"Rock/Pop" MultiChannel shining in Stereo

[esldude]

10 hours ago, shtf said:

 

Perhaps you think it bizarre because you're unable to fathom that as in my example of the arbitrary 50% remains 50% regardless of format content volume.  

Yes, this is illogical and unfathomable.  Unless you have some more subtle point that needs more explaining this is simply ridiculous. 

Quote

 

How bizarre is to conceive that an arbitrary 50% of the music info of single guitar note remaining audible at the speaker is still 50% of a single guitar note regardless of format content volume?  That is so long as the distortions induced along the input signal processing path remain unchanged.

It is very bizarre.  It doesn't fit any explanation of how the gear including speakers work.  Or the practical experience of that gear. 

Quote

 

Why is that bizarre?  Throw in an superior pair of interconnects and speaker cables and maybe that arbitrary 50% audibility changes to 54 or 55% remaining audible at the speaker.  Add superior line conditioning and maybe it’s now 66 or 67%.  Or if per chance the replacement is inferior, maybe that arbitrary 50% remaining audible drops to 47 or 46%.

 

If an actual improvement occurs in a system, what exactly do you think is happening to generate that improvement?  I’m reminded of a poli-sci professor who said on the first day of class, “IMO, you’re all A students – and it’s all downhill from here.”  Some of us laughed while others expressed their confusion or concern.  I could easily say the same thing of a transistor radio or SOTA-level playback system.  They are all 100% accurate – until you turn them on and push play.

 

The above is pointless babbling.  

Quote

 

You say my premise makes no sense and is bizarre, yet you, using my same grading example, then say you believe roughly 90% of the input signal retains its fidelity all the way to the speaker.  Although you are far from alone in your thinking, there certainly is no evidence to support such a high percentage.  I think of one gent calling himself the audio expert who claims all cables and all components retain 100% of the input signal's fidelity.  If there was any truth to such claims, perhaps aside from proper speaker positioning, then there would be very few, if any improvements remaining to be made to any playback system.  Comments in perhaps any high-end audio forum would suggest the contrary is true.  That some form of improvement may be found under perhaps every corner of the carpet.

The evidence is being able to measure distortion, noise and other contributions that corrupt the input signal.  They are no where near 50 %.  The entire playback chain from source to speaker input will all add up to maybe 1%.  The speaker may add another percent or three.  Various radiation patterns may do damage which is difficult to put a percentage upon.  Your wire feeding the speaker is NOT going to fix those issues. 

Quote

 

Moreover, if one is intimately familiar with live music and playback music it is not uncommon for them to also realize the rather large gulf that separates the two, even with today's state-of-the-art level playback systems.

Having done some recording I agree.  But for altogether different reasons.  A base level reason is the limitations of stereo recording and playback.  The other is the tiny percentage of recordings that are relatively unprocessed and simple enough to have a chance to show what is possible.  Something far less than 1% of all recordings. 

Quote

 

For example.  Back in a more innocent age:

 

1.  Robert Harley, editor-in-chief of The Absolute Sound stated in the Mar/Apr 2009 issue, "I believe that something catastrophic occurs at the recording mic's diaphragm so that much of the music never makes it to the recording."  Paraphrased.  Harley went on to explain that his premise was based on a somewhat silly experiment conducted by Ed Meitner and it was Meitner who actually came to this conclusion and Harley I suppose just went along for the ride.  Meitner (and Harley) was just wrong about the cause but Harley was accurate when he used the word catastrophic to describe the effects.

 

2.  About a year earlier, Jonathan Valin, senior editor of The Absolute Sound wrote, "We are lucky if even our very best playback systems are able to capture 15% of the magic of the live performance."  Paraphrased.  I've known a few who not only agreed with Valin’s statement but went further to say 15% was perhaps a bit too optimistic.  BTW, I translate magic to believability.

 

If you think my premise of an arbitrary 50% remaining audible is bizarre, then what is your explanation for various improvements you read that others encounter or perhaps even you encountered?  Or the only-marginal sonic differences between Redbook CD and hi-rez?  Or why the supposed "hi-rez" MQA format has fallen on its face when compared to its introductory out-of-this-world performance claims?  Why even today some 18 years after the introduction of hi-rez formats, few if any in a high-end audio forum would bat an eye if somebody opened a new thread asking something like, "Is there a sonic difference between Redbook CD and hi-rez?".

Placebo effect plain and simple is the explanation for the bulk of the above.  The loss at the microphone is not very great for the signal that hits the microphone.  The big loss is that sampling two points in a huge space and reproducing that simply has severe limitations to recreate the entire event.  Your microphones aren't that bad, but they aren't capturing all the needed information.  It isn't there to be had in a stereo recording.  

Quote

 

You say MCH is a much better resolution that 2-ch because of the inherent limitations of 2-ch.  Again, there can exist a 5.1 channel recording format or even a 55.7 recording format and yes, there can be some sonic benefit to such configurations.  But 50% audibility at the speaker still remains 50% regardless of the number of channels.  Even if there’s volumes more information being processed.  So I understand there’s a hint of truth to this claim but resolution is definitely not the right word to use. 

You are unhinged about the 50%.  Simple information theory shows more channels of the same quality (even the 50% you babble on about) have the potential for more information than lesser channels.  And listening to the well done recordings in a good MCH setup make that abundantly clear.  You can hear it for yourself. 

Quote

 

Besides, who could possibly make such a claim and maintain credibility?  Meitner?  Harley?  Valin?  IOW, if nobody was even aware that there’s far more music info embedded in a given recording than what we hear at the speaker, how could anybody know the type of presentation a well-thought-out 2-ch system is truly capable of reproducing? 

 

BTW, my premise of accurately retrieving 100% of the music info embedded in a recording but due to distortions inducing a much raised noise floors with only percentages of the original input signal remaining audible at the speaker can explain nearly every last deficiency, known and unknown, in high-end audio. .......snippage of MQA irrelevancy..............

 

Again, although my percentages are a bit arbitrary this is why hi-rez formats (regardless of format content volume) and why more channels like multi-channel actually do little or even nothing to improve real sound quality or get us closer to the live performance.

Most of the above is demonstrably wrong headed.  You make the claim nobody could know what a well thought out 2 channel system is capable of reproducing and yet you couldn't make your claims unless you knew.  You explanation is hung up in some recursive crazy loop here.  

 

I'll also add, that your goofy ideas would point you toward mono.  You can say all the same things and determine in your world that mono is best because no one knows about the arbitrary 50% and the additional channels do nothing to change the arbitrary 50% problem.  Mono is king.  

 

 

[Le Concombre Masqué]

11 hours ago, Ralf11 said:

the real issue is whether you think you can do a better job at home, than the mastering engineer did at the console

 

"So you need to ask yourself 'Do I feel lucky today?'"

"Well, do ya?"

I wish all the sarcastic, sometimes preposterous, comments that now abound don't shy away music lovers to unearth their packages, rip the MCH and give MCH to Stereo a try; afaic failed exemples are the minority, such as the recent White Album, better to my ears in 2.0 hires. 

 

The matrix pipeline in HQP is set once ; now I do no job but feeding a mch instead of a stereo one.

 

Convolving separately 6 channels in HQP consistently yields much better results than processing the 2.0 mix with sound characteristics (delineation of bass lines, smoothness and naturalness of vocals, soundstage depth...) in a league beyond mastering differences.

 

I feel like I have just met true Hires.

 

Now, how it fits with Nyquist etc......

 

Maybe esldude will elaborate on "Simple information theory shows more channels of the same quality (even the 50% you babble on about) have the potential for more information than lesser channels.  "

 

 

[Miska]

On 12/1/2018 at 5:46 PM, Kal Rubinson said:

How did you arrive at these settings?  Is there any reason to believe that they are generally applicable given that the original multichannel mixes vary so greatly in their balance/distribution? 

Edit:  @Miska

 

These are based on mix-down coefficients from the Dolby Digital aka AC-3 standard. In addition, I've done quite a bit of experimentation with these values and found them work nicely for the multichannel material I've found, including DSD (which is 5.0 channel most of the time, so without the LFE).

 

One good set for experimentations has been also the Genesis boxes with DVD + SACD for all albums. Both discs, all layers and audios ripped. I ripped the DTS multichannel PCM tracks into multichannel FLACs. For music experimentation, there's quite a bunch of multi-channel DSD material and also PCM from 2L.

 

When it comes to sonic differences, for large part of music content, stereo version is not based on mixdown, but instead either just feed from front channel mics of multichannel setup, or from a separate stereo pair. So stereo mixdown from 5 or 6 microphones provides a different kind of view on things. I have also found that in many cases the mixdown sounds better. Neither one is wrong, they are just different views to the same music.

 

[esldude]

11 minutes ago, Le Concombre Masqué said:

 

 

Maybe esldude will elaborate on "Simple information theory shows more channels of the same quality (even the 50% you babble on about) have the potential for more information than lesser channels.  "

 

 

In most simple terms, I have pipe of a given size at a given pressure and it flows 1 gallon(liter) per minute of liquid.

If I have two such pipes I get 2 gallons (liters) per minute of liquid. 

If I have 5 such pipes I get 5 gallons (liters) per minute of liquid. 

 

Or take another approach. I have a two channel stereo system setup optimally.  Right and left. 

There are limitations for reproducing hall ambiance.  I add some channels in the rear for hall ambiance and it is coming from a proper direction for that.  It in no way reduces the front two channels, but adds additional info that isn't contained with in them.  Adding this other info into the existing front channels will not be as good.  

[Miska]

5 minutes ago, esldude said:

Adding this other info into the existing front channels will not be as good.  

 

You anyway listen it with two ears...

 

If you've ever tried good 3D audio with headphones you know that it sounds so real that you are truly convinced sound is coming all around you and you can pinpoint a sound source behind you for example.

 

I do a lot of headphone listening, and especially for that, the multichannel mixdown is so much better in many cases... It is not 3D audio, but better than the headphone cross-feed systems that attempt to put some space back to ordinary stereo recording.

 

For some headphone fun, one could for example check out these:

https://www.nativedsd.com/homepage/binaural_dsd_music

 

[STC]

Here is another thread about the same album. BTW, out phase content can give sense of surround. 

 

https://www.quadraphonicquad.com/forums/index.php?threads/using-the-stereo-image-to-get-5-1-with-jimi-hendrix-electric-ladyland.17064/

[mansr]

9 hours ago, Kal Rubinson said:

AFAIK, DST is lossless.

It is.

[Kal Rubinson]

4 hours ago, Miska said:

When it comes to sonic differences, for large part of music content, stereo version is not based on mixdown, but instead either just feed from front channel mics of multichannel setup, or from a separate stereo pair. So stereo mixdown from 5 or 6 microphones provides a different kind of view on things. I have also found that in many cases the mixdown sounds better. Neither one is wrong, they are just different views to the same music.

Well put.

[Kal Rubinson]

4 hours ago, Le Concombre Masqué said:

Convolving separately 6 channels in HQP consistently yields much better results than processing the 2.0 mix with sound characteristics (delineation of bass lines, smoothness and naturalness of vocals, soundstage depth...) in a league beyond mastering differences.

As compared to stock stereo, I assume.  Of course, it involves corrupting most of the directional/spatial information that the 6 channel vehicle preserves. ?

[Kal Rubinson]

4 hours ago, esldude said:

Or take another approach. I have a two channel stereo system setup optimally.  Right and left. 

There are limitations for reproducing hall ambiance.  I add some channels in the rear for hall ambiance and it is coming from a proper direction for that.  It in no way reduces the front two channels, but adds additional info that isn't contained with in them.  Adding this other info into the existing front channels will not be as good.  

Amen.

3 hours ago, Miska said:

I do a lot of headphone listening, and especially for that, the multichannel mixdown is so much better in many cases... 

Have you compated that with a Smyth-rendered experience?

[ARQuint]

34 minutes ago, Kal Rubinson said:

 

4 hours ago, esldude said:

Or take another approach. I have a two channel stereo system setup optimally.  Right and left. 

There are limitations for reproducing hall ambiance.  I add some channels in the rear for hall ambiance and it is coming from a proper direction for that.  It in no way reduces the front two channels, but adds additional info that isn't contained with in them.  Adding this other info into the existing front channels will not be as good.  

Amen. 

 

However, when you do that, you're adding a sort of generic ambience - kind of like those buttons on receivers that promise a "stadium" or "cathedral" or "night club" acoustic. Discreet m-c, done right, gives you a specific venue's ambient signature—the Concertgebouw is different than Mechanics Hall is different than the Bayreuth Festspielhaus, etc. IMO, the "sense of occasion" that a good surround recording can provide derives in part from the evocation of a unique place.

 

Andy Quint

 

Senior Writer

The Absolute Sound

[Miska]

45 minutes ago, Kal Rubinson said:

Amen.

Have you compated that with a Smyth-rendered experience?

 

There's no Smyth in Finland.   And I haven't noticed it at Munich High-End either. So I really don't know...

 

I have compared it to other headphone processing and I think it is at least as good in many cases.

[Miska]

1 hour ago, Kal Rubinson said:

As compared to stock stereo, I assume.  Of course, it involves corrupting most of the directional/spatial information that the 6 channel vehicle preserves. ?

 

I would say it enhances that informations and puts in some of the information that two channel mic array lacks due to physics/math.

 

OTOH, lot of recordings never originated from single stereo pair of microphones. Like the Genesis box set, or DSOTM, etc. Those were always multitrack studio productions that never had any real directional/spatial information. All that is artificially created in studio. Original DSOTM mix was quad channel, like many other things of the day. Most of the music I listen are studio productions anyway...

 

[Miska]

49 minutes ago, ARQuint said:

However, when you do that, you're adding a sort of generic ambience - kind of like those buttons on receivers that promise a "stadium" or "cathedral" or "night club" acoustic. Discreet m-c, done right, gives you a specific venue's ambient signature—the Concertgebouw is different than Mechanics Hall is different than the Bayreuth Festspielhaus, etc. IMO, the "sense of occasion" that a good surround recording can provide derives in part from the evocation of a unique place.

 

You mean M/S processing? Yes, it works very well. The first hifi playback system I was involved with as a kid was M/S system (Orthoperspecta) that had three loudspeaker setup and needed only two channel amplifier. Center channel was playing sum signal of stereo source and side channels, connected in opposite phases, were playing difference signal of stereo source.

 

One can also make a "native" recording for this system using M/S microphone technique too - which also works well for stereo, but requires simple decoding for stereo use.

 

[Le Concombre Masqué]

3 hours ago, Kal Rubinson said:

As compared to stock stereo, I assume.  Of course, it involves corrupting most of the directional/spatial information that the 6 channel vehicle preserves. ?

Don't think it matters much with Hendrix, Talking Heads (downmixed remain in Light is gorgeous) , Seal, Bjork, Steely Dan, King Crimson... you're probably right if I was to take say a Jordi Savall SACD

[Kal Rubinson]

2 hours ago, Miska said:

I would say it enhances that informations and puts in some of the information that two channel mic array lacks due to physics/math.

You would say that since you are speaking from the position of a two-channel listener and the two channel mic array benefits from any such help.  OTOH, playing back the discrete 5 channel recording via a multichannel system preserves the proper directional information in those additional channels and I am speaking from the perspective of a listener who benefits from that preservation.

[Kal Rubinson]

3 minutes ago, Le Concombre Masqué said:

Don't think it matters much with Hendrix, Talking Heads, Seal, Bjork, Steely Dan, King Crimson...

Sure.  Not my concern.

[Le Concombre Masqué]

25 minutes ago, Kal Rubinson said:

You would say that since you are speaking from the position of a two-channel listener and the two channel mic array benefits from any such help.  OTOH, playing back the discrete 5 channel recording via a multichannel system preserves the proper directional information in those additional channels and I am speaking from the perspective of a listener who benefits from that preservation.

I'm not on the verge of investing in a 5.1 system made of components in the league of my 2.0 (+ would definitely need a new room/house) but just curious : how many records of musical top value meet your criteria ? I don't care for first rate audiophile recordings of second rate interprets. I bet the Savall are good, the Italiano on Pentatone are excellent in Stereo and I bet the MCH is great too but... how many technically satisfying MCH of interprets in the league of Savall or the Italiano do you know of?

[Kal Rubinson]

13 minutes ago, Le Concombre Masqué said:

I'm not on the verge of investing in a 5.1 system made of components in the league of my 2.0 (+ would definitely need a new room/house) but just curious : how many records of musical top value meet your criteria ? I don't care for first rate audiophile recordings of second rate interprets. I bet the Savall are good, the Italiano on Pentatone are excellent in Stereo and I bet the MCH is great too but... how many technically satisfying MCH of interprets in the league of Savall or the Italiano do you know of?

That is a highly subjective analysis but I'll take a shot at it.  Out of the 37K+ multichannel files that I have on my server (about half my collection), I would guess(!) that more than half of them are of top musical value.  That is not to say that I might not have a musical preference for a stereo recording of any given piece but, since I do have enjoy having duplicate interpretations of many pieces, these multichannel ones would be in my collection even if they were only stereo.

 

I am not insisting that everyone must have a multichannel system and I appreciate that the vast majority of us live under various constraints which force us to make particular choices.  What I am insisting is that discrete MCH is inherently superior to 2 channel.  What anyone does about that is up to them.

[Miska]

53 minutes ago, Kal Rubinson said:

You would say that since you are speaking from the position of a two-channel listener and the two channel mic array benefits from any such help.  OTOH, playing back the discrete 5 channel recording via a multichannel system preserves the proper directional information in those additional channels and I am speaking from the perspective of a listener who benefits from that preservation.

 

Yes, sure. Certainly playing multichannel material through multichannel system is better. I was looking this only from perspective of two channel playback system, especially headphones in this case. Not comparing two channel speaker system to a multichannel speaker system...

 

[esldude]

3 hours ago, ARQuint said:

Amen. 

 

However, when you do that, you're adding a sort of generic ambience - kind of like those buttons on receivers that promise a "stadium" or "cathedral" or "night club" acoustic. Discreet m-c, done right, gives you a specific venue's ambient signature—the Concertgebouw is different than Mechanics Hall is different than the Bayreuth Festspielhaus, etc. IMO, the "sense of occasion" that a good surround recording can provide derives in part from the evocation of a unique place.

 

Andy Quint

 

Senior Writer

The Absolute Sound

I was thinking of real multi channel. Not some faux multi channel. So the ambiance would be recorded from the real location.

[gmgraves]

On 12/1/2018 at 7:43 AM, Kal Rubinson said:

Especially if you play it in multichannel.? 

To be frank, my experience with down-mixing good multichannel to stereo has been consistently disappointing.

I should think that it would be. The "design" of the mix between a two-channel and multichannel version of the same recording is not designed so that everything in the MCH version will just "fold" seamlessly down into a coherent two-channel rendition of the same title. That's not to say that perhaps it shouldn't. In an ideal world, I would think, a MCH  version of a recording would be exactly like the two-channel version except that the hall ambience that is present in the 2-channel, while also in the MCH version, has been broken away from the front channels and redirected to the other channels so that it mimics the sense of space one gets when  one is in a larger venue (than one's listening room) listening to the music live. But I suspect that if the record companies did it strictly that way, MCH wouldn't be very impressive and many in the target audience wouldn't be enthused enough to spring for the extra hardware. The only MCH recordings that I have heard that I thought were done close to correctly were Ray Kimber's "IsoMike" MCH recordings. Those recordings were so popular that he stopped making them! I suspect that the realistic approach to soundstage and ambience retrieval that Kimber's unique mike technique brought to the table, was simply not impressive  enough for the average MCH enthusiast. It was however, very natural (IMHO, anyway).   

[Kal Rubinson]

1 hour ago, esldude said:

I was thinking of real multi channel. Not some faux multi channel. So the ambiance would be recorded from the real location.

But with the mixdown, it would be coming from the wrong direction.?

[Kal Rubinson]

1 hour ago, gmgraves said:

In an ideal world, I would think, a MCH  version of a recording would be exactly like the two-channel version except that the hall ambience that is present in the 2-channel, while also in the MCH version, has been broken away from the front channels and redirected to the other channels so that it mimics the sense of space one gets when  one is in a larger venue (than one's listening room) listening to the music live.

I see what you are saying but I look at it another way.  It is not just the ambience present in the 2 channel vs. the MCH recording.  Each captures the entire ambience but each mic captures it from its particular point of view.  What is lost in going from, say, 5 to 2 channels, is not the amount of ambience but, rather, the directional information that the additional mic channels contribute.  Mixing from 5 to 2 loses (or corrupts) the directionality of the transferred ambience.  Up-mixing from 2 to 5 cannot recover it without considerable psycho-acoustic computation. 

[Miska]

1 hour ago, gmgraves said:

I should think that it would be. The "design" of the mix between a two-channel and multichannel version of the same recording is not designed so that everything in the MCH version will just "fold" seamlessly down into a coherent two-channel rendition of the same title. That's not to say that perhaps it shouldn't. In an ideal world, I would think, a MCH  version of a recording would be exactly like the two-channel version except that the hall ambience that is present in the 2-channel, while also in the MCH version, has been broken away from the front channels and redirected to the other channels so that it mimics the sense of space one gets when  one is in a larger venue (than one's listening room) listening to the music live. But I suspect that if the record companies did it strictly that way, MCH wouldn't be very impressive and many in the target audience wouldn't be enthused enough to spring for the extra hardware. The only MCH recordings that I have heard that I thought were done close to correctly were Ray Kimber's "IsoMike" MCH recordings. Those recordings were so popular that he stopped making them! I suspect that the realistic approach to soundstage and ambience retrieval that Kimber's unique mike technique brought to the table, was simply not impressive  enough for the average MCH enthusiast. It was however, very natural (IMHO, anyway).   

 

The standard surround mic setups are quite good and seem to be used for example by 2L and probably many others:

https://www.dpamicrophones.com/mic-university/immersive-sound-object-based-audio-and-microphones

 

They have a suitable mount and even different bundles with mics:

https://www.dpamicrophones.com/accessories/surround-decca-tree-mount

 

Various mix setups are part of multichannel specifications and thus have to be taken into account when producing multichannel content.

 

One advantage of multichannel to stereo mixes is when the multichannel or stereo content is not recorded with omnidirectional microphone. Cardioid and such mics have poor bass response, so using additional omni optimized for low frequencies (LFE channel) improves bass response of the recording. In addition if the mics are not diffuse field compensated, they provide uneven frequency response to different directions. Again mixing down multichannel material provides more even frequency response to all directions. In addition, doing this mix at player side one can adjust amount of space information in the mix to his liking by adjusting the mix gains.

 

 

P.S. From purely technical point of view, there is also slight SNR (mostly mic self-noise) improvement in making such stereo mix.