How Does a Vinyl Record Even Sound Good at All? Talking With Lewis Hopkin, Vinyl Mastering Engineer, Part Two

How Does a Vinyl Record Even Sound Good at All? Talking With Lewis Hopkin, Vinyl Mastering Engineer, Part Two

Written by Russ Welton

We began our interview with Lewis Hopkin of Stardelta Audio Mastering in Issue 164. Here, he details the factors which are essential to perfection in mastering a vinyl record, in this second part of an exclusive insider interview series.

Lewis Hopkin: Let’s break down the different stages that genuinely make a premium record.

Mastering and Cutting

This stage is completely dependent on the skills of the mastering engineer. This is real “make or break” for the resultant overall quality. The professional operation of the equipment, critically, combined with the understanding of the musical imperative of what is being put on the disc, is essential. Technical corrections have musical consequences. This [should be] handled to minimize any changes to the intention of the music.

Cutting music onto a lacquer is not a 1:1 process. The medium reacts audibly to the musical input and the playback reflects this. You could refer to this as the “sound” of vinyl. It’s the opposite of a digital recording in the fact that with digital what you record and play back should sound substantially the same. With a cut, this can be very different, and [it] relies on lots of variables. The level at which you are cutting has a big bearing on how the medium responds to the signal you give it. 33 rpm and 45 rpm sound different. For example, at 33 rpm the lateral tracking motion of the playback stylus being somewhat slower creates the illusion of more bass. I say illusion as most of this is just the self-noise of the stylus cantilever picking up the rumble at a lower frequency and feeding that into the music on playback. There are numerous things like this that can only be ascertained by test cutting and careful listening and making corrective changes as required.

Cutting equipment has to be calibrated so precisely to minimize any errors made by the engineer in the EQ decision-making process. This entire subject could be a book in itself in terms of [the] magnitude and breadth of the topic.

Lewis Hopkin.

Lewis Hopkin.

 

RMS, or Average Level

In simple terms, the higher the average level, the more space on the record will be used.

More dynamics equals a lower average level, but often a higher peak-to-average level. This means loud dynamic sections followed by quieter passages. Technically this is more problematic to cut but uses less disc space than [using] a simple high average level. Lots of vinyl mastering decisions are based on these calculations.

It’s physically harder to fit a consistently higher-amplitude recording on vinyl. Why? Because you need more space as the pitch (or lateral spacing on the disc) can’t be varied to [cut to] a finer pitch and has to stay coarse, in order to mitigate the grooves running into each other with a lack of land [the space between the grooves].

On a recording where there is less dynamic range, you have to use more disc space if you have a higher average level. As a comparison, if you have a heavily-compressed version and a hardly-compressed version of the same thing, you could have as much as 25 percent difference in space utilization. A highly-compressed record may be easier to cut in one sense; ([it will have a] more predictable [overall] level), but it will be compromised in that it will end up with a quieter cutting level, as the RMS level means [there will be] no choice but to drop the level to fit the music on the side. So, a more dynamic record will always give you more choices as a vinyl mastering engineer [when doing] the cut. Having said this, if the original intention of the music was to be heavily compressed with low dynamic range, then that intention can be translated to vinyl, but requires skill in a different respect in order to fit the program length onto the side at a decent level. As always, real vinyl mastering is a delicate balancing act between the technical and artistic demands of the job.

Sibilance

You may be forgiven for thinking that vocals naturally reproduce very well on vinyl, but there is almost certainly nothing more difficult to get right when it comes to vinyl cutting and mastering. The cutting stylus and playback stylus are different shapes and sizes. The cutting stylus can cut groove geometry that even the best stylus can’t accurately track. So, if you cut a sibilant vocal, then you create a saw-toothed element to the groove that the tip of the playing stylus simply can’t trace accurately. It skips across [the] tops of the teeth and produces a ripping distorted sound.

The solution is to slow the ramp of the sawtooth down by using high-frequency limiting, or [some kind of] de-essing. If de-essing the record (at multiple frequencies on different bands with varying degrees of subtlety) is going to significantly destroy a good part of the [music] just to make it technically correct, I will always ask if the recording can be de-essed in the mix, to provide a better master to begin with. If that isn’t an option, then a very painstaking process of manually slicing out the sibilant parts from the audio and processing just those parts is undertaken.

At a pressing plant or mastering studio [that doesn’t pay attention to this], a record cut by an [inexperienced] operator will just have the high frequencies removed, or heavy-handed broadband HF limiting added to attempt to stop the tracing distortion. An [experienced] vinyl mastering engineer will do this right. It takes a long time. The result is fantastic. Say no more!

 

Stardelta Audio Mastering. From the Hutchinson Homes & Gardens website.

Stardelta Audio Mastering. From the Hutchinson Homes & Gardens website.

 

Stereo Bass

It’s physically possible to [cut stereo bass] to a degree but let’s be honest here. Stereo bass… I mean it doesn’t sound pleasant at all anyway so it’s one area I think we can all live without on vinyl or digital! Let’s define that. I don’t mean [having] some low-frequency content in the sides, that is going to somewhat present in a lot of multi-microphone recordings. That most certainly needs to stay, as it’s part of the sound and it can be cut with no technical issues if it’s handled right. I’m talking about synthesized bass with stereo effects. That stuff you can’t cut as physics won’t allow it, but a version of it can be cut with [the] bass [being] mono below a certain frequency. If it’s handled right, it won’t be obtrusive and can be made technically right for vinyl without destroying the original artist’s intention.

Out-of-Phase Content

A record still has to still contain out-of-phase content to some degree; this will inform the sound of the record. A lot of the texture and color of recordings comes from the wandering phase of the content. A skilled mastering engineer can make that work well on vinyl with no compromise and keep the integrity of the vision, despite this being much more challenging to cut, and requiring careful test cuts, with visual inspections with a microscope of the grooves being cut.

Hysteresis in Metering and Listening With Your Eyes…

All of the metering of the stereo signal, particularly on the lathe, will be prone to hysteresis (the difference between the signal and what the meter is showing), and not tell you what is going on precisely enough to make decisions [based on metering] alone. A proper vinyl mastering engineer will use metering tools as guidance, but [also will be] test-cutting the difficult audio passages, and inspecting [the grooves using a] microscope. The test cuts will also be listened to using a very average consumer grade cartridge and stylus to ensure that the engineer hears what the average listener might hear. Also, this should be checked with a more high-fidelity playback system to ensure it meets both criteria. Vinyl mastering is the last bastion of a really physical approach to audio engineering. An obsession with metering and listening with your eyes is going to land you in big trouble here. A “360-degree” approach is needed.

Inner Diameter Decisions

A record decreases in fidelity as the stylus travels towards the center. The groove is one continuous concentric spiral. The reason fidelity decreases is because the [distance the stylus has to travel] on the inner diameter is over three times shorter in length than the length [the stylus travels on] the outer diameter of the disc. Quite simply, the record is spinning at the same speed, but at the inner diameter the same duration occupies a shorter length.

It’s pretty much the same as with analog tape speed. A faster speed increases the length (or tape) covered by the same duration [in time], leading to greater fidelity. This is why the inner diameter of a 45 RPM record has better fidelity than a 33 RPM record.

How far do you want the audio to spread [physically]? How many discs can the LP be spread over? The inner diameter decisions are sometimes taken out of your hands. If you can only cut a single-disc LP then the decision is already made for you. However, if you take the decision to spread [the music] over several discs, those compromises are then taken away.

The track order could be changed if there’s a track that is better-suited to being [located] on the inside; that is, something that is gentler such as an acoustic guitar piece or ambient drone [music]. If money is no object, a double-vinyl album instead of a single LP is better for fidelity, but is also significantly more expensive in all the reproduction costs incurred with spined sleeves and gatefold options, and you have to keep getting up to turn it over! Not always the right decision.

Cutting Levels vs. Program Length

This is a decision based on what is possible. The longer a record is, the quieter it must be. That is just physics – however [it will] not necessarily be of lesser quality if done correctly. How long is too long? I feel that in order to minimize compromise in level, 22 minutes of music on one side of a record is as far as it should go. It’s technically possible to cut quite a bit longer than this, but my feeling is that beyond 22 mins the compromises stop being worth it.

It’s a real skill to cut a 25-plus minute side and have this really sound fantastic and be technically good in terms of making a great record. It’s something I have to do regularly, and most of the time it’s going be good, but very quiet in terms of cutting level. But ideally, no longer than 22 minutes on each side at 33 rpm.

The absolute success of cutting great records relies inherently on the proper calibration of the cutting equipment. It also relies entirely on the signal path at every signal stage, including the digital-to-analog conversion journey (or tape path) right through to the cutter head and through the stylus into the lacquer.

What I have built in-house at Stardelta is above the highest tolerances of the original manufacturer’s standards, which were already exceedingly high. I have modified nearly every system on my two lathes here, and the entire signal path is modified or built in-house to a standard you could not buy off the shelf.

Cutting stylus lifespan is also a major consideration at a professional mastering facility (not [at] a [typical] pressing, plant so I hear). So, we constantly listen to our cutting styli during test cuts, and the quality of the cut as the stylus hours progress. Styli do become noisier as they are used, due to wear on the facets, and stuck-on debris. We clean the stylus before and after each cut and inspect it in the microscope. Styli are expensive these days, but we don’t get high hours out of them, as we can hear issues after 10 or so hours normally. We measure stylus cutting hours to quantify [their] lifespan but not arbitrarily, always in conjunction with our critical listening for optimal performance. I’ve heard stories of people using styli for 50 hours or more which quite frankly sounds horrifying to me…I can’t imagine that based on what we hear that could be even vaguely possible if you wanted any level of great fidelity!

I think a lot of people just assume that if the stylus just cuts what looks like a clean groove and sounds like it has a bearable amount of groove noise then it’s still ok…Well technically it is ok, but it doesn’t sound the same as when [the stylus] had fewer hours on it. Even the best styli get short shrift at Stardelta. Why? Because, we really are listening.

******

In the next part of this series, we will consider some more of the inconvenient truths of creating a perfect record. What’s the most difficult part of the record to cut and press? What effect does the electroplating part of the process play? What are “groove fry” and stylus heat? All this and more will be revealed.

Header image from the Hutchinson Homes & Gardens website.

Back to Copper home page

1 of 2