Audio Science: Does it explain everything about how something sounds?

[Fitzcaraldo215]

If the recording was a simple 2 mike recording, I do think you can get pretty close to reproducing what was picked up by the microphones, and that will be pretty close to what was heard when played over your better systems. That however is a different animal to saying the recording system is able to reproduce the original event. There are considerable limitations in sampling only two points in space then using the resulting recording to reproduce that sound field in another location.

I agree, even if room issues on the playback end have been successfully addressed. The revelation for me in playing back the classical music I love was the discovery 8 years ago of discretely recorded, hi rez multichannel sound. That simultaneous discovery of 3 technologies - hi rez, multichannel and room correction - was by far the biggest sonic leap forward I have experienced relative to the live sound I hear in the hall at concerts, which I attend frequently. No stereo I have heard since then, even those well into multi-100 thousands of $$, comes as close. I listen to little else today. Fortunately, there are thousands of hi rez Mch recordings on SACD and BD-A and -V.

Is it perfect yet? No. But, I think I am on the right track. Perhaps if Auro 3D becomes viable for music reproduction, I might be able to take another step in this direction.

From what I have heard, stereo as we know it is very much a limiting factor in attempting to reproduce the live event.

 

[stehno]

I agree, even if room issues on the playback end have been successfully addressed. The revelation for me in playing back the classical music I love was the discovery 8 years ago of discretely recorded, hi rez multichannel sound. That simultaneous discovery of 3 technologies - hi rez, multichannel and room correction - was by far the biggest sonic leap forward I have experienced relative to the live sound I hear in the hall at concerts, which I attend frequently. No stereo I have heard since then, even those well into multi-100 thousands of $$, comes as close. I listen to little else today. Fortunately, there are thousands of hi rez Mch recordings on SACD and BD-A and -V.

Is it perfect yet? No. But, I think I am on the right track. Perhaps if Auro 3D becomes viable for music reproduction, I might be able to take another step in this direction.

From what I have heard, stereo as we know it is very much a limiting factor in attempting to reproduce the live event.

I agree with your last statement. But only if you exchange the word “stereo” with “playback systems”.

This deficiency you allude to has perhaps zero to do with the number of channels. Rather, this deficiency has everything to do with the percentage of music info that remains audible at the speaker in comparison to the total amount of music info embedded in the recording.

For sake of argument let’s say that 90% of the live performance is captured at the recording and if we pick an extreme example like only 25% of the music embedded in the recording remains audible at the speaker output causing a very unmusical presentation. If you change your configuration from 2-channel to say a 4 channel config. and now there’s a smile on your face. But why? Surely you’re not thinking that 25% x 4 channels equals 100% of the recorded music, are you? Even if you converted to a well-engineered 23.5 channel config. and your PB system was able to play 23.5-ch. you’re still only hearing 25% of all the music info embedded in the recording.

Simply because if the distortions invoking a much raised noise floor are not addressed at the source, the percentage of music info remaining audible at every channel remains the same. 25%. Hence, I suspect the only reason you might be smiling is because of the DSP-like phenomena of having multiple sources of music coming at your ears and convincing you that you are hearing more of the music. But since a soundstage is most always well in front of your listening seat, 2 speakers will generally suffice to create a wall of sound much like live music flowing out from the soundstage. Provided of course the vast majority of music info embedded in a given recording remains audible at the speakers.

Multi-channels, DSP’s, high-rez formats, and acoustic treatments do absolutely zero to decrease a given system’s noise floor and hence, cannot increase this percentage of music info that remains audible at the speakers. IOW, it all boils down to a noise floor level which determines the percentage of music that remains audible at the speakers. It's a percentage thing.

For example. If you’re listening to a 5 minute 54MB-sized Redbook 16/44k format recording or a 300MB-sized high-rez 24/192k recording on the same system, the percent of music output at the speakers still remains exactly at 25%.

Aside from the grossly exaggerated 25% number I used here, I attest this is exactly why even 15 years after the introduction of high-rez formats, there is no consensus that high-rez formats are superior to Redbook. IME, higher-rez recordings generally are moderately more musical, but that’s due primarily to the quantity of information stored and reproduced. If a single blatt of a horn (1 note) consumed say 10k in a Redbook recording and 60k in a high-rez recording, even though I’m still hearing only 25% of that single note, and whether it's 25% of 10k or 25% of 60k, 25% audibility of a single note is still only 25%. But certainly 25% of 60k is capable of sounding a bit more musical since the higher resolution obviously includes a greater quantity of music info than 25% of a 10k note. But make no mistake, regardless of the quality of the format I’m still hearing only 25% of the note. Again, simply because high-rez formats of any kind have zero to do with lowering much raised noised floors induced by various distortions.

BTW, this is also why Meridian’s new MQA format cannot offer any greater levels of musicality over today’s high-rez formats. Unless of course Bob Stuart injects his vast neuroscience research (DSP) into the mix. But then MQA becomes a box of chocolates, meaning you never know what you’re gonna’ get, except that you’re not going to get an increased percentage of music remaining audible at the speakers from a format alone. That’s a scientific impossibility. I think.

 

[Kal Rubinson]

Rather, this deficiency has everything to do with the percentage of music info that remains audible at the speaker in comparison to the total amount of music info embedded in the recording.

For sake of argument let’s say that 90% of the live performance is captured at the recording and if we pick an extreme example like only 25% of the music embedded in the recording remains audible at the speaker output causing a very unmusical presentation. If you change your configuration from 2-channel to say a 4 channel config. and now there’s a smile on your face. But why? Surely you’re not thinking that 25% x 4 channels equals 100% of the recorded music, are you? Even if you converted to a well-engineered 23.5 channel config. and your PB system was able to play 23.5-ch. you’re still only hearing 25% of all the music info embedded in the recording. ............................................

Where do you come up with these numbers?

 

[stehno]

Where do you come up with these numbers?

What does for argument's sake mean to you?

 

[amirm]

For sake of argument let’s say that 90% of the live performance is captured at the recording...

A recording is an unknown process with infinite variables. There is no way it can ever be quantified as a number like this. As I noted in the other thread, the recording is secondary art created by the recording/mixing engineer. It is not a photocopy of the event. Here is a quick example of that in visual domain:

How would you put a percentage on the right painting versus the real photograph on the left? You can't.

We need to erase the notion of "live" from our vocabulary. Everything needs to start at what is already recorded. Not before.

 

[Fitzcaraldo215]

I agree with your last statement. But only if you exchange the word “stereo” with “playback systems”.

This deficiency you allude to has perhaps zero to do with the number of channels. Rather, this deficiency has everything to do with the percentage of music info that remains audible at the speaker in comparison to the total amount of music info embedded in the recording.

...

There are only two possible general issues in reproduction and playback. One is the alteration of the accuracy of existing information as picked up by the mikes in the recording/playback chain - distortion, frequency alteration, noise, adding or obscuring information, listening room distortion, etc. The second issue is whether sufficient information is picked up by the mikes at the outset at the live event and transmitted through the chain at all.

Multichannel is focused on the second issue, and it obviously provides a pathway for more information to be acquired by the enhanced array of mikes at the live event and flow through the chain to our ears. The live sound field in the concert hall is omnidirectional, consisting of direct sound from the performers on stage plus enveloping reflected sound coming from all around the listener. That enveloping reflected sound is inseparable from the direct sound by our ear/brain, unless it is sufficiently delayed in time to be perceivable as distinct echo or reverb. There has been much psychoacoustic research to confirm all this.

In any case, stereo provides only a fraction of that live information. Today's 5 or 7 channel discretely recorded Mch provides a much higher percentage of that information, pure and simple. Stereo also distorts some of the reflected hall sound by redirecting it at the listener from just the front channels - the wrong angles. Mch today is 2 dimensional, however, we may be on the threshold of capturing and reproducing more of the information in the vertical dimension, e.g. Auro 3D, which current Mch systems lack.

The audio industry has been continuously improving on the first sound quality issue I cited since its inception, and no doubt it can be incrementally improved somewhat further still for both stereo and Mch playback, though progress has been slow and will continue to remain so, I expect. But, improving the quality and accuracy of available information via 2 channels cannot completely overcome lack of more information captured in the hall and played back accurately in our listening rooms via many more channels.

As a footnote, there have been 2-channel recording and playback systems that theoretically capture and provide the omnidirectional sound field of live performances. They include coincident pair milking, like the Blumlein technique, and binaural milking via a dummy head for headphone listening. Although around for decades, they have all been a commercial flop and comparatively few recordings of these types exist. There are many reasons, but one is the distortion in the mikes/pickup systems themselves relative to the less distorted, spaced omni mikes used as the backbone of discrete Mch recording.

The proof, of course, is in the listening and making comparisons to my experience at live concerts to the best of my recollection of them, which I attend on average about twice/month. To me, Mch comes much closer to reproducing live concert hall sound than any stereo I have ever heard. And, for me, greater "musicality", whatever that actually means, is primarily a function of that comparison of recorded vs. live.

 

[microstrip]

(...) The proof, of course, is in the listening and making comparisons to my experience at live concerts to the best of my recollection of them, which I attend on average about twice/month. To me, Mch comes much closer to reproducing live concert hall sound than any stereo I have ever heard. And, for me, greater "musicality", whatever that actually means, is primarily a function of that comparison of recorded vs. live.

Are you addressing classical or acoustical music? Can you list a few examples of multichannel recordings that you consider that surpass any stereo performance?

 

[PeterA]

We need to erase the notion of "live" from our vocabulary. Everything needs to start at what is already recorded. Not before.

Really? Then what is to serve as a reference, if not the sound of live, unamplified music? One's imagination of what something should sound like or what some chosen panel of listeners prefer? Should we also erase the notion of memory, ie how me remember a guitar sounding the last time we heard one? Does there not need to be some agreed upon reference upon which we can all agree and use as a foundation to advance the art?

 

[microstrip]

Really? Then what is to serve as a reference, if not the sound of live, unamplified music? One's imagination of what something should sound like or what some chosen panel of listeners prefer? Should we also erase the notion of memory, ie how me remember a guitar sounding the last time we heard one? Does there not need to be some agreed upon reference upon which we can all agree and use as a foundation to advance the art?

Just posted it at another thread debating "how real does it sound"

Many times we refer to non amplified music when referring to life comparisons. And IMHO then SOME aspects of life can become part of our references.

 

[853guy]

A recording is an unknown process with infinite variables. There is no way it can ever be quantified as a number like this. As I noted in the other thread, the recording is secondary art created by the recording/mixing engineer. It is not a photocopy of the event. Here is a quick example of that in visual domain:

How would you put a percentage on the right painting versus the real photograph on the left? You can't.

We need to erase the notion of "live" from our vocabulary. Everything needs to start at what is already recorded. Not before.

Yes and no.

You're right, a recording is - to the consumer - largely an unknown process with infinite variables.

And no, it's not a photocopy of the event, and yes, sometimes it's almost always a secondary art in which the producer and engineer(s) make decisions based on their own aesthetic.

But painting and photography are not time-based art forms. They're static, and not dynamic. Music is a time-based art form, and is constantly "moving" against time in the domains of both pitch and amplitude.

What music shares with painting and photography is intent. In both the examples above, someone attempted to capture something about the essence of the female form and telegraph that ("communicate it") via a particular medium (film stock/paint). Music is the same, but the producer and engineer are (usually) not the genesis of that intent. The artist is. And I've yet to meet a musician who writes and performs music for any other reason than to touch on something about the human experience and share that with others in a way that is unique to them.

And it's that aspect of the art form that is of primary importance (to me) - the producer/engineer are (usually) just there in the service of helping to telegraph it, they're (usually) secondary.

Live is live, a recording is a recording - I completely agree. But as I've said before in an earlier post, music and sound are not the same. You say, "Everything needs to start at what is already recorded." No, it starts with the artist trying to communicate something - it might be banal or it might be profound - but generally at some stage, someone made music ("art") and it got recorded as sound. Playing back the sound is the (relatively) easy part - the music/art part... I'm not so sure.

 

[bonzo75]

Are you addressing classical or acoustical music? Can you list a few examples of multichannel recordings that you consider that surpass any stereo performance?

Most - try Channel Classics Ivan Fischer Mahler 2. Or the Blu ray for Claudio Abbado Lucern festival Orchestra Mahler 5. Play it in MCH, will be better than much more expensive 2ch. Upsampled to Auro 3d is my favorite, by far

 

[Steve williams]

Really? Then what is to serve as a reference, if not the sound of live, unamplified music? One's imagination of what something should sound like or what some chosen panel of listeners prefer? Should we also erase the notion of memory, ie how me remember a guitar sounding the last time we heard one? Does there not need to be some agreed upon reference upon which we can all agree and use as a foundation to advance the art?

I agree Peter.......

there has to be a reference somewhere and for me it is live unamplified music

 

[Robh3606]

Then what is to serve as a reference, if not the sound of live, unamplified music?

Hello Peter

Every recording is different as is the sound at every seat in a concert hall. Having a universal reference is just an illusion. There is no universal reference. We listen to recorded music not live performances. They are completely different things that for some reason seem to get confused as being some how equivalent where one can be used as the standard against the other.

Rob

 

[jkeny]

My take on this dilemma is that we have a high level model/analog of the various sounds we have encountered in our experience of the physical world. This model/analog isn't the details of an oscilloscope or FFT a plot of freq Vs amplitude but some other representation that is specific enough to allow us to compare what we hear with what we expect to hear based on the stored model.

This is the universal reference that I believe exists & it's pretty much the same for everyone (but it does depend on one's exposure to real sounds/audio)

Look at Olive's statement that we all tend towards the same quality of speaker reproduction!

 

[esldude]

Multi-channel clearly provides more information. Regardless of the total getting thru vs the original 5 channels is more than 2 channels of information.

I have never conceptually come to terms with what is the proper method of capturing multichannel. Typically in minimalists set ups you record with one mike per channel. Usually in some sort of a tree. Maybe not a coincident technique, but near that. For stereo, the whole idea makes use of peculiarities of directional hearing to fool us into hearing the sound in front of us and in more or less the proper direction. Those tricks don't really work close to properly except the frontal 60-90 degrees. Seems other than the front 3 channels, any side and rear channels should be placed close to the actual sides and rear of the space being recorded. So you get something of a slice of the soundfield to have some chance of mimicking it upon reproduction in your listening room.

Some of the multi-channel techniques coming out now aren't using any perceptual shortcuts in tricking us so much as placing extra speakers to place a real source of sound in the real point in space or at least much closer to it than happens with stereo or even old 5 channel surround. These are using tricks to be able to pan something they add to the location they want it to be. Presumably if one recorded with mikes spaced around the points at the edges of the space it might reproduce with proper delays and from something close to proper angles upon playback. Of course in fact that is naive (so is normal stereo). JJ has talked about the need for pressure and velocity information for soundfield recreation which is really what we are talking about here. If you don't approach sound field recreation upon playback, you are always doing an impressionist painting of the real event, and various tastes or preferences are possible meaning picky adherence to the real event becomes a fool's errand.

 

[amirm]

Really? Then what is to serve as a reference, if not the sound of live, unamplified music?

That's like asking me why world peace doesn't exist . It is just that way. Audio recordings do not comply with any reference themselves. Nor is that the goal for vast majority of anyone creating recordings. Everything you do in your system just compounds the problem of being removed from the original presentation.

Fortunately, we are not performing forensic analysis with audio. We are trying to get enjoyment out of listening to music. Something we all do everyday with our systems.

One's imagination of what something should sound like or what some chosen panel of listeners prefer?

There is no notion of what something should sound like. Because what was heard by the talent in the recording studio alone is guaranteed to be different than what you are hearing due to differing loudspeakers and rooms, and seating/speaker locations. There is no way the tone, timbre, imaging, etc. is anywhere close to being the same. After all, there are thousands upon thousands of recording/mixing venues, and millions of systems we have. No way can there be a match there.

So where does that leave us? Simple: whatever creates maximum enjoyment wins.

Should we also erase the notion of memory, ie how me remember a guitar sounding the last time we heard one? Does there not need to be some agreed upon reference upon which we can all agree and use as a foundation to advance the art?

There should be. There is one for video. We can at the limit replicate exact colors seen by the camera. That is because the recording equipment is calibrated to a standard and if we do the same in our home video systems, we get the same experience. Audio however, has never had that and nothing is in the works to give it that. It is not a hard problem to solve but the industry has not chosen to solve it, nor are consumers asking for it.

Our brain is quite creative fortunately. You can hear a guitar on a clock radio and the brain instantly knows it is a guitar. You almost can't degrade the audio enough for the brain to fail to do that. The brain also seems to recognize "good sound" without reference. This is why the handful of demo tracks used at audio shows work. They sound great to great many people so manufacturers use them to demo systems.

The other way to think about this is that the art itself, the music, better be a large part of the enjoyment. Not fidelity. I enjoy music I like at all fidelity levels. You have to get it to AM radio fidelity before I stop liking it. In that regard, the aim should be to find music that is great. And have fidelity add the next level of enjoyment to it. If you are going to enjoy music only because you think it is "live" or "real" then in my opinion the priorities are backward.

 

[amirm]

Live is live, a recording is a recording - I completely agree. But as I've said before in an earlier post, music and sound are not the same. You say, "Everything needs to start at what is already recorded." No, it starts with the artist trying to communicate something - it might be banal or it might be profound - but generally at some stage, someone made music ("art") and it got recorded as sound. Playing back the sound is the (relatively) easy part - the music/art part... I'm not so sure.

That "art" includes huge modification of sound prior to getting frozen in the recording media. That is why this is the art we need to worry about replicating, not what happened "live." Here is an example of track being recorded at Nashville's famous blackbird studio:

Look at Dawn Langstroth the signer in that made shift vocal booth. She is wearing headphones. What "live" sound is she hearing? Or the musicians? All that is happening here is gathering raw elements. Some elementary mixing is done to allow them to play the track but that is not at all what we are going to hear on the album. Tons more manipulations happen with the artist and label buyin before we get the real art. That art has the talent's blessing and they know full well, with full intentions, that it has nothing to do with the live sound. Indeed for a ton of music, the artist would commit suicide if the recordings of their voices and instruments were not "'sweetened."

We need to measure the album based on enjoyment scale not some made up notion about it approximating live music. Technically and logically nothing is being done in our systems to present a live experience.

 

[Orb]

And this is also compounded by the multiple track-channels that are mixed/multiple mic arrangements.
We talk a lot about the mastering, but never really how the audio mixing and multiple track-channels/mics setup influences the "live" event and when released as stereo.
Cheers
Orb

 

[stehno]

I once asked JV if he could put a percentage on this idea of our best systems capturing the musical event, and he simply told me at the time, "No".

Peter, if you’re still in contact with JV, you might ask if he ever made that quote? The JV quote I provided I’ve made certain to keep that issue of TAS, circa 2007-2008 but I probably won’t be able to find it. Around 2010, I even inquired with Harley at TAS for permission to use one of Harley’s 2009 quotes as well as JV’s quote from a year earlier.

Haven't we all turned around while watching a movie believing the phone was ringing? Is the reproduction somewhat believable, or are we being asked to identify "is it live or a recording" while blindfolded?

The phone ringing during a movie and causing us to “turn around” isn’t a good example. If so, then my cat freaking out and looking around when my “Rooster” alarm sounds on my iPhone is equally valid.

There are many things that come into play during a movie. First and foremost is that your visual senses are consumed by what’s on the screen and your brain’s processing is focused on the visual input and the plot. Depending entirely on how consumed you are by the visual and the plot and your emotional engagement will determine how you react to any startling sound. Hollywood is well aware of this.

But I remember reading about 15 years ago somebody said, regarding home theater, if you want to improve the perception of your sound system, improve the picture quality. I found that to be a very true statement as it seems the visual processing instantly takes precedence over the auditory processing.

 

[stehno]

A recording is an unknown process with infinite variables. There is no way it can ever be quantified as a number like this. .

Really, Amir? Maybe we’re thinking of two different things.

When I upload a 1 track Redbook CD into my macbook and informs me there is one 54.6MB track of music, should I not trust the accuracy of my computer? We’re talking digital data here right? If I migrate that track to a portable SSD and connect the SSD to my CDP and it also tells me that track contains 54.6MB, could I not then take the leap of faith and assume that 54.6MB of storage represents almost exactly 100% of the recording? Is that not a valid quantifiable number and percentage? Or are you thinking about all the many musical characteristics and potential combinations thereof contained within that 54.6MB? Even so, we still know there is no more music info than 54.6MB for this one recording.

We also know from decades of successful backups / restores in the computer industry that reading (and writing) digital data, reading accuracy is darn close to 100%, even with a $20 CD reader. So is it not safe to assume from that the percentage of digital music info read by the CDP is quite close to 54.6MB or 100%? Is that not relatively quantifiable?

However, once read and our components begin processing/converting that 54.6MB or 100% music info things get a bit dicey. If you’ve ever experienced a single assuming valid/real audible performance upgrade e.g. a burn-in, a cable or component swap, etc. wouldn’t that one experience alone suffice as evidence that at least prior to the upgrade there is a degradation of (a percentage of the 100% music info remaining audible) the 54.6MB of original music info read by the CDP that remains audible at the speaker.

I suppose another valid question might be, if you’ve never observed a single audible improvement, is that evidence that your PB system is already processing 100% of that music info from the source to the speaker output? Or might that be evidence of something else?

What if one experienced 10 or 20 valid audible improvements? How does that play on the percentage of music info remaining audible at the speaker prior to those improvements? After those 10 or 20 improvements are they done? Have they reached sonic nirvana and are the speakers now outputting 100% of the 1 track of music info with nothing more to improve? Our emotions might tell us, wow, it sounds so much better, there can’t be anything more to improve. That may be, but only until the next improvement comes along.

What if one realized well over say 300 valid audible improvements in their current system alone? What would that say about the 54.6MB (100%) volume of music info read verses percentages of music info that remained audible at various points in time?

I’m not a science-minded type, nor very trusting of measuring instruments in this regard as I evaluate everything by ear. Hence, I’m unequipped to measure what percentage of that 100% music info has retained its fidelity that was read by the CDP. Perhaps a science-minded type is equipped to do so, but I doubt it.

As I noted in the other thread, the recording is secondary art created by the recording/mixing engineer. It is not a photocopy of the event. Here is a quick example of that in visual domain:

How would you put a percentage on the right painting versus the real photograph on the left? You can't.

This is a little embarrassing but my sensitive measuring instrument tells me there is a 96.4% accuracy in representation between the two and I’d be satisfied with either.

Seriously though. I think it safe to say a photocopy of a document or even a high-resolution photograph falls well short of the 100% mark. And though we're a bit off topic I'm the first to admit a counterfeit can never perfectly match the original. But for better or worse I think any one of these counterfeits mentioned, including recorded music, for better or worse suffice as facsimiles or representations of a live event. Nothing more, nothing less.

We need to erase the notion of "live" from our vocabulary. Everything needs to start at what is already recorded. Not before.

Agreed and I hope I’ve not indicated otherwise.

Since I have zero control over what makes it to the recording or the quality thereof, my focus always remains only on the recording as the starting point and the speaker output as the end point, the deficiencies of processing the recording, etc. While always bearing in mind that behind every recording was a one time live event.

Accurately reproducing the recorded information is our only Holy Grail. Beyond that, it is entirely outside of our scope and dependent on others to perform due diligence. But I think you'll find that with even some grossly inferior recordings the professional side has actually done a pretty good job overall.