What's wrong with SACD, and how to fix it

GoldenBoy, maybe you are not reading some of my posts and Malc's post. Think logically for one minute. The definition of analougous is simply (again from The American Heritage Dictionary, Fourth Edition 2000):

"Similar or alike in such a way as to permit the drawing of an anaolgy."

Read some of the examples that I have used and that Malc S has used. In nature a soundwave or acoustic wave emanates from a source through space (utilizing the principle of molecular excitation) in such a fashion that the original shape and size of the soundwave or acoustic wave is transmitted in an analougous manner. In other words, from point 'A' to point 'B', the basic shape and size of the acoustic wave is transmitted in such a manner that there is an analougous representation of the original shape and size of said acoustic wave at each point in space between point 'A' and point 'B'. Remember that we are only talking about words (i.e. analoug and acoustic waves) that merely symbolize physical phenomena (sound propagation). There really is nothing more to read into this whole debate simply because the words are mere conventional idioms that we are using to communicate a specific idea regarding a natural phenomenon. We could use other words to describe the natural phenomenon that we call acoustical propagation, but there is no denying that in nature these waves are transmitted through space (wheather air or liquid molecules) in an analougous manner. It would be silly to try to use other words that are not a conventional part of our language.

 

I have no problem with the words, but the fact that people are somehow trying to equate analogue recording techniques as being the natural way of sound reproduction, and that is just not correct. I will not agree that sound is analogue by nature. Sound is what it is; it is analagous to nothing. To use the example of the drum again, sound waves in air or water or vibrations in stone cannot be created without the drum, likewise without the air or stone or water the drum cannot create a sound wave. The sound of the drum and the medium through which it travels are one and the same. You cannot have one without the other. Sound is physics and physics just is. It is neither analogue nor anything else.

This also gets back to my point that you cannot seperate 'sound' from the condition of hearing and how our brain interprates it, because that is sound. Without someone to hear it, is there sound?

Here is what MSN Encarta has to say about sound:

(Dictionary):
sound

noun*
1.*something audible:*something that can be heard
not a sound in the whole house
the sound of gunfire
2.*physics*acoustics*vibrations sensed by the ear:*vibrations traveling through air, water, or some other medium, especially those within the range of frequencies that can be perceived by the human ear.*At sea level and freezing point the speed of sound through the air is 1,220 km*(760 mi) per hour.
3.*sensation of vibrations:*the sensation produced in the ear by vibrations traveling through air, water, or some other medium
4.*electronic engineering*reproduced music or speech:*the music, speech, or other sounds heard through an electronic device such as a television, radio, or loudspeaker, especially with regard to volume or quality

(Encyclopedia):
Sound

physical phenomenon that stimulates the sense of hearing. In humans, hearing takes place whenever vibrations of frequencies from 15 hertz to about 20,000 hertz reach the inner ear. The hertz (Hz) is a unit of frequency equaling one vibration or cycle per second. Such vibrations reach the inner ear when they are transmitted through air. The speed of sound varies, but at sea level it travels through cool, dry air at about 1,190 km/h (740 mph). The term sound is sometimes restricted to such airborne vibrational waves. Modern physicists, however, usually extend the term to include similar vibrations in other gaseous, liquid, or solid media. Physicists also include vibrations of any frequency in any media, not just those that would be audible to humans. Sounds of frequencies above the range of normal human hearing, higher than about 20,000 Hz, are called ultrasonic.

 

You are almost there. If you think about it just a little bit harder I think that you will get it.

 

I've already got it; I just don't agree with you.

 

This one is simple folks: perception is reality. So if you perceive digital as flawed and not real "sound", but 0s and 1s captured, well, that's it. It can never be "analog" in any way to someone.

I tend to agree with Goldenboy. Our ears have preferences, but sound is sound, whether it's generated by a kazoo, a pure digital path from my CD player into my receiver with S/PDIF, or a turntable into phono inputs. The speaker (or sound output device on, say, the kazoo) then sends that to my ears to let my body interpret things.

I'll bet people had these arguments of going to bazillion track multi-track recording, too. "It's not real if it's not recorded live!" We've gotten over that, thankfully.

These threads entertain me

 

Let's assume GoldenBoy is right for the sake of argument....

Isn't still possible that analog is more inherently natural due to continuous sampling?

The reason I ask this is because I get to listen to master tapes in the studio (many of which supposedly have lower performing metrics than digital media) and I am constantly amazed how good and natural and realistic they can sound if they are from a good session. As good as hirez can be, there is something better and right about a good tape.

 

I listen to some reels of that caliber, and I've recorded both digitally and analog for my own stuff. Bottom line is that if it's recorded well, it will sound good. Recorded badly - it's gonna sound like poo no matter what you do to it. We just did some recording at 48/24 through great preamps, etc., direct to HD and it sounds phenomenal to my ears in the studio.

And let's be honest: how many people play the master tapes at home? No one, except maybe the artist. WHat we hear on LP, CD, tape, 8-track, SACD, DVD-A is a processed version of that.

I also somewhat discount listening in a studio because studio monitors and such are some degree different beasts than your average person's speakers at home. I've had mixes that sound great in the studio but suck at home. Same mix, different equipment.

Analog AND digital color sound recorded onto media (tape or bits on a disk somewhere) due to the way they store it. Analog may be more pleasing to your ear, but I've A/Bed reels I've archived at 96/24 to the original, and the 96/24 doesn't sound harsh to me or that I'm missing things.

Your mileage may vary of course. I trust my ears, which tell me my opinion is 100% correct

 

I find that master tapes sound better than 24/96 archives but I think you are right to rely on your own ears. The problem I have with 24/96 is it still sounds harsh in the upper frequencies and cymbals still sound a bit scratchy.

 

I've read that in a listening test, most folks found the greatest improvement in PCM sound to come from the jump between 96 and 192. Can't remember who did the test, why, or when -- sorry.

 

Let me once again try to clarify my position on the difference between "analogue" and "digital" as they pertain to acoustic waves.

First, let me say that what I am talking about isn't anything to do with whether an analogue system is fundamentally better than a digital system.


Since it is a concept that we are all familiar with lets consider the chain of events that occurs when you hear someone beating a drum.

i) the drummer sets up vibrational modes in the skin of the drum

ii) these vibrations cause the air next to the skin to compress and expand in tandem with the motion of the skin. Thus the air pressure variations share almost exactly the same relative amplitude variations as a function of time.

iii) these pressure variations in the air are propagated outwards. However, due to energy dissipation, the amplitude of the pressure variations diminishes as the distance from the source (in this case the drum) increases.

iv) these pressure variations enter the ear canal of someone within earshot of the drum. The eardrum is moved back and forth by these variations causing vibrations whose characteristics (ie the amplitude as a function of time) are similar to those that were occurring on the drum-skin moments before (the time delay depends upon the speed at which the medium, in this case air, transmits acoustic waves).

v) the vibrations set up on the eardrum are conveyed to the inner ear via three bones connected together that each resonate in sympathy with the eardrum.

vi) these vibrations enter the cochlea which contains fluid and millions of tiny hairs which connect to the brain via nerve endings. The fluid propagates the acoustic wave which causes the little hairs to sway back and forth - the amplitude as a function of time by which these hairs are being moved back and forth is fundamentally linked to the that of the original drum-skin.

vii) the hair's movements trigger electrical signals which are carried to the brain which interprets them and enables the listener to experience the sensation of hearing the drum.



Now, let's introduce another element into this acoustic chain. Suppose the drum is being played within a soundproof booth, but that there is a microphone in the booth connected to an amplifier which is in turn connected to a loudspeaker outside the booth. Now the sound of the drum can no longer be heard directly via the air but it can still be heard thanks to the microphone setup.

What is happening?:

a) the acoustic waves caused by the drum are propagated to the diaphragm in the microphone in the same way as the sound of the drum reached the listener's eardrum in my above example.

b) the diaphragm moves back an forth with similar amplitude variations as a function of time as those of the drum-skin.

c) the diaphragm causes an electric signal to be generated due to the magnet/coil setup within microphone. This signal has voltage amplitude variations as a function of time that are similar to the amplitude variations as a function of time present in the drum-skin.

d) These voltage variations are amplified and then fed to the loudspeaker causing the cone to move in and out with similar relative amplitude variation as a function of time as the those of the original drum-skin.

e) the process from here to the brain of the listener is the same as from points ii) to vii) above.


So, the only difference we have between the two example above is that the energy propagated by the original drum-skin has been converted from mechanical energy to electrical energy and back again.

No one would disagree with the statement that the electrical signal in the second case is what we call an analogue signal. While every other part of the chain is normally referred to as just "sound" there is nothing fundamentally different about the continuously variable nature of the amplitude variations represented by the waves be they electrical or mechanical. In this way we can think of all of the waves in either chain from drum-skin to brain as analogue signals. That is, each wave is a continuously variable representation of the original vibrations set up in the drum-skin by the drummer.

Now let's think of the soundproof booth example again but his time let's introduce an analogue to digital converter in the electrical chain followed by a digital to analogue converter (note that I have not mentioned the idea of recording any of the signals in these example).

The introduction of the A/D and D/A converters means that some new entries are need between c) and d) above:

c') this signal is converted to digital by sampling the waveform discretely in time and calculating the amplitude of each sample creating a list of sample amplitude values.

c'') this stream of numbers, while representing the amplitude at the sample points of the input waveform, does not itself constitute a wave. The energy that has been propagated via different means since the drum-skin's motion began this chain of events is no longer in the chain. The energy has been dissipated but, of course, the sample amplitudes were measured before the energy associated with that part of the signal was lost and so we still have a way of recreating the energy variations of the original signal.

c''') the conversion back to analogue recovers as closely as possible the input analogue electrical signal and thus we return to the situation whereby we have energy being propagated in our chain.

At which point we can return to d) above.


So, the inclusion of a digital link in our chain introduced a fundamental difference from the other examples I have talked about here. That difference is that while in the other examples the energy of the motion of the drum-skin was propagated to the brain via various mechanical and electrical systems, in the example that incorporated the digital system the energy itself was exchanged for information about the energy at that point in the chain.

This is the fundamental difference between analogue and digital that I am talking about. A purely analogue system conveys a signal (eg the vibrations of a drum-skin) via the propagation of energy, be it mechanical or electrical, just as occurs in nature. A system that incorporates both analogue and digital components conveys a signal by using energy propagation within the analogue component and by using information about the energy within the digital component. This hybrid form of signal transmission does not occur in nature.


[As I have stated before, thanks to the sampling theorem (and thanks for this must go to Harry Nyquist), digital systems can be incorporated in analogue systems very effectively and I in no way mean to imply categorically that digital = bad.]

 

I respectfully disagree with this analysis. In the analog system you describe, the "energy propagation" stops at the amplifier. The amplifer's energy comes from the wall socket; the input is just a control signal. Of course the input signal necessarily has energy, but that's equally true of digital signals.

The brain isn't powered by acoustic energy, either. You didn't say it was, but it seems implied by the analog=energy propagation idea.

 

Maybe I don't understand how analogue amplification works then. I was under the impression that although the input signal was amplified (ie the amplitude was increased by a factor known as the gain) using energy from the mains, the energy from the input signal was still propagated through to the output signal too.

Afterall, if the energy fluctuations in the input analogue waveform aren't transmitted through the system (albeit with a proportional increase in amplitude depending on the gain controls) then how do we end up with a signal directly related to the input signal?!

I think you are missing the point here really. That is, in the analogue system the waveform is at all times represented by a continuously variable analogue of the original time-varying amplitude of the surface of the drum-skin. In the digital system this is no longer the case.

Of course, but I am talking about the different ways an acoustic wave reaches the listener depending on the system in place - the connection to the brain itself will never be changed in any of those systems so it is not a variable we need to consider in this discussion!

 

Very interesting opinions in this thread.

I understand that debating objectivism vs subjectivism isn't allowed on Steve's forums, and I'll respect that. But surely, if the difference between analogue/digital and redbook/hi-res is as pronounced as some people here think, it should be very easy to produce positive results in a double-blind/ABX test between the formats, no? You just might surprise yourself.

Heck, you might become the first person on the planet to actually prove you can hear a difference between 44.1/16 and higher resolution digital.

 

What does this mean? I think everybody that I have demonstarated my SACD to can "HEAR" the real difference between redbook and high-rez!!!

 

Reviewing this idea, I really don't think that it is such a good idea simply because of the bass issue. It has been proven time and time again that vinyl does not have the same bass extension as does digital, and if you make a DSD copy from a vinyl copy then you are wasteing a generation of resolution (copy of a copy). If you make a DSD copy of the original master tape then you will avoid one link in the recording chain and you will also get the bass extension that is present on the original master tape.

 

Ah, but this whole discussion got started with the assertion that sound was 'naturally' analogue and in this context I don't feel that the brain and how it processes sounds can be removed from the conversation.

 

[delete me - double post! ]

 

In my opinion any discussion on the nature of sound requires us to separate the two distinct processes involved - the first being the generation and transmission of acoustic waves and the second being how the brain receives and processes the characteristics of the acoustic waves.

These two process are completely independent in the sense that, philosophical arguments aside, acoustic waves are generated and transmitted irrespective of whether or not they are detected by the brain of a living organism. Similarly, the brain can give the sensation of hearing something without acoustic waves being present. This last point may seem a little odd but just think about your dreams - you routinely have the experience of hearing everyday sounds in your sleep despite the absense of the actual acoustic waves that normally cause them.

All I am trying to say with my contributions to this thread is that the natural processes involved in the generation and transmission of acoustic waves can be described as analogue (see earlier posts for details) and that any artificially contrived analogue system for propagating acoustic waves is fundamentally related to the natural process of acoustic wave propagation. In contrast, once you introduce a digital component into your artificially contrived system it can no longer be said that this system is fundamentally related to the natural process of sound propagation.

Of course, this doesn't mean that a system incorporating a digital component can't sound natural. However, I think it is fair to say that a purely analogue system will inherently produce a natural sounding result whereas a digital system will only produce a natural sounding result if the sampling theorem is rigorously followed (which is often not the case - but that is another thread....).

To avoid being misunderstood let me just clarify what I mean by "natural sounding" in the above paragraph. I do not mean to imply that just because sonething is "natural sounding" that it must follow that this means it will sound good! What I mean is that the sound will have all of the infinite variety we experience in natural sounds.

I think it is extremely misleading to suggest that a digital process is just as natural as an analogue one in the world of acoustic wave reproduction. The key point here is that a purely anlogue system relies on processes that are fundamentally the same as the processes involved in natural acoustic wave generation and propogation whereas a system incorporating digital processes does not.


As ever, I am not trying to say that this distinction somehow proves that digital components are not suitable in a sound reproduction system. However, I think it is vital to realise this distinction if we are to have a meaningful discussion about "analogue" and "digital" audio technologies.

 

Looking at the question empirically, the trend over the entire history of recording has been less "natural" methods leading to more "natural" sound. An acoustic (pre-electrical) recording certainly does not "have all of the infinite variety we experience in natural sounds". And yet an acoustic recording best fits your conception of "natural".

I don't agree that analog "will inherently produce a natural sounding result". Even in the HiFi era, analog is full of unnatural sounds. What's natural about tape overload distortion, for example?

 

You know Malc, this post seems to assert a lot of things that I find to be off kilter. In my understanding of the process of recording a sound that occurs in nature, (be it a thunder clap, a human voice or a motorcycle engine revving up), there are a lot of "conversions" that take place in the chain of events. The conversion of mechanical energy into electrical energy, the transmission of said eletrical energy through electronic circuits, the conversion of electrical energy back into mechanical energy, these are all processes that occur in the recording process. Now if you add a process in this chain of events, that being the conversion of electrical energy into digital information, then you are still achieving the same results; the recording of a sound captured from a physical source
So here is an example of the recording chain;
1. Mechanical energy captured by a transducer such as a microphone or a piezoelectric cell
2. Mechanical energy converted into electrical energy
3. Electrical energy stored on some some of recording medium such as magnetic tape
4. Electrical energy read off of storage device
5. Electrical energy amplified through a series of electronic circuits
6. Electrical energy reconverted back into mechanical energy by some kind of transducer device

This chain of events is what I assume to be the normal process of recording a sound in the "analouge" domain. But let us look at the same chain of events if one is to add the factor of digital into the equation.

1. Mechanical energy captured by a transducer such as a microphone or a piezoelectric cell
2. Mechanical energy converted into electrical energy
3. Electrical energy converted into digital information
4. Digital information stored on some some of recording medium such as magnetic tape
5. Digital information read off of storage device
6. Digital information reconverted into electrical energy
7. Electrical energy amplified through a series of electronic circuits
8.. Electrical energy reconverted back into mechanical energy by some kind of transducer device

Sure we have added a couple of steps to the chain of events, but since the chain of events already included the conversion of states of energy from one state into another state anyway, we have not really altered the big picture of recording, simply the process of storage. In truth, I really don't think that we can go straight from step '1' to step '3' without the intermediary step of converting the mechanical energy into electrical energy, no matter how hard we try. But even incuding this factor I really don't see how one process could be more "natural" that the other.

 

A couple of points. First, we can indeed go straight from step 1 to step 3 without the intermediary step of converting the mechanical energy into electrical energy. This is how all recordings were made during the first forty years of sound recording, i.e., acoustically. Second, it's not the number of conversions that matter, but how much information is lost (or distortion added) at each step. That's what this debate should be about, IMO. Digital and vinyl recording both lose information and/or add distortion at different steps in the conversion/re-conversion process. How far can the loss be minimized, and how offensive to the ear are the added distortions, in each case? That's the real issue.

 

Luke,

I think the main problem here is semantics!

I tried to clarify what I meant by "natural sounding" but I clearly didn't do a very good job. I have been reluctant to go into to much detail for fear of hijacking this thead but here is what I mean in a little more detail.....

In nature, any sounds that are produced are generated from a combination of factors so many in number as to be practically infinite. This is why, no matter how many times someone hits a drum the sound created each time will never be identical. Even if someone was to keep hitting the drum for the lifetime of the universe (OK - lots of reasons why this couldn't happen but bear with me!) the pattern of acoustic waves generated from each hit will never be identical - ie the sound may be similar each time but there will be infinite variety in the sound produced thanks to the chaotic nature of the factors behind the sound's creation.

You asked "What's natural about tape overload distortion, for example?". I'll tell you - it will never be exactly the same however hard you try to recreate the same distortion twice. And it's not just that the same distortion will produce very slighly different sound due to playback conditions constantly changing (due to temperature variations etc...), no, it's more profound than that. It is because the distorion itself is never the same twice. This is because the factors behind the distortion that occurs in an analogue system is generated by a hugely complex combination of factors. In this way, analogue systems produce audio with infinite variety built in at it's core and this is exactly the same infinite variety that exists in real world acoustic waves.

Digital processes do not have any variety in them whatsoever. This isn't necessarily a bad thing and, in fact, under the right conditions this can be used to our advantage. However, the point I was making is simply this - digital systems, unlike analogue systems, do not in themselves produce the infinite variety found in real world acoustics.

Realising this is the first step to understanding why not rigourously following the rules of the sampling theorem results in unnatural sounding audio from digital systems. In contrast, whatever problems there may be in an analogue system, the audio will still sound "natural" as I define the term above. I have gone into this at length elsewhere on this forum - if you are interested try there links:

1. Digital/Analogue debate part I

1. Digital/Analogue debate part II

3. vaguely relevant thread

As you can probably tell I love talking about these issues so if you don't mind your original thread topic being hijacked then I am happy to carry on if you are.....

 

Rocky,

you may have noticed that in my examples I specifically left out the added complication of the storage (ie recording) of the propagated waves from an acoustic event to simplify my arguments.

What you are basically asserting is that with the added steps your second example is just as "natural" as your first. I would argue that, on the contrary, while each stage in the first example has the characteristic of infinite variety (see my last post) inherent in any naturally occuring acoustic system, steps 4 and 5 in your second example most certainly do not. Therefore I would argue that your second example does not represent a "natural" chain if events.

Do I believe that because your second example does not represent a "natural" system that it must therefore be somehow wrong? No!

Do I believe that because your second example does not represent a "natural" system that there is a danger that without due care and attention unnatural sounding audio may result? Yes!

In your first example I am happy that audio resulting from such a system would at the very least sound natural - although this doesn't mean it has to sound any good (see my last post for more detail on what I mean by "sound natural").

 

Wise words .

I would add that understanding the nature of sound (ie the generation and propagation of acoustic waves) and how this relates to the processes involved in both analogue and digital audio is an essential part of the debate you describe.

 

Folks, I think we have gotten badly off point here. The main difference between analog and digital recording/processing is that digital, by design, does not even attempt to record (capture) all of the information. Yes, analog recording clearly has limitations (as does any recording/processing technique), but it attempts to record all of the information. The only way digital recording can work is to bandwidth limit the incoming signal (the Nyquist theorem). This is a direct consequence of not trying to record/process all of the signal. The debate is can human beings tell the difference between a signal (music) that is recorded/processed using analog techniques and digital techniques. There is one consensus of opinion that thinks that humans can clearly distiguish CD from analog, that DVD-A is closer to analog and SACD is even closer to analog and that analog being closes to live performance (i.e. no recording/processing and no digitization).