Audio Mastering


I thought I’d share what I learned in the process of having my Canons album mastered. First of all, what is audio mastering? The credits of almost every album say “mastered by so and so,” but what does that mean? For much of my listening life, I’ve had the vague notion that mastering is about refining the sound of album before you release it. In fact that’s basically right, but there are some important distinctions to make.

First of all, mixing is when you have multiple audio tracks relating to the same song (like the outputs from different microphones positioned around a performer or distributed across the performers in an ensemble) and you want to combine them into a single track. Mastering happens after mixing, once all the inputs have been combined and you’re ready to optimize the result as a whole. In the case of Canons, my collaborator had already mixed the inputs from the multiple microphones that were positioned around his harpsichord during the recording sessions. This mixing was done at recording time, without preserving separate outputs from the individual mics, so there was no option to do further mixing, only mastering.

Mastering can be done on a single song, but one of the specialties of mastering engineers is that they take a comprehensive view of an album and make it flow consistently from song to song. For this reason it’s important to finalize your song order before you start a mastering session. The engineer is going to make sure the transitions are smooth for the specific order you’ve chosen.

Since my Canons album includes a whopping forty-five pieces it was quite a challenge to order them. The ordering process turned out to be one of those time sinks that catch you by surprise in a project right when you think you’re nearing the finish line. I considered grouping the canons by style (traditional vs. contemporary), or grouping them by affect (bright vs. dark), or trying to sustain a sense of contrast and variety throughout. I chose that last approach. While the pieces employ a range of tonal centers and many of them take significant trips away from their initial tonality, I wrote quite a number of them with C as a starting pitch, mostly for notational convenience, and not really thinking about the aggregate effect. Now considering how all the pieces worked together in an album, it took some effort to break the monotony that arises when you’ve got several tracks in a row that all begin in the same place. The bigger challenge was that to really get to “know” how an order works, you’ve got to listen to the whole album with sustained focus; when you make a change to the order you can try to preview its effect by just listening to the new transitions that are created, but to do it right you’ve got to rest, come back with fresh ears, and listen to the whole album again, and there’s no way to rush this.

Back to mastering itself: the process depends on accurate hearing, which depends on accurate playback. You want to get the clearest, least biased sense of what’s in the recording so you can adjust what’s actually there and not what seems to be there when you listen with biased equipment. For example, a recording might seem bass-heavy when you listen on your home stereo, but that might be because your stereo itself is bass-heavy; if you reduce the bass to get things sounding right on your home system, you’ve made an adjustment that’s not ideal for someone else’s system. An aim of mastering is to get the album sounding good on as many systems as possible, which starts by knowing the album’s “true” balance.

This is one of those things in life where equipment really matters. Whereas a blues guitarist can make amazing sounds with a $10 guitar from a pawnshop—no need for fancy gear—a mastering engineer probably does need a sophisticated (and likely expensive) studio to do what the job demands.

In the photo at the top of this post, I’m sitting in the engineer’s chair at Jeff Lipton’s studio, Peerless Mastering, in Newtonville, MA. Jeff describes the studio as an extension of himself, so it was a privilege to sit in his chair. As he tells it, Jeff got started in mastering back when he had a totally different career in software. He started helping some friends make their recordings sound better and he kept going. Soon, all the proceeds from his then day-job were being used to buy mastering equipment and develop studio designs. The current incarnation of his studio was planned by two acousticians and the walls and furniture were built to specs that had to be accurate to within an eighth of an inch.

I’m sitting at the one point in the room where my head makes an equilateral triangle with the tweeters of the left and right monitors (only the left is shown in the photo). This is supposed to be the one point in the room where you can hear the most accurate and uncolored version of what’s being reproduced. Take a step back and indeed, it sounds different.

One of the things that helped our session go smoothly was that I had chosen a strong and representative opening track: the canon Tourmaline. We got Tourmaline sounding great in the first hour. Throughout the rest of the session, the engineers (Jeff with assistance from Maria Rice) continued referring to that opening track as a benchmark.

The engineers went through a process that I found quite startling as a newbie. They’d play one or two seconds from the benchmark track and suddenly switch to a random point in some other track to see how these two unrelated samples compared. Having spent a few years of my life slaving over these compositions, it was fascinating and jarring to hear bits and pieces of them juxtaposed in random orders, as you can hear in this clip I recorded with my phone. Tourmaline, one of the most energetic tracks on the album, is being compared with one of the slowest and most contemplative, Lapis Lazuli:

Before I approached Jeff about this project I had assumed that mastering a harpsichord album would be fairly straightforward and that the same “settings” could be applied to all the tracks. The harpsichord is a dynamically uniform instrument where the performer has no control over the volume of each note, and besides, all the recordings were made with the same instrument in the same church with the same microphone configuration. But in our first phone conversation, Jeff asked whether the recordings were made on the same day with the same weather conditions. The answer: no. Did all the pieces have the same mood? No. So there would still be subtle differences from track to track that should be considered in mastering. Would there be enough variation to warrant more than an hour or two of work? Our session lasted twelve hours: I arrived at the studio at 3pm—this was Fri Dec 30, 2016—and left at 3am the next morning!

This leads to something else I learned during the process. I used to think that in an alternate life, if most of my time weren’t spent grappling with music—the writing and playing of it—I’d enjoy being an audio engineer. But now I don’t think I have the stamina for it. As much of a thrill as our twelve-hour session was—I’d call it my very best moment of 2016—I periodically found myself exhausted and losing any ability to hear the subtle differences that were being considered. To be able to do twelve-hour sessions day after day with attentive ears and a focused mind, that strikes me as an Olympic kind of ability. As we were nearing midnight on the day of our session I hinted that I might go home as it seemed the engineers had everything under control by then and I probably wouldn’t be of much more use feedback-wise. Maria joked, “So you mean you don’t want to stay here till 3am?” That’s their life in the studio.

How did the album improve in this process? Basically it became brighter, more present, more “alive.” It is almost as if someone with a beautiful voice but a mild cold suddenly overcame the cold and revealed the best of how they can sound. When I listen to the mastered album on a decent set of speakers I feel like the harpsichord is closer to me—the sound is more immediate and intimate—but yet seems to resound more fully. Matt said that the mastered tracks sound much closer to the way he hears the harpsichord as he’s performing. He also feels that certain subtleties in the musical compositions which were previously in the background have now moved to the foreground

What are technical secrets of how this was achieved? I don’t really know. I made a point not to pester the engineers with questions about all their equipment and how they were using it. Though they seemed willing to talk shop, I wanted to stay focused on the actual audio comparisons rather than the underlying mechanisms.

I did learn one technical thing: in a good mastering room, changing the playback volume by a fraction of a decibel can have a dramatic impact on how things sound. Sometimes you’ll think that a more substantive effect was applied (like a dramatic change to the equalization) but all that really changed was the volume. Mastering engineers play subtle tricks with volume alone, boosting a track by a quarter of a decibel to give it more prominence in an album, or reducing it by that much so it doesn’t stand out. The effect is subliminal since most listeners cannot consciously detect volume differences that small (unless they’re in a controlled studio environment doing A/B comparisons) but nevertheless these difference do have an impact on perception. One of the questions we considered for each track was whether it should occupy a more active position in the album or one of more repose.

Overall I’m quite proud of the sound we achieved on this album, which is to say I’m proud of Matthew McConnell’s work in performing and recording it, and proud of Jeff and Maria’s work in mastering it. It strikes me that we arrived at a harpsichord sound that’s different from the norm.

I think the norm in solo harpsichord albums is to do an up-close recording that emphasizes clarity and detail while minimizing the room feel, the distinct “voice” of the performance space. Some albums in this vein err on the side of being clinical. Now I actually do like this more typical sound, this crystalline but almost sterile sound where you feel your ear is up against the instrument’s strings which are vibrating in a neutral space, but as it happens, our sound in Canons stands quite apart from that. Matt’s recording setup has captured lots of the hall resonance at First Baptist Church in North Adams, MA. Because the hall itself is “speaking” so prominently in these recordings, the harpsichord sounds more massive and resonant than I’m used to hearing it on record.

I flipped through my collection of Bach harpsichord recordings and took note of their sonic personalities. Since I hadn’t listened to many of these in a while it was like becoming reacquainted with old friends. One of the first Well-Tempered Clavier albums on harpsichord I ever heard was Kenneth Gilbert’s and I remembered it as having a particularly bright and brilliant sound, which it does. I’ve got a set by Bob van Asperen. A set by Peter Watchorn. A set by Colin Tilney. A set by Ralph Kirkpatrick. A set by Rebecca Pechevsky. (I’m not counting what’s probably my favorite Bach album of any, the Landowska WTC, because it’s an old recording on a Pleyel harpsichord and not easy to compare with anything else.) In listening again to all these amazing albums, the recorded sound struck me as admirably crisp in most cases, and often quite resonant, but still thinner than the sound I had come to know in working with our canon recordings for so many months. Not considering the music or performances but focusing only on the sound of the harpsichord in these recordings, I feel Canons captures the instrument in its fullness. Not better or worse, just something less commonly heard, where the instrument has a more supported voice—where you really sense the power of the instrument resonating, singing through a sacred space. When we began the mastering process I told Jeff I thought the prominence of the hall in the recordings might pose a challenge—something to be tempered or restrained—but in fact I think it became a defining asset of this particular view of what the instrument can be. And it turns out to be an appropriate sound for our album, because the music consists of only two contrapuntal lines—a thin texture which blossoms when rendered in a voice of considerable weight.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s