Upmixing - where are we at? Have people compared upmixers?

[sfogg]

My experience with L7 is that it is outstanding at adding ambiance to the surround speakers, giving a "live concert-hall"
type experience. But over the years since I've found I preferred a number of the more modern upmixers that I found better
at emulating a real discrete multich source.
YMMV

YMMV but....

Mode Adjust -> L7 Music (or even 'L7 Film', not 'L7 Music Surround'... that one is much more subtle and built just for ambiance extraction, not discrete steering)

Soundstage -> Rear (defaults to neutral, this change makes the surround balance even with the fronts as it removes -3db attenuation of the surround/rear channel that is built into the default neutral position)

Rear Delay Offset -> 0ms (defaults to 15ms, this is beyond the time alignment of all speakers. This is for the Haas precedence effect. Setting this to 0 makes the surround activity more prominent, increasing it changes the sense of depth of the acoustic space)

Surr Rolloff -> Off (Defaults to 7kHz and is applied to the extra material steered by L7. Turning this off makes the L7 steered material full range which mirrors discrete MC source)

Front Steering -> This really depends on what you want for your center. I want a center with a well defined soundstage. For me Film (on the MC-4/8/12) works great. On earlier Lexicon's (DC1/2, MC1) the Film setting wasn't as stable with music and it tended to narrow the soundstage a little on some sources. The 'Music' setting is a little bit less aggressive steering and "Music Surround' uses minimal steering and is much more like a passive fill.

 

[Sal1950]

YMMV but....

Mode Adjust -> L7 Music (or even 'L7 Film', not 'L7 Music Surround'... that one is much more subtle and built just for ambiance extraction, not discrete steering)

Soundstage -> Rear (defaults to neutral, this change makes the surround balance even with the fronts as it removes -3db attenuation of the surround/rear channel that is built into the default neutral position)

Rear Delay Offset -> 0ms (defaults to 15ms, this is beyond the time alignment of all speakers. This is for the Haas precedence effect. Setting this to 0 makes the surround activity more prominent, increasing it changes the sense of depth of the acoustic space)

Surr Rolloff -> Off (Defaults to 7kHz and is applied to the extra material steered by L7. Turning this off makes the L7 steered material full range which mirrors discrete MC source)

Front Steering -> This really depends on what you want for your center. I want a center with a well defined soundstage. For me Film (on the MC-4/8/12) works great. On earlier Lexicon's (DC1/2, MC1) the Film setting wasn't as stable with music and it tended to narrow the soundstage a little on some sources. The 'Music' setting is a little bit less aggressive steering and "Music Surround' uses minimal steering and is much more like a passive fill.

Has that been available on any new gear in the last 10 years?

 

[dlaloum]

Has that been available on any new gear in the last 10 years?

JBL Synthesis AVP/AVR's only to the best of my knowledge

 

[Sal1950]

JBL Synthesis AVP/AVR's only to the best of my knowledge

Good call, I had to dig deep into the owners manual of the SDP-58 to find Logic 16 listed.

 

[sfogg]

Has that been available on any new gear in the last 10 years?

Not that has the adjustability. My understanding is that the post Bedford, Ma Lexicon's had the dumbed down non-adjustable versions. Apparently, when it was expanded to more than 7 channels the engineers (which weren't the same ones that designed/programmed it as David Griesinger and the Bedford team was gone by then) may have screwed up the tweaking of it.

Older HK receivers had L7 too and it wasn't the same as what was in Lexicon processors. The HKs lacked the tuning. Some of the HKs also did really weird things on 5.1 DD/DTS sources running 5.1 L7. In them it actually downmixed the multichannel source to 2 channel then ran L7 on it and threw away the LFE. People didn't even realize this was happening because L7 expanded the 2 channel back out to 5/7 channel so well and had more activity in the surrounds then running plain DD/DTS decoding.

Which is sort of my point. If one hasn't used the full blown Bedford,MA versions of L7 they don't know what it could really do and how it can be tweaked to taste. That is why some still use it and/or want something equivalent. I can take a multichannel source (via HDMI or multichannel analog) fold it down to 2 channels and play it back in L7. Doing that L7 retains a great deal of the directionality of the discrete source to the point that if you ran a channel identification test you would still have individual voices in each channel. It isn't nostalgia, it really was that good.

 

[sfogg]

JBL Synthesis AVP/AVR's only to the best of my knowledge

Yup, but likely not the same. Haven't heard it but by many accounts on AVSForum Logic16 is essentially broken. There was also a version called Logic 7 Immersion which added height channels but it lacked the adjustability too. Don't know how well that one worked as it was maybe only in one processor.

JBL Synthesis systems added code into their SDEC processors to add height channels for the JBL rebadged MC-12. I used the same logic from the SDECs to add stage height and regular height channels to my MC-12 but that was when I was bringing the QLI-32 online and have been using that since getting it up and running.

 

[Sal1950]

Which is sort of my point. If one hasn't used the full blown Bedford,MA versions of L7 they don't know what it could really do and how it can be tweaked to taste. That is why some still use it and/or want something equivalent. I can take a multichannel source (via HDMI or multichannel analog) fold it down to 2 channels and play it back in L7. Doing that L7 retains a great deal of the directionality of the discrete source to the point that if you ran a channel identification test you would still have individual voices in each channel. It isn't nostalgia, it really was that good.

All quite possibly true. But I believe it's also very true that when any "processing" has all sort of options available to tweak, the final sound is something that has been fine tuned to that listeners preferences. So of course he's going to love it, right?

 

[sfogg]

You could potentially tweak it to be closer to what you want it to sound like but one will still run into the limitations/design choices of the core processing, tweaking isn't going to change that.

For example, no amount of tweaking gets PLII or PLIIX to do what L7 does, either up front or in the surrounds. PLII simply doesn't handle the center channel as well as L7 if you want to use the center channel (to get rid of the artifical comb filtering on vocals from L/R) while still maintaining a wide cohesive soundstage. Nor can it be adjusted to be as discrete in the side and rear channels as the surrounds are more correlated. That difference is audible in multiple ways. For steered discrete sounds you have less channel separation between speakers and for more ambiance type things the correlation reduces the envelopment and feeling of space that L7 gives. PLII sounds more closed in, not as open.

No amount of tweaking gets Trifield to be even close to what L7 or PLII/PLIIx does in the surrounds. What Trifield does up front is very well done, and works in a different manor that either L7 or PLII.

Likewise, I can't tweak L7 to be able to do what QLI is doing either in the front channels or the surrounds. In some ways I can get it a little closer (removing the LPF on the surrounds for example) but no amount of settings is going to suddenly make a 7 channel matrix steering processor sound the same as a processor that is doing signal decomposition and redistributing the streams across dramatically more speakers.

 

[sfogg]

I'm not sure if they own the tech or have just licensed it. Check out: https://camdenlabs.com/

I sent them an email as a longshot to see if there was any way of adjusting the aesthetic engine in the QLI-32.

To my surprise, I received an email from Dr. Gil Soulodre, the creator of QuantumLogic. He confirmed that the QLI-32 was tuned for movies as that was what it is marketed for. He is leaving on vacation but said when he returns and has a little time he was going to look through his old notes and source code for the QLI-32 to see if there is a way of tuning the aesthetic engine for 5.1 or 7.1 music sources.

 

[Sal1950]

You could potentially tweak it to be closer to what you want it to sound like but one will still run into the limitations/design choices of the core processing, tweaking isn't going to change that.

Well I don't have the ability to use it myself, gear or software, so I can't make any debate on the quality of the final results. To tell you the truth I also don't have the inclination to submit myself to all the mental strain of going thru all that work trying to make a 2 ch source into something resembling a real discrete surround file. Thankfully, of the music I love the most, since the early 1970s more and more of it has appeared in real Quad, 5.1, Atmos/Auro and other high definition, discrete forms. Otherwise until it does, it's just a click or two for choosing between DS, DTS, Auro and a few more to see if one satisfies my desires. They mostly do, but that's just me. YMMV

 

[sfogg]

I think you misunderstand how I used L7, I am not tweaking every song. I tweaked L7 to give me the results I wanted, for all my music. No need to click through multiple modes to find if one satisfy my desires for that album or song. I use to do that 30ish years ago with my Fosgate-Audionics processors as the results were not consistent mode to mode or song to song

Moving to the Lexicon’s greatly reduced that with the exception of center steering setting on the Dc-1/MC1 processors as I would change that sometimes on albums .

MC-12 L7 fixed that as it was a new version of L7 and that was more consistent and stable with the strongest setting for center steering. Pretty sure it is doing some of analysis behind the scenes and adjusting itself too.

 

[dlaloum]

I think you misunderstand how I used L7, I am not tweaking every song. I tweaked L7 to give me the results I wanted, for all my music. No need to click through multiple modes to find if one satisfy my desires for that album or song. I use to do that 30ish years ago with my Fosgate-Audionics processors as the results were not consistent mode to mode or song to song

Moving to the Lexicon’s greatly reduced that with the exception of center steering setting on the Dc-1/MC1 processors as I would change that sometimes on albums .

MC-12 L7 fixed that as it was a new version of L7 and that was more consistent and stable with the strongest setting for center steering. Pretty sure it is doing some of analysis behind the scenes and adjusting itself too.

The Fosgate processors were the ancestors of PLII... Fosgate developed PLII, and his deal with Dolby means that every licence for PLII sold sends some $ to his estate.... which is also why Dolby stopped including it in the default package after they developed the Dolby Surround package.
PLII was included pretty much universally until about 10 years ago.
However, I agree with you that L7 was (is) superior to PLII / Fosgate - but it all becomes very academic now that they are all no longer available other than on increasingly rare vintage gear!

 

[sfogg]

Fosgate improved his processing greatly with PLII. It is much more consistent than his earlier work.

Agree the good L7 is in vintage equipment. It is crazy how little money one can get that for now. I have a number of DC/MC-1 that will go cheap.

It is a shame Harman doesn’t get it working and back into more current products and potentially licensing it to others. It is why people should keep pushing for it.

Ditto with QuantumLogic.

 

[Sal1950]

which is also why Dolby stopped including it in the default package after they developed the Dolby Surround package.

So you don't believe DS was built off the PLII base?

 

[dlaloum]

So you don't believe DS was built off the PLII base?

If they built it off the PLII code base - they would owe the Fosgate estate licencing fees.

So (in theory) DS was built new from the ground up. - and it is pretty decent, but not as good as PLII for music (Fosgate focused on perfecting it for music)

 

[Magnus]

The Fosgate processors were the ancestors of PLII... Fosgate developed PLII, and his deal with Dolby means that every licence for PLII sold sends some $ to his estate.... which is also why Dolby stopped including it in the default package after they developed the Dolby Surround package.
PLII was included pretty much universally until about 10 years ago.
However, I agree with you that L7 was (is) superior to PLII / Fosgate - but it all becomes very academic now that they are all no longer available other than on increasingly rare vintage gear!

I dunno about academic given the abundance on eBay and low prices compared to when they were new ($200 instead of $5000 and $250 instead of $7000). I keep buying up the rare vintage Lexicons.

I now own 5 Lexicons and 1 JBL AV-2. I've got 4 remotes (just picked up another extra DC-1 remote in case one goes bad; I buy the Lexicon remotes when I can get them plus program my own universal ones in the home theater room). Three Lexicons are in use in three different rooms. Logic 7 is worth it. It's the absolute best music upmixer I've heard. It works wonders with my Carver AL-III ribbons + Klipsch Surrounds (6.3 layout with no center and Sonic Holography on the Carvers to boot, giving it a full 15 foot width of the room front/back giant circle. Hell, it does with 6 speakers what it takes 11 speakers to do in the home theater.

 

[sfogg]

If they built it off the PLII code base - they would owe the Fosgate estate licencing fees.

That is a really intersting question. Jim Fosgate didn't ever give Dolby any code, he wasn't a DSP designer at all. Dolby took his analog circuit and modeled DSP around it. Dolby wrote that code. I wonder if they could take that work and build off it without it infringing on the license?

 

[Newman]

[to dlaloum] So you don't believe DS was built off the PLII base?

It seems to have such a quantum leap in technology and even in approach, that it is hard to credit PLII as being anything more than a 'spiritual father'. But I am the last person to claim any deep insight into these technologies.

...So (in theory) DS was built new from the ground up. - and it is pretty decent, but not as good as PLII for music (Fosgate focused on perfecting it for music)

In the Audioholics interview with Roger Dessler, Roger makes it pretty clear that DSU is next-level stuff and PLII/x was a crude beginning, technologically speaking. IIRC he said, paraphrased, that DSU really knows how to deconstruct a sound into its different components compared to PLII handwaving; the way it deconstructs across 20 frequency bands is like having 20 decoders in parallel; it has an amazing ability to discretely isolate and position sounds. And the interview was not particularly neglecting music.

When I, many years ago and filled with excitement for the potential of MCH to enhance music, tried PLII/x both (Movie) and (Music) with music, I couldn't reach for the off button fast enough (with rare exceptions). The adjustable centre spread was definitely helpful, but it wasn't long before I settled back into straight 2CH for 2CH music.

Whereas, trialling DSU+CS for 2CH music for 6 months, so far, I am pretty close to 50/50 depending on the recording itself. I cannot predict which way I will end up going in the long term, and I am still gradually expanding my speaker configuration, but the journey this far with DSU+CS on music is positive regarding ambience and mixed regarding program material, rather than firmly negative....and highly recording-specific. (In fact, if people aren't concluding that their positivity or otherwise about any upmixing technology varies greatly with the recording, and their impression of its success is fixed for all recordings, then something is wrong. And when comments are posted along those lines, I tend to think it's a safe bet that some non-sonic bias is unconsciously driving them.)

cheers

 

[Magnus]

If I turn off the overhead decoding for DSU, I hear very little difference with movies between it and PLIIx. I'm not crazy about either for music, but DSU with overheads works better than Neural X for music (opposite for movies).

Logic 7 is still the best I've heard for music upmixing. I'd Iike to hear Quantum Logic, but the theater units are hard to come buy and difficult to connect at home. I was hoping my newer WRX would have it in the car since it's Harmon Kardon, but Subaru only has it in the Ascent.

 

[sfogg]

In the Audioholics interview with Roger Dessler, Roger makes it pretty clear that DSU is next-level stuff and PLII/x was a crude beginning, technologically speaking. IIRC he said, paraphrased, that DSU really knows how to deconstruct a sound into its different components compared to PLII handwaving; the way it deconstructs across 20 frequency bands is like having 20 decoders in parallel; it has an amazing ability to discretely isolate and position sounds. And the interview was not particularly neglecting music.

When I, many years ago and filled with excitement for the potential of MCH to enhance music, tried PLII/x both (Movie) and (Music) with music, I couldn't reach for the off button fast enough (with rare exceptions). The adjustable centre spread was definitely helpful, but it wasn't long before I settled back into straight 2CH for 2CH music.

Whereas, trialling DSU+CS for 2CH music for 6 months, so far, I am pretty close to 50/50 depending on the recording itself. I cannot predict which way I will end up going in the long term, and I am still gradually expanding my speaker configuration, but the journey this far with DSU+CS on music is positive regarding ambience and mixed regarding program material, rather than firmly negative....and highly recording-specific. (In fact, if people aren't concluding that their positivity or otherwise about any upmixing technology varies greatly with the recording, and their impression of its success is fixed for all recordings, then something is wrong. And when comments are posted along those lines, I tend to think it's a safe bet that some non-sonic bias is unconsciously driving them.)

cheers

Roger was saying it was deconstructing the audio into correlated vs not correlated sound. In other words to common material that is in phase or out of phase portions or specific to L or R in the source. It is not breaking the audio down into individual instrument streams or anything like that. That is literally what QLI does.

Steering by frequency band isn't new with DCU. Circle Surround was doing that back in the 1990s, though with fewer bands.

As far as DSU for music Roger said the surrounds in the early version were bright/annoying, pointed out some odd effects for music and "while your bugging them tell them to put ProLogic II in there."

As far as your 50/50 for music vs 2 channel with DSU, to me that points to a lack of consistency in DSU. Or potentially, being so accustomed to the sound of 2 channel (ex. comb filtering on central vocals) that you reject it when they are reduced/removed. My comments about consistency doesn't mean every album sounds the same. The results, of course, vary by mix. What I mean is that the processing consistently returns results that I prefer to unprocessed 2 channel. I've been listening that way for 27 or 28 years now and don't go back to 2 channel in my listening room. That was not the case with other processing methods, they didn't consistently give results that I preferred to 2 channel.