Analog vs Digital

Exploring the Digital Divide in detail.

Vinyl records (left) and music CDs (right) illustrate the analog-digital Yin-Yang.

A central feature of the difference between physical brains and computer simulations of those brains is the difference between analog and digital. These form a truly exclusive Yin-Yang pair.¹ It’s not so much they are opposites in a black-white, positive-negative sense, but that they differ as much as those. We might say digital and analog are mutually exclusive.

I think recorded music offers a good starting point in exploring that difference.

Nearly all recorded music today is digital. Live music is still (generally) analog,² but unless one makes the effort of using vinyl (or taped) recordings of analog performances and mixes as well as analog playback gear, recorded music is almost certainly digital music.

What happens to the music between the performers being recorded and the playback of that recording provides an excellent example of digital dualism, the divide between physical and abstract.

Analog Music

Before digital music, musicians sang and played instruments, both of which make physical vibrations in the air and nearby objects. Even performers using electronic gear ultimately need speakers to create air vibrations.³ When we listen to live music, these air vibrations move our eardrums, pass into our inner ear, and ultimately end in our brain. In a very real sense, the musicians push and pull on the air, and this is mechanically transmitted to your eardrums as if a string connected them to you.

The critical point here is that every part of the “string” vibrates in a way that directly represents the sound. Louder sounds have stronger vibrations, higher pitched sounds have faster vibrations.⁴

In analog systems, there is a direct correlation between the information content and the “shape” of the vehicle carrying it — which makes the information readily recognized.

Whether it be the vibrations of the vocal cords, the instrument bodies, the speaker cones, the air, or our eardrums, we can always recognize analog sound for what it is.

An oscilloscope trace of a brief instant of the jazz music I was listening to while working on this post. The trace shows the electrical audio signal for a tiny fraction of a second. This electrical signal from a microphone in my iPad matches the variations in air pressure that we hear as, in this case, a saxophone solo.

Recording

To record these vibrations, engineers use microphones that convert the air vibrations to electrical voltages that match them. These electrical vibrations pass from the microphone through wires to the recording console and eventually to a magnetic recorder that converts the electrical vibrations to magnetic vibrations frozen in a tape coated with fine iron particles (i.e. tiny magnets).

The tape can be played back at any time, and the magnetic vibrations converted back into the electrical vibrations. These are amplified to make them electrically strong enough to drive speakers. These move the air, reproducing the air vibrations of the music.

So, with strictly analog music, live or recorded, a series of mechanical systems acts as a “string” from the musician to your ear (and hence your brain). There is a very real physical connection from the musician to you. A physical connection that at all stages directly resembles the music it represents.

The key point here is that if we could examine the air vibrations, or the electrical vibrations, or the magnetic ones, we’d find they all resemble the music. We would be able to see when the music was loud or soft, low-pitched or high. We would see something that looks like the music.

Digital Music

With digital music, at some point the vibrations are converted to numbers. There are digital microphones that convert the air vibrations directly to numbers, or the analog-to-digital conversion (ADC) may happen later in the audio chain. Once it does, everything is numbers from that point until the digital-to-analog conversion (DAC) re-creates the analog voltages for playback in speakers or headphones.

When digital music began in the late 1970s, only the studio tape recording was digital. Everything else was analog, but the tape recorder converted analog signal to numbers and recorded the numbers. On playback, those numbers were converted back into electrical vibrations. This was because magnetic tape and vinyl records are the most limited parts of the audio chain in terms of audio range and fidelity. Replacing the magnetic tape with a digital system removed a major source of audio distortion.⁵

The M79 32-track digital recorder introduced by 3M in the late 1970s.

Digital expanded to the point it now includes the entire audio chain.⁶ Home systems can be digital until the numbers are converted to analog electrical signals to drive the speakers. If the music was recorded with digital microphones and processed digitally throughout, this music would have the absolute minimum of generational loss and — electronically speaking — would be the most accurate possible.

Yet many audiophiles dislike digital music.⁷ Unlike 8-track tapes and cassettes, vinyl records never died and might be said to be making a bit of a comeback. More and more artists are releasing vinyl copies of their work. Ironically, vinyl has the worst audio characteristics in terms of frequency and dynamic range. And merely handling it, let alone playing it, can degrade the record.⁸

Closeup of the grooves of a vinyl record (analog, because we can see the music).

Why do presumed connoisseurs still favor vinyl? I think part of it is psychological. Those who grew up listening to vinyl recordings have trained their brains to hear music that way. The brain can learn the difference between live and vinyl and adjust our comprehension of recorded music accordingly.⁹ But when presented with what is technically a more accurate recording, the mind may reject the experience as “wrong”.

That said, digital recording does alter the music in ways that analog recording does not. For one, it can introduce digital artifacts with no connection to the music. Converting to digital requires cutting music into tiny slices of time, and the size of those slices is a critical factor in fidelity. But there is a tradeoff in data bandwidth. The smaller the slices, the more data there is to handle. This compromise may account for the dislike some have for digital music.¹⁰ They may be hearing something real, be it added artifact or altered audio.

Analog recording does introduce distortion, but almost always linked with the original sound in some way. As mentioned above, some recording producers like the soft, warm distortion overdriving magnetic tape provides. And some recording engineers like the more subtle warm distortion tube-based systems provide over solid-state systems.¹¹

It’s All Numbers

An important disconnect between reality and digital lies in the question, what numbers? When music — in the form of analog electrical vibrations — is converted to numbers, exactly what numbers do we use? How big should they be? How often do we “take a sample” — slice the air pressure at that instant into a number? Are we using absolute values or relative values? How do we encode these numbers?

For instance:

00092d20: 210c 542a 243e 97de f405 cab3 530f bc47
00092d30: c05d 2093 890b 5e37 8b11 41f9 2e55 1a96
00092d40: 7fe2 3fee 210c 5461 043f 1027 d0f9 0420
00092d50: d050 0e21 0c54 1c0d 9684 1501 8420 e7d4
00092d60: 82c0 0da0 9850 bc02 72fd b8f0 a028 46ab
00092d70: 6a77 a1e6 b2a2 d82c e6ab 7122 fe01 f9cb
00092d80: 4a21 b020 c7e4 7c85 7e51 7567 af0a 7ddf
00092d90: 3b70 e91f e9f7 cfa3 4152 5a13 e3f5 b4c9
00092da0: fca5 d253 ce46 0846 1010 83c4 20d7 b2f9
00092db0: 0ed9 b680 89c0 210c 540d 8e99 0862 2184
00092dc0: c0a7 c66a e586 0deb 241b 86d0 1599 8166
00092dd0: 63b0 24f0 5cc7 5244 c115 a784 d988 f2ad
00092de0: 2aad 76cd 019b 0d92 3340 370e 2b3d 3b38
00092df0: 23aa d6c9 5436 e33d 0543 1b08 afd0 a80d
00092e00: bcb9 e44c a844 ba74 3213 8371 8e58 e05b
00092e10: 35ed 0789 ed7a 5c74 2cef 532e 3d39 36e7

Is the above: [1] random gibberish I whipped up with a simple Python script; [2] part of a Windows program; [3] part of a YouTube video; or [4] part of Amy Winehouse’s Back to Black? It could be any of the above. From the context and phrasing, obviously the last one, but there is no readily apparent way to know those numbers are music, let alone to “name that tune”. In contrast to analog, we can’t directly see the pitch or loudness. No aspect of the music itself appears in these numbers.

There are many decisions associated with this “mapping” of music to numbers, and what matters most here is the arbitrariness of those decisions. As described above, when we look at analog music anywhere along the chain, we see the music. When we look at digital music, all we see are numbers.¹² What those numbers mean requires knowing the map. Without the map, the numbers are essentially random.

This is exactly what the term abstract means. The numbers represent music according to an arbitrary scheme defined by audio engineers. The only connection to the music is through that external map. In contrast, analog systems directly and physically represent the sound throughout the audio chain.

In short, analog systems are physical, digital systems are abstract.

Digital Dualism

Because they are abstract, digital systems require a physical implementation layer. This means digital systems have two layers: the abstract digital layer and the physical implementation layer. This forms the digital divide, the dualism I mentioned in Digital Emulation.

This dualism enables a key characteristic of abstractions: that they can be physically realized in multiple ways. A simple example is the calculator. All the different calculators, mechanical and electronic, are physical implementations of the abstraction — the algorithm — of “simple calculator”.

In future posts in this series, I’ll dig more deeply into the digital divide and into what I think are important differences between computation (which is digital) and evaluation (which is analog).

Until next time…

1. As opposed to “cup” Yin-Yang pairs such as empty-full that are comprised of a thing and the lack of that thing (hot-cold, light-dark, and so forth).

2. Most synthesizers, and many audio effects boxes, are digital. Parts of the sound system may also be digital.

3. And sometimes floor and wall vibrations.

4. In fact, what we consider “sound” involves vibrations too fast for us to perceive as vibrations (we perceive them as sound).

5. Ironically, some musicians and audio engineers like the soft distortion of tape, and some bands deliberately overdrive tape to get that soft distortion.

6. Which has a lot of advantages, but many audiophiles still resist it. Oddly, the first digital recordings were of classical music with the intention of unprecedented audio fidelity that audiophiles were expected to love.

7. I have a friend who will only listen to uncompressed digital music (WAV files but not MPEG files). Compression does add considerable distortion.

8. I’ve heard of systems that use a laser to probe the record groves so — as with a CD — there’s no physical contact with the media.

9. Somewhat similar to how our eyes adjust to the color temperature of indoor versus outdoor light.

10. Or had. I suspect newer generations raised on digital music don’t care as much.

11. Because tubes emphasize the even harmonics while solid-state emphasizes the odd ones.

12. Specifically, just lots and lots of 1s and 0s.