Dvd-audio & sacd are true multichannel formats: a DSP is NOT the same thing! (1 Viewer)

[Rich Malloy]

Kevin, my opinion was very much like yours until very recently. My poor experience with DSP-derived surround music left me quite stereocentric. My ears only started opening up with DVD music releases, but even here most were not to my liking. Even the very best of these, say the DTS track on Diana Krall's "Live in Paris" DVD, are not "pure" mixes, in that they're not an attempt to precisely recreate the sound of the performance. Rather, piano, guitar, etc., often "float" within the mix and are occasionally reproduced discretely in an unnatural channel. Still... I was starting to come around.

And I hasten to add that I very much enjoy the "artificial" surround mixes for appropriate material (nearly always studio-based recordings). SACDS like Beck's "Sea Change", Miles' "In a Silent Way", Bowie's "Heathen", PF's "Dark Side of the Moon", etc., all have instruments discretely mixed into rear channels. For me, this is a perfectly valid musical expression and entirely appropriate for certain recordings.

But not the classical recordings you describe. I'm afraid I can't help you on the DVD-A end, as I'm unfortunately SACD-only for the moment. However, if you get the chance to hear Alison Krauss + Union Station's "Live" SACD (recorded directly to DSD with mics setup to capture all dimensions of sound... not a remixed approximation), or Telarc's SACD of Vaughan William's "A Sea Symphony", or the MT Thomas/SFSO's SACD of Mahler's 6th, or just about any performance recorded by Telarc or Chesky for SACD (either on stage or "live" in the studio -- for example, Ana Caram's "Blue Bossa") , you'll get an excellent demonstration of SACD's capability of accurately reproducing a live performance.

And the timbres... my god, the timbres.

 

[RichardHOS]

Kevin,

I think the posts above adequately answer your challenge to provide "logic" to support this position, but since it was addressed to me I feel there is some detail I can still contribute. I'll have to rehash a bit of what was already stated though, so bear with me.

Assume there are only two microphones in the concert hall recording the live performance, in a suitable stereo configuration (M/S, X/Y, wide separation, etc.). These microphones will pick up information from directions (even tight figure-8 mics) other than straight ahead... where the performers are located.

The sounds this mics pick up from the sides and rear are the reverberations of the performers' sound bouncing off of the walls, ceiling, etc. Those reverberations look like time, frequency, and level shifted "copies" of the original direct sound coming straight from the stage to the microphone. They are, in short, distorted reflections.

Those reflections contain all the information about the size, geometry, and material of the environment (though two mics isn't quite enough to capture all that information). In traditional stereo systems, all of that information is recorded on the left and right channels.

I'm sure you're with me up to this point. The problem is that in stereo reproduction, your left and right speakers are having to recreate not only the direct sound that should come from in front of you, but also all those distorted reflections that should not come from in front of you. The traditional solution has been to damp the early reflection points in your listening environment, purposefully bounce sound all around your room to the back wall, and then let that bounced around sound coming from behind you and to the sides of you recreate the ambient environment.

Don't you recognize the inherent flaw in this approach? It is, in fact, a double flaw. First, you can't prevent the recorded distorted reflections from coming directly from your L/R speakers to you... so you already have sound coming from in front of you that shouldn't. Second, you can't prevent the recorded direct sound from bouncing all around the room as well, mixing in with the reverberant sounds. So where does that leave you? You hear the ambience created by your room, with some hints of the original venue. All in all, it is amazing how good this approach can sound.

But there is a better way. This is where DSP comes in. You see, processors are getting faster, accoustic experts are getting smarter, and programmers are getting more experienced. It now possible for a DSP algorithm to examine a data stream, and actually look for those reflected sounds. It's easy in theory, difficult in practice, but they are getting better and better at it. The essence of what is done is as follows: the data stream is analyzed and waveforms that appear to be time-delayed copies are identified (the better algorithms correctly identifying these copies, despite the frequency (EQ) and level shifting, more often than the less capable ones); these waveforms are digitally removed from the main L/R channels, leaving primarily the direct sound coming from up front; further analysis is done on the remaining "ambient" sounds to attempt to reconstruct their general directional information; and finally the results are routed to the appropriate surround speakers.

Of course, it's more complex than that, but that is the basis of the process. The end result is a reproduction that sounds less like your room and sounds more like the recording environment. Poorly done, it sounds... poor. Well done, it surpasses 2-channel playback... and it's getting better.

The missing piece of the puzzle is the micing techniques at the recording. From a simple L/R microphone pair, you can reconstruct left/right directional ambience information, but front/rear is exceptionally difficult (especially if you don't know the exact polar pattern of the microphone used). Three microphones would allow more precise triangulation (with higher horsepower DSP required, but that seems to be a given with the passing of time).

The ultimate goal would be an increase in awareness of recording engineers as to the desires of the audio-world. If they used a 7 channel blumlein isolated micing technique, little processing would have to be done, and if mixed "straight" channel for channel onto a multichannel format the reproduction would be nearly pristine in its convincing reality. Unfortunately, they have to cater to the largest common denominator, which is often cars and boomboxes. So for now, we're stuck with 2 channel micing (often), and mixing engineers listening in environments less accurate than they could be.

Ah... the agonies of progress.

 

[Rich Malloy]

Thanks for that, Richard! Fascinating reading.

 

[Shawn Fogg]

Kevin,

"As Shawn noted, it will be "my concert hall", and not an accurate reproduction of the actual concert hall of the performance. But this would only be annoying if I had sat in that specific hall "

The problem here though is your have your rooms sound superimposed over the recorded hall sound as well. And the halls original sound is only coming from your two speakers. It can get lost in your rooms sound.

With surround the side and rear speakers play back the halls sound directly to you so it is a stronger influence then your rooms sound. To make this better yet the deader your room is the better.

" I will open my mind to the fact that multi-channel reproduction can be made to faithfully recreate live music (there - I said it)."

It is just like the transition from mono to stereo. Gimmicky recordings can turn you off. But when you here some well done surround (either discrete or by DSP) it should make it clear what the next step is in music reproduction.

Shawn

 

[Kevin_R_H]

Guys,

Don't get me wrong - I'm not throwing up my arms in defeat .

All I'm saying is that if there are multi-channel discs which utilize the rear/side speakers for "spatial information", then at least I accept the theory that they have the capacity to equal (and on occasion surpass) stereo. I fully intend on buying a few of these recommended discs to see for myself.

But this potential advantage comes at what cost?

For me to have purchased "surround amps/speakers" that match my mains would cost, well, at lot more than I had to spend. And if the quality of these "supporting components" are not of equal quality, then wouldn't this discrepency detract from the presentation?

You see, I'm afraid that this theory of "multi-channel audio excellence" remains nothing but a theory except for those audiophiles with an unlimited budget. Because certainly $10,000 worth of pure stereo gear should greatly outperform $10,000 worth of 7.1 audio.

Stereo:
$5,000 for 2 speakers
$5,000 for stereo preamp and power amp

Multi-Channel:
$5,000 for 7 speakers (plus a subwoofer)
$5,000 for pre/pro plus a 7-channel amp.

For someone to start a $10,000 DVD-A or SACD audio system from scratch, I would hope they remember that it's quality that counts, not quantity.

Of course, I'm making the assumption that multi-channel is not so advanced that it has the capability to make "bad gear" sound magically perfect. If that's true, then maybe stereo is dead.

Kevin

 

[Shawn Fogg]

Kevin,

"I fully intend on buying a few of these recommended discs to see for myself."

Don't just listen for what they do wrong... listen to all they do right.

"Because certainly $10,000 worth of pure stereo gear should greatly outperform $10,000 worth of 7.1 audio."

In some areas it should, but in other areas the surround system should have it beat.

" I would hope they remember that it's quality that counts, not quantity."

If you follow that logic through you should have spent your money on the best speaker (singular) and monoblock amp you could afford. Obviously stereo (quantity) gives you advantages over ultimate quality (all money on one speaker).

Shawn

 

[RichardHOS]

Shawn, your concern is a very valid one.

Of course, I'm making the assumption that multi-channel is not so advanced that it has the capability to make "bad gear" sound magically perfect.

Not literally, of course, but IMO "bad gear" for surrounds with a capable processor and quality mains would probably do a more satisfying job of recreating the original ambience than your room would.

 

[Kevin_R_H]

Richard,

Unless I've missed something, this most recent post of yours seems to have an inherent flaw, and that is your harping about the reflected sounds from my stereo speakers, as if the result is unlistenable. But:

1) A full-blown stereo image is a magical thing. A well-designed stereo system can absolutely take your breath away. I know. And you don't need to keep your head in a "locked position" to be amazed.

When you say that "no matter how great your mains are, they don't sound that terrific if you have to listen to them reflected off of various surfaces", I can't help but laugh. In fact, I've got so many thoughts running through my brain, it would take 14 paragraphs to put them down. Let's just say I couldn't possibly disagree with you more.

2) Speaking of reflections, it seems to me there would be many more reflected sounds from a multi-channel system. If the listening room is not chock full of treatments, it seems these multiple reflections should cause many more breakdowns than just from two speakers.

Rear speakers bouncing off side walls; Side speakers bouncing off rear walls; Front speakers bouncing off side walls; all of these bouncing off the ceiling (if it is flat). Or do multi-channel recordings "stick to the walls like mud"?

+++++++++++++++++++++++++++++++

Shawn,

Your posts have been great so far, but don't stoop down to say how one amp / one speaker should be the ultimate bang for the buck. You seem to be smart enough not to bore us with that.

+++++++++++++++++++++++++++++++++++++

Look guys, plain and simply, I am absolutely shocked at your distaste for quality 2-channel rigs. If either of you know of someone in Tampa who would be willing to come to my home to listen, I would love to have him/her report back. And I would particularly love someone who dislikes 2-channel, hates tubes, and is revulsed by vinyl. That way I can kill three birds with one stone .

Kevin

 

[Shawn Fogg]

Kevin,

" but don't stoop down to say how one amp / one speaker should be the ultimate bang for the buck. You seem to be smart enough not to bore us with that."

I didn't make the quality over quantity argument, you did. If you don't like the implications of that argument don't make it. Just like moving from one speaker to two improves the experience so does adding additional speakers.

"Look guys, plain and simply, I am absolutely shocked at your distaste for quality 2-channel rigs."

For me that is because I've heard a better way to listen to music.

"And I would particularly love someone who dislikes 2-channel, hates tubes, and is revulsed by vinyl."

Been there, done that, got the T-shirt.....

Shawn

 

[John Kotches]

Kevin_H said:

Look guys, plain and simply, I am absolutely shocked at your distaste for quality 2-channel rigs. If either of you know of someone in Tampa who would be willing to come to my home to listen, I would love to have him/her report back. And I would particularly love someone who dislikes 2-channel, hates tubes, and is revulsed by vinyl. That way I can kill three birds with one stone.

Kevin:

Who said we have a "distaste" for quality 2 channel? Perhaps we simply prefer multi-channel reproduction to stereo.

BTW, I'd like to thank you for lapsing into cliche` for us with your closing ditties, which basically amount to the equivalent of "you haven't heard my stereo system". Well you haven't heard our multi-channel systems either.

I can't speak for Richard or Shawn's personal experience, but I've heard some incredibly good stereo rigs, and in the end I am left wanting, since the illusion breaks down far too easily.

You are certainly entitled to your preference, but do us all the favor of actually listening to the two algorithms that have been pointed out to you repeteadly within this thread, that just happen to "get it right" when it comes to stereo DSP modes, Meridian's Trifield and Lexicon's Logic-7.

I found the experience with Trifield very eye opening to the point that it was the "difference maker" between the Krell HTS-7.1 and the Meridian 568.2MM in my purchase decision.

Regards,

 

[RichardHOS]

Kevin...

Speaking of reflections, it seems to me there would be many more reflected sounds from a multi-channel system. If the listening room is not chock full of treatments, it seems these multiple reflections should cause many more breakdowns than just from two speakers.

Rear speakers bouncing off side walls; Side speakers bouncing off rear walls; Front speakers bouncing off side walls; all of these bouncing off the ceiling (if it is flat). Or do multi-channel recordings "stick to the walls like mud"?

[/b]Multichannel playback systems require a fundamentally different approach to room treatment than two-channel playback. With a straight stereo setup, you must allow sufficient reverberance in the room (with adequate diffusion, hopefully), or the result will be lifeless. With multichannel, you shoud deaden all reflection points at ear level, not just early reflection points, so that the room is much more "dead," allowing the surround channels to provide ambience information.

Fortunately, the same room approach can be used for theater, discrete multichannel music, and derived (DSP) from two-channel multichannel music playback. It is only straight stereo that has the different room requirements. This is why so many find it difficult to achieve a great two channel system and a great HT in the same room. It isn't the equipment that causes the problem... it's the room.


Look... I know where you're coming from. I'm in the same boat - I can't afford, at the moment, the processor I want. However, I have heard the "best of both worlds" and IMO there's nothing a good two channel room can do that a good DSP/multichannel can't do equally well, and the reverse isn't so kind. Since I'm currently constructing my HT/music room, it only makes sense to look forward a few years. It's a whole lot easier to upgrade processors than it is to upgrade the room layout and treatment. Cheaper too.

 

[RichardHOS]

A bit more clarification on this statement of mine:

"no matter how great your mains are, they don't sound that terrific if you have to listen to them reflected off of various surfaces..."

Imagine it this way: how would your mains sound if you had walls arranged such that you couldn't hear the direct sound from your speakers at all (their sound was absorbed by sufficient arrangements of panels between you and them), and could only hear their sound as reflected off of some typical wall surface? Their apparent imaging would shift to that reflection point, so you could put the speakers anywhere you wished such that the reflection point is where you would normally place your mains. Do you think this "wall reflection" approach would be a truly hi-fi sound?

How much lower quality of a speaker do you think would equal, using direct sound, the quality of this wall reflected sound of your mains? Interesting question at least.

 

[Craig_Kg]

" Shawn stated that stereo requires a dead room. "

No I didn't......

Shawn

My apologies Shawn. I quoted from Richard's earlier post and then John referred to you in his reply (dunno why). Looking back some more, Richard may have been talking about using DSPs on 2 channel music but it wasn't clearly stated that way.

 

[Craig_Kg]

OK, given all the discussion above, does this mean that multichannel speakers (and maybe amps) can be less accurate in the time domain than stereo speakers since the spatial cueing is spread across more sources? This would reduce the price concerns for upgrading to a multichannel music system.

 

[RichardHOS]

That's a complicated question.

One important consideration, IMO, is that if you are summing low frequency information from each channel to a common sub(s), then time/phase coherency is much more important. If you have a separate sub (or full range speaker) at each location, than that issue is much less important (i.e., so long as each speaker is phase coherent within itself, they don't necessarily have to be phase coherent with one another). That's not the most practical, or most cost effective solution though.

In short, I think the requirements for phase coherency in the mains are the same for multichannel and 2-channel, since this is where your primary imaging and soundstage clues come from, and is where the timbre of the instruments/vocals comes from. In general, a phase error in a speaker will manifest itself as an amplitude response error. The fidelity of surrounds in this regard is not nearly as important as the fidelity of the mains (since, as stated above, a mid-fi direct source surround speaker is still of higher quality than a reflected sound from the mains).

I'm not sure if that really answers your question though. In general, I think your mains + sub need to be time/phase aligned to each other and within themselves as much as is possible. I think each surround speaker needs to be time/phase aligned within themselves as much as is practical in a speaker of limited low frequency extension (natural rolloff of the surround speaker will cause pretty bad phase shifting and horrid square wave response), but they don't have to be "phase coherent" with the mains. Time alignment would help, which simply means digitally delaying the appropriate signals.

 

[Craig_Kg]

But since the spatial cueing for stereo is pretty much a function of phase and amplitude differences between the 2 speakers, won't a more direct source (centers and surrounds) allow the spatial image to be created merely by amplitude differences? I'm talking of the phase shifts that speakers usually exhibit over their frequency response even while remaining relatively flat in amplitude.

Summing bass for a sub should be relatively simple if the channels are recorded correctly (and even simpler for stereo being DSPed).

I totally agree about adjustable channel delays being required for multichannel music sources to allow for those of us who can't set an ITU speaker arrangement. Why this isn't availiable in all DVD-A and multichannel SACD players is a continuing puzzlement to me - as you basically stated, only a digital delay of sampling frequency * delay period of each digital stream is required.

 

[RichardHOS]

All speakers have large phase shifts througout their frequency range, regardless of speaker, enclosure, or crossover deisgn. However, humans are not sensitive to absolute phase, so we can't really hear that effect. What you are talking about is relative phase between drivers in a single enclosure. Relative phase offsets will manifest itself as amplitude response variations.

Therefore, a speaker flat in combined amplitude response will have the desired relative phase response.

Phase and amplitude response are intricately linked... you can't separate the two.

Hope this helps.

 

[Kevin_R_H]

Okay, I've read back through my posts, and first of all, I want to apologize for the tone of them. It was unwarranted.

My second post above touched on value, and I don't think this can be emphasized enough. There are many variables in audio reproduction. Some are:
> male and female vocals
> acoustic strings
> woodwinds
> percussion
> horns
> electric instuments
> bass response
> midrange smoothness
> treble (and above)
> clarity
> attack/decay
> dynamics
> frequency response
> instrument separation
> imaging (of voices/instuments within a given soundstage)
> soundstaging (width and depth)
> ability to recreate a given musical venue

I'm sure I'm leaving out a dozen or so, but you get the gist. Our posts up to this point have only dealt with one of these (the ability to recreate a given musical venue). And in my 2nd posts, I admitted my mistake when I stated that multi-channel was inferior at this capability.

But I don't think it is unreasonable to again put forth the stance that $5,000 worth of 2 speakers will likely have better fidelity than $5,000 worth of 7 speakers (don't get me wrong, I know of some $700 speakers that sound better than some $2,500 speakers). And that $5,000 worth of stereo amps should also be of higher quality (2 channels to drive vs 6; 2 channels to process vs all the surround stuff). There's a cost associated with the extra demands put upon multi-channel amps/preamps.

Based upon the criteria I listed, I would expect that a 2-channel system should only be matched/exceeded in "bass response" (because of the dedicated sub), and "musical venue recreation" (based upon all your posts). But to me, recreating instuments/voices in a realistic manner should be the fundamental goal of any music lover. And if the 2-channel system is better at a number of the other criteria, wouldn't the overall musical experience likely be better as well? I kind of think it would for me.

If not, I assume that all 2-channel systems must sound the same to you, as do all multi-channel systems. The quality of components add nothing to the experience. Just throw more speakers at it, and things will get better.

If fidelity isn't really important to you, and you are just looking for the "coolest, gee whiz" effects, then I guess my stance doesn't apply. Also, as stated before, if you are financially rich, and have unlimited funds to spend on your system, then my stance is also invalid. My philosophy only applies to us working stiffs with a finite budget.

+++++++++++++++++++++++++++++++++++++++

Shawn, instead of me making a "jerk comment" to you as I did before, I will actually respond to your post. While having one speaker / one amp would undoubtably allow for higher quality components than stereo, it is unreasonable because there is currently no hi-res (SACD/DVD-A) "mono" software that I'm aware of. If you want to take this argument one step further, then our buyer should spend all $10,000 on a single monoblock amplifier. Higher quality component, but silly and unusable on its own.

In fact, I could be just as silly in attacking your stance (more = better) by stating that the ultimate in audio nirvana in your world would be 2000 speakers crammed into one listening room. Of course, because of our budget, each speaker in this scenario would only be able to cost $2.50 apiece, so I assume this is not your intention.

+++++++++++++++++++++++++++++++++++++++

John,

I'm sorry about the final paragraph in my previous post. As you pointed out, it was uncalled for.

My first post was simply to challenge Richard about "venue recreation". I admit I was premature in my judgement.

My second post was to acknowlege my error, and to speak about "value".

My third post (primarily) also challenged Richard's perceived attack on the viability of stereo. I felt his post implied that "stereo s*cked", and nothing could be done to change it. He backed off this on his subsequent post, and simply stated that while he has heard wonderful stereo rigs, he has heard even better multi-channel setups.

John, as I stated in my first post, I fully accept that thousands of folks prefer multi-channel to stereo. I have no desire to try and change their minds. But I do want to stress that "Hi-Resoulution enhanced CDs" does NOT equate multi-channel playback. These discs (currently) have a stereo signal (that is also enhanced). And my second point is my opinion that for multi-channel to approach the "quality of musical reproduction" of a given stereo system, it doesn't come cheap (in my case would likely double the cost of the overall system).

Kevin

 

[RichardHOS]

My third post (primarily) also challenged Richard's perceived attack on the viability of stereo. I felt his post implied that "stereo s*cked", and nothing could be done to change it. He backed off this on his subsequent post, and simply stated that while he has heard wonderful stereo rigs, he has heard even better multi-channel setups.

[/b]I'm sorry my posts came off that way... it wasn't intended. Stereo reproduction "lacks" only relative to a newer, superior method of sound reproduction. It can still sound as fantastic as it always has.

Again, your concern over price is a valid one. Any new advance is generally cost prohibitive at first. Dolby digital was expensive to begin with (I paid ~$600 for an outboard Marantz processor) relative to ProLogic, but now can be found in budget $200 receivers. "Good" DSP will, in a couple of short years, filter down into the sub-$1000 pre/pro range. In fact, before long is will essentially be "for free" in all good preamps, just as DD is now.

Also, IMO, roughly $1000 on surround speakers and $1000 on surround amplification with proper room treatment (that should be no more expensive than the proper room treatment for stereo listening, just different) would offer superior ambience than wall reflections of an incorrect signal.

So today, yes, it is very possible that a good multichannel system could double the cost of a good two-channel system (add $6000 for a pre/pro, and you're talking about an ~$8000 increase). In the near future, the gap will decrease to just a couple of thousand dollars. And for anyone that wishes to also have an HT setup in their homes, and thus would already have the surround amplification and speakers, this upgrade to multichannel music would, for all practical purposes, be "free."

That's the way I look at it currently anyway. I compare not just the cost of the additional speakers required for multichannel music, but the cost of having separate dedicated rooms for two channel music and multichannel music/HT. When you figure in the price of duplicating your mains, subs, two channels of amplification, and source, it appears cheaper to me to simply "upgrade" the two channel system to multichannel.

If you only have a two-channel setup in your home, then yeah... it's gonna cost you to upgrade it!

 

[Shawn Fogg]

Kevin,

"But to me, recreating instuments/voices in a realistic manner should be the fundamental goal of any music lover. "

Of course, and I think those get better reproduced in surround sound. Esp. vocals because instead of relying on phantom imaging (and the attendant comb filter which screws up the sound) you can have a single speaker reproduce the majority of voices. Of course the quality of the center channel needs to match your L/Rs and the steering for the center needs to be good for this to work well.

" Of course, because of our budget, each speaker in this scenario would only be able to cost $2.50 apiece, so I assume this is not your intention."

There of course is a balance to be struck. But again, you made the quality over quantity argument... not me. I was pointing out that you aren't following that argument since you listen in stereo. Obviously increase your quantity there improved your overall listening experience... so don't dismiss the possibility of multichannel doing the same. It certainly did for me.

"If not, I assume that all 2-channel systems must sound the same to you, as do all multi-channel systems. The quality of components add nothing to the experience. Just throw more speakers at it, and things will get better."

Kevin, we have stated many times that the quality of the DSP itself is vital to making this work. Do you doubt we don't value quality in the rest of our systems too?

John,

" but I've heard some incredibly good stereo rigs, and in the end I am left wanting, since the illusion breaks down far too easily."

Yes, and for me every time I hear a great two channel system my thought is always the same.... 'but how would it sound in surround?' Now that Trifield has won you over I suspect you may be the same.

Shawn