Some of our readers will be old enough to remember when sampling arrived on the music scene, and drum machines, for the mainstream it was in the 1980s.

When both appeared some were up in arms about the impact of these technologies on the livelihoods of music makers. Some predicted the end of drummers, orchestras, in fact any kind of live musical performance.

As time has shown none of those doom-laden predictions came true. There are still plenty of drummers, orchestras and live musicians appearing on recordings.

Now it seems that the latest predictions about the demise of a craft are directed at automated online mastering services. This article is not aimed at discussing the merits of these systems, instead we want to suggest why we believe that automated mastering isn’t likely to kill off pro mastering, in fact it may well grow the market.

What Is Automated (AI) Mastering

Perhaps the first significant online mastering process to arrive on the scene and make a splash is LANDR, suggested by some as the devil incarnate. LANDR (pronounced lander) is the product of a Quebec-based AI company MixMagic founded in 2014.

Since the early days of LANDR it has grown from online mastering to offer a more complete package;

• 2014: AI mastering

• 2017: Music distribution

• 2018: Music samples library

• 2019: Rent-to-own VSTs and plug-ins

• 2020: Collaboration tools & services marketplace

• 2021: Proprietary VSTs and plug-ins

For LANDR the big battle has always been to establish credibility amongst the pro community, latterly it has received more positive feedback and has been used on top projects. Short of time before the Grammys, Gwen Stefani’s production team used the web-based program LANDR to master the song “Make Me Like You.”

More recently Plugin Alliance Founder, Dirk Ulrich, announced Mastering Studio, another AI-based mastering service. Given their track record for delivering great software, we have little doubt this service should offer an excellent result and prove to be popular, given their existing customer base.

Automated mastering uses AI (artificial intelligence) to simulate the decisions made by a mastering engineer. With those algorithms, it aims to deal with EQ, compression, and loudness, preparing tracks for final distribution either via online platforms or more conventional formats such as CD and vinyl.

What Automated Mastering Isn’t

Without wishing to state the obvious, it’s not a human. It’s a computer making its best guess about what to do with your music. It’s not going to sit down with you and have a chat about the tracks.

AI features in all parts of our lives, be that Siri, Google or Apple Home devices. AI uses machine learning to get better at what it does, the more data it has to work with and the more feedback it gets. However, as with any input based system, bad data in equals bad data out, there are limitations. That’s why we don’t chat with Siri all day, in fact most of the time we are shouting at our AI services as we grapple with their shortcomings.

Automated mastering is great for automating many of the audio processes mastering engineers undertake, but don’t confuse process with craft. That’s not to say automated mastering is bad, in many cases it does a great job, but it’s working with rules, not exceptions. In other words, all AI based systems work in generalities not specifics.

To unpack this a little more, let’s use the example of AI-based mixing plugins. You put one on a track and it’s smart enough to know you’ve put it on a bass guitar track. It should be able to spot the key and tempo you are playing in, that’s easy. It might be able to work out what kind of strings you have round or flats. It might also figure out from the timbre of the sound if you’ve used a pick or a finger style. However it won’t know it’s a 1970s MusicMan being played by a bass player who tends to play the third string harder than the rest and who has intonation problems on the 12th fret plus some rattle off one of the frets.

Most of the time the plugin will do a great job and some of the specifics won’t matter, but some times they will. If a real person were mixing it they might spot issues and take action to deal with any problems they may bring in the mixing process.

AI isn’t about specifics, it’s about generalities.

Writing for the MIT School of Engineering Carolyn Blais writes;

“In a paper published last year, titled, “When Will AI Exceed Human Performance? Evidence from AI Experts,” elite researchers in artificial intelligence predicted that “human level machine intelligence,” or HLMI, has a 50 percent chance of occurring within 45 years and a 10 percent chance of occurring within 9 years.”

and…

“And of course, there’s more. What about the relationship between creativity and intelligence? Scientists in Germany have trained computers to paint in the style of Van Gogh and Picasso, and the computers’ images aren’t all that bad. But, is teaching a machine to mimic creativity true creativity?

When it comes to raw computational power, machines are well on their way. And there’s no doubt that they will continue to make life more pleasurable and easier for humans. But will a machine ever write the next Tony Award winning play? Or break into an impromptu dance in the rain when an unexpected shower strikes?”

Secondly, don’t confuse AI based mastering services with the online services offered by studios like Abbey Road or Metropolis. The services offered by big studios or online mastering engineers are using real mastering engineers, often with a list of credits to die for. The only online part of the process is that you are sending your audio to them via a file transfer service and then receiving their work back in the same way.

Online mastering services are not AI mastering services.

What IS Mastering?

Check out an extract from an article by mastering engineer Mike Thorne (The Brand New Heavies, KT Tunstall and Procol Harum);

Well … it depends. The meaning of the word has gone the same way as the definition of a record producer. It’s developed into a catch-all for much more than just creating production masters from a mix.  

Originally, mastering was a technical process - the last step before the physical manufacture of a record - cutting a vinyl lacquer from mixes on tape. EQ was applied to compensate for the limitations of vinyl, and a limiter was used to prevent the vinyl cutting head from being destroyed, but little creative work was done to the mixes.

Jump to the present day, and we have the somewhat controversial “stem mastering”, machine-based AI mastering (e.g. Landr or eMastered) and everything inbetween.

Before we delve deeper into what mastering is (and isn’t), let’s see how some of the most renowned mastering engineers define their role:

“[Mastering] is the finishing school for the album project. It’s the last stop to get right any concerns that the artist, producer or label has with the sound.” Ted Jensen (via Atlantic Records website).

“Mastering is about trying to make that connection with the emotional message of the music easier for the listener to experience.” Bernie Grundman (via Red Bull Music Academy website).

“The purpose of mastering is to maximise the amount of musicality that’s kind of inherent in the tape that’s been mixed.” Bob Ludwig (via Performer Magazine website).

Enhancing the sonics, emotional impact and musicality of the mix is a common theme among mastering engineers. But let’s not forget the less glamorous technical side that is crucial too.

Read the entire article by Mike Thorne here

Will AI Replace Mastering Engineers?

We asked two professional mastering engineers, Katie Tavini and Mike Thorne, about AI based online mastering services on a recent podcast with Julian. This is a transcribed extract from the show.

Mike: They're definitely quicker and cheaper, and for some situations, that's what somebody would want. It's the fast-food version of mastering, isn't it? I don't have strong feelings about them. I think it's probably a good thing because it opens mastering up to more people. I don't think it's going to have an impact on what I do because what Katie and I do is a bespoke service for people. It's a totally different thing, but there's room for everything,

Julian: So you're saying that some mastering is better than no mastering, even if that mastering is automated?

Mike: Yeah, I think so.

Katie: I've not experienced them personally. I feel like I should perhaps try and see what comes out.

Julian: My experience is brighter and wider, but that might just have been what I heard...

Katie: It's not really mastering, though, is it? Because it's not doing that quality check. So I think the fact that they call them mastering services is a little misleading. However, if they add to the sound in a positive way, then that's only a good thing.

Julian: The thing that always occurs to me is that if you're using an automated service, then if it were to come back unchanged, you'd feel disappointed. Whereas with a mastering engineer, they would say, "you nailed this. This needs nothing." That's a whole different thing.

The reason that professional mastering engineers aren’t worried about AI mastering is simple, the people wanting to use AI services probably wouldn't have ever considered spending money on a pro mastering engineers, there’s no work to lose as they didn’t have that work in the first place. Top London restaurants like The Ivy or Claridges wouldn’t worry if a McDonalds opened on the same street. It’s just the same with AI mastering, it serves a different market on the whole, it suits a certain type of person with a different budget. Both can happily co-exist together.

In many ways AI mastering is doing for music production what the Portastudio for recording in the 1980s. AI mastering offers access to something out of reach for many music makers.

Summary

AI mastering isn’t the end of the world, it’s an exciting development for the music production industry. Like all technology based systems, it can only get better.

However, predictions of the death of the craft of professional mastering are unlikely to come true, no more than drum machines killing off drummers, or Wallmart killing off a bespoke tailor. Each service has its place and anything that gives accessibility and choice to music makers is, on the whole, a good thing.

At the start of this article we talked about early drum machines, plenty of great records have been made using them, but after 40 years of development they haven’t come close to the performance of a real drummer, they are good, but hear the feel a real drummer applies and that’s a different thing. Both co-exist happily in the music production eco-system, in fact many drummers have embraced drum machines as part of their own music making tool-kit. It doesn’t have to be a binary choice.

Computer or human? Process or a personal service? You pays your money you takes your choice!