Heavy Vinyl Production And the Unpredictability of Random Processes

Those in the business of producing the highest quality remastered recordings on LP are crashing smack into a problem endemic to the manufacturing of the vinyl record — randomness.

Record producers can control many of the processes (variables) that go into the making of a high quality record. But they cannot control all of them. The word for such a situation, one with random, uncontrollable aspects, is “stochastic.”

Taking the liberty to paraphrase Wikipedia liberally, we would explain it this way.

A stochastic, or random, process, is the counterpart to a deterministic process. Instead of dealing with only one possible way the process might develop over time, in a stochastic or random process there is some indeterminacy described by probability distributions. This means that even if the initial condition or starting point is known, there are many possibilities the process might go to, but some paths may be more probable and others less so.

In other words, although some of the variables can be controlled, there will always be some element of randomness that makes the final result predictable within limits, but not predictable precisely.

A thought experiment related to the classic Drunkard’s Walk is illustrative here. Imagine a drunkard walking down the sidewalk. On his right is a wall, on his left a gutter. He can’t walk very well, being drunk, so there is a good chance he will bump into the wall. Bouncing off the wall he will probably not be stopped but instead be able to continue on his way. However, if he falls into the gutter he will very likely not be able to get back up and will end up stuck there.

We know very well how the wall works, and we know very well how the gutter works, but we simply cannot predict how far down the street the drunkard will get before he falls into the gutter because of the unpredictable nature of his path. The average distance can be determined experimentally easily enough. But each time he starts his walk there is really no telling how far he will get.

Record making is like that.

We can put new shoes on the drunkard so that he stumbles less. Likewise we can sweep the sidewalk of debris that might cause our drunkard to slip and fall.

What we can’t do is specify the path the drunkard will take, just as we can’t insure that those pesky vinyl molecules will align themselves in the grooves of the finished record the way we want them to.

Which means that a small portion of the pressings will contain something approaching 100% of the musical information originally inscribed on the acetate. Most will contain a smaller amount, and that group will be followed by a very small group that will contain a much smaller amount. These groups will all be distributed in the classic bell curve fashion.

Testing the Hypothesis

Recently we had the opportunity to crack open two brand new sealed copies of a recently remastered Heavy Vinyl pressing. We were told that with this record every effort was made to produce the highest quality product and to maintain the highest quality control throughout the production process, such that every copy produced, from first to last, would be more or less identical. Not just in terms of surfaces, but sound quality too.

To accomplish this feat the producer used the real master tape (we were told), had a top mastering engineer do the mastering at a highly-regarded studio, then had a well-known audiophile pressing plant in Germany make the record, using the highest quality vinyl compounds available, in presses that meet the highest standards in the industry, operated by highly skilled professionals.

Long before any stamper could possibly be worn out it would be replaced. All the metal mothers and stampers would be made in a way designed to eliminate any possible variation. One and only one complete run would be made; if another was needed at some later date the whole process would have to be started over from scratch using the same strict quality controls.

No corners would be cut, nothing would be left to chance. Each one of the finished records would reflect the exceptional efforts brought to bear at every stage in the process. Every copy would be quiet, the sound would be of the highest audiophile quality, and, more to the point, every copy — from number 001 all the way to number 7,500 — would sound as good as any other.

Our Experience

As you might suspect, our opinion as to the possibility of these results being achievable is that they are not.

Bear in mind that this is an opinion supported by the playing of thousands and thousands of records, including sometimes more than a hundred of the same title, and having them all sound different.

So we proceeded to test the proposition that by exercising maximal control over all the known variables of record production, using the most exacting standards at each and every step, two copies of the same record, chosen at random, would sound the same.

We picked a song, cued it up and listened to it for a minute or so. Then we put the other copy on our table, cued up the same track and let it rip.

Falsified in Fifteen Seconds

Immediately the sound was different and, importantly, quite a bit better. The first big cymbal splash was brighter and more life-like, with more extended high frequencies. The bass was better too: deeper, as well as more solid and easier to follow. All of this was evident within the first fifteen seconds of playing the second copy. So much for controlling the variables.

The random variability inherent in the record making process cannot be overcome by best practices and high standards. The process is complex, not well understood, and surely stochastic; some parts of it can be controlled but not all the parts of it can be controlled, which means that the finished product will have some unavoidable element of randomness.

Learning to Listen

Critically auditioning literally thousands and thousands of records during our 35+ years in the record business has convinced us that the only way to find good records is to clean them and play them. Not one at a time, but by the dozens. That single, simple approach, tedious and time-consuming though it may be, has the potential to teach you more about records and sound than any other.

We might even go so far as to say that no other approach to learning how to listen critically — none that we are aware of anyway — works remotely as well.