1R

Quadraphonic research by Duane Cooper in the early 1970's showed that an antiphase signal (at levels between -20 and -10 dB) in the opposing speaker could "pull" the stereo image of the dominant speaker, moving it away from the center and pulling it more to the edges of the soundstage. This "wide-stage" effect of introducing antiphase crosstalk is routinely used as a special effect on studio recordings in order to move the image outside the L/R pair.

In practice, you'll probably want to adjust the center-speaker level for each recording, probably over a range of:

Maximum: 0.707L + 0.707R
Minimum: 0.1L + 0.1R

In the minimum case, all the center speaker is doing is improving the tonality of the center image by removing the comb-filter effect of 2-speaker stereo, which results in tonal alteration of solo singers and instruments. The comb-filter effect (look it up in the AES literature, along with Cooper's research) is caused by the delayed arrival of the more-distant speaker diffracting around the head when listening to central images - this is what also causes the image elevation of central sources in 2-speaker stereo.

It's also very very important that the center speaker have exactly the same phase response and polar pattern as the L and R speakers, otherwise there will a lot of listening fatigue. In practice, this means the same crossover, drivers, and driver layout as the L and R speakers. The dissimilarity of the Center speaker is why home theater system are so unmusical - and low-fi amps don't help either.
Lynn Olson
James Romeyn
Posts: 1
Joined: Wed Jul 24, 2013 3:19 pm

Re: Matrix Math

Post by James Romeyn »

Mr. Olson,
I have the utmost admiration for your contribution to audio and music enjoyment. I read all your audio journal postings at least once, some more than twice. It's a great privilege to see you post re. "3-channel" because I am enthusiastic and constant Trinaural user and supporter since 2007, since then having no love for the ubiquitous stereo format.

My Trinaural opinions and conclusions are based on hundreds of hours of experiments, possibly more than anyone short of my friend the deceased inventor (RIP). I am full time audio and music professional.

SST production including Trinaural Processor continues in the capable hands of young brilliant designer/Wyred4Sound principal "EJ." Current Trinaural MSRP is $2500 USD.
Lynn Olson wrote:Let's get the math right first. The correct 3-channel decoder is:

Left channel = 0.967L - 0.259R
Centre channel = 0.707L + 0.707R
Right channel = 0.967R - 0.259L
James was always secretive about Trinaural’s “algebraic� formula. Is there reason to believe Trinaural employs the formula above?

Polk's SDA (Stereo Dimensional Array within the speakers) and Carver's original Sonic Holography (line level) have no C Ch. Do they follow your math for the L/R channels only?
The purpose of the anti-phase crossfeed into the L and R channels is to prevent excessive "monoing", or center-heaviness of the stereo image, which is what happens with the old Paul Klipsch matrix, dating back to the 1950's:

Left channel = 1L
Centre channel = 0.5L + 0.5R
Right channel = 1R
Indeed, James repeatedly mentioned Trinaural C Ch is “algebraically processed� and not summed L+R mono, which he states does not work.
…In practice, you'll probably want to adjust the center-speaker level for each recording, probably over a range of:

Maximum: 0.707L + 0.707R
Minimum: 0.1L + 0.1R
Above really piques my interest. James stated that once the user has properly calibrated C Ch level, said level should be left alone and not adjusted.

My personal experience contradicts James, and rather confirms your suggestion.

To remotely trim C Ch and Bass levels I employ a particular Japanese mass market HT receiver with pure analog 7.1 line stage:

Stereo source analog output L/R > Trinaural 3.1 output (3.0 optional) > four inputs of remote control line stage > LCR main speaker amps + sub active xo/sub amp/four distributed subs (proprietary setup instructions posted for public use)

In personal A-B test the above 7.1 line preamp was indistinguishable from my $7500 SST Ambrosia line preamp.

Lastly, per Jim Smith's criticism, it appears that Trinaural benefits attenuating the C Ch several dB with spaced omni recordings ala Telarc.
In the minimum case, all the center speaker is doing is improving the tonality of the center image by removing the comb-filter effect of 2-speaker stereo, which results in tonal alteration of solo singers and instruments. The comb-filter effect (look it up in the AES literature, along with Cooper's research) is caused by the delayed arrival of the more-distant speaker diffracting around the head when listening to central images - this is what also causes the image elevation of central sources in 2-speaker stereo.
Above too is highly interesting.

Largely because of a unique radiation pattern, my speaker system, even in stereo, has as good or better center image and stage performance as I’ve heard doing this professionally and as a hobbyist for over forty years.

For comparison sake I often switch back from Trinaural to stereo.

Because overall system performance is so high, in stereo I quickly adjust to the lack of C Ch, almost immediately wondering if the C Ch is worth all the cost, effort, and space.

Even in “lowly� stereo my system performs not only admirably, but on par with the best cost no object systems auditioned over a lifetime of listening to the world’s best. (Hint: my “reference� is not stereo, but rather Ray Kimber’s best IsoMike playback system at CES, eight floor to ceiling Soundlab stat panels (two per corner), proprietary DSD 4.0 recordings, oodles of Pass Class A power amps.)

I performed the above test again just last night. (While listening in stereo I drape a thick comforter over the C Ch speaker, thus eliminating reflections that would otherwise minimize spatial effects.)

Then, immediately upon switching back to Trinaural, the texture and tonality of all the sounds between the speakers, especially soloists, is magnitudes better. One can't "know" the phenomena you describe above unless one hears it "fixed". It's like counterfeiting. You have to know the real currency to recognize the fake. One who has not heard a properly setup Trinaural system has no idea how wrong is stereo's "phantom" center.

I’d never willingly return to stereo for this reason. Over and over again my tests confirm exactly what you state above re. C Channel performance.

However good one might think is their system performance, I dare you to hear the above comparison and willingly choose the stereo format. You’ll pen volumes of a love once lost (for the stereo format).
It's also very very important that the center speaker have exactly the same phase response and polar pattern as the L and R speakers, otherwise there will a lot of listening fatigue. In practice, this means the same crossover, drivers, and driver layout as the L and R speakers.
I agree with the above.

James opens the door for users to employ C Ch speaker different from the matched L/R speakers, when he states that the C Ch speaker could be better quality than the L/R (but apparently never vice versa).

Fine points to “add to the mix� as my friend Nathan likes to say:

Circa 2007, just prior to purchasing my Trinaural, I was at VMPS loudspeakers. Per James’ advice above, we setup VMPS’ then largest-most costly planar speaker (about $5500ea) for the C Ch, and two least-costly planar model ($800ea) for the L/R. Performance was among the best Trinaural systems I’ve heard. It worked wonderfully on many levels, and there was no downside. (The speakers employed the same mid bass/planar mid/planar tweeter.)

This year I had the high pleasure of reading Jim Smith’s Trinaural opinions. Jim suspects Trinaural C Ch quality is inferior to its L/R quality. One imagines this could result from L/R content being subtractive while C Ch content is additive.

My LCR main speakers are matched high quality stand mounts. All three were at identical height and tilt back. Setup is exactly per Bongiorno’s diagram (I experimented a lot with placement and my favorite setup is per Bongiorno’s diagram).

Possibly because of the power of Smith’s suggestion, possibly because of his industry status, possibly because it was true, I seemed to notice a slight “dullness� and reticence in the C Ch vs. a more clear pristine quality in the L/R.

Then I seemed to recall that Bongiorno suggested slightly elevating the C Ch relative to the L/R. I re-read Trinaural instructions but found nothing about this. Possibly Borngiorno told me this in person, possibly another Trinaural user suggested it, possibly I imagined it.

I elevated the C Ch speaker 1-3/4in. On these 2-way stand mounts the mid bass is above the tweeter. It turned out best to maintain the same horizontal axis and the same position front to back.

The above change seemed to completely eliminate the slight but constant difference described earlier. Performance is greatly enhanced. The prior inconsistency is now replaced with a highly transparent and consistent performance across the front.
The dissimilarity of the Center speaker is why home theater system are so unmusical - and low-fi amps don't help either.
I could not more agree. The sum total reason for C Ch speakers whose architecture differs from L/R speakers, is to protect profit stream for flat panel display, and absolutely nothing else. This is a glaring audio industry failure.

An interesting Trinaural benefit: Trinaural coupled with perforated retractable screen makes for a dual use music/HT system that saves money, saves space, exceeds stereo format performance, and with proper speaker siting and selection also exceeds audio performance built around flat panel display.

HT specifies (especially Lucas’ vaunted “THX�) extremely “dry� acoustic with minimal radiation pattern for LCR front mains. Conversely, HT originally specified maximum diffusion for surround speakers (effective since TrueHD Dolby specifies mono pole speakers).

As you once wrote, dry front acoustic + high dispersion surrounds is antithetical to music as it exists in nature. Such antithesis leads to quick and extreme listener fatigue.
suqun1
Posts: 4
Joined: Thu Dec 26, 2013 8:48 am

Post by suqun1 »

You won't catch me using one of those pesky transistory hearing aid gizbobs.
Post Reply