Pro Audio Files

Train Your Ears Become a Member

Interview with Justin Evans of LANDR Mastering

Article Content

Recently I met Justin Evans, VP of Product and Innovation at MixGenius/LANDR.

He was kind enough to answer some questions about the online mastering service, music technology, and direction in which digital audio may be headed.

What is your musical background? How did you become involved with software development and where did the idea for LANDR come from?

I moved to Montreal in the late 1990’s, into what was an extremely fertile and generous music scene. At that time there was an incredible techno scene, really deep and multi-generational jazz-improv community, and one of the greatest indie/post-rock communities in the world.

I played a lot of improvisational and experimental music, and had a regular DJ night for quite a while. Playing such uncommercial, audience-challenging music full time meant that I got pretty broke during that period!

Some musicians from the ensemble I led and I started an internet marketing firm that wound up being quite successful — we did some nice work for big clients and also wound up launching a bunch of successful startups.

Eventually I got tired of helping other people build their dreams and decided to leave agency life to make my own startup. That decision led to LANDR.

There are probably a lot of engineers who are either skeptical or threatened by a service such as LANDR, how would you address them?

There was a lot of noise when automatic focus came out on cameras. Now I think you would be hard pressed to find a professional photographer that doesn’t use it sometimes.

From what I’ve read, an immense amount of research went into developing LANDR over the course of many years. Can you elaborate on this process and describe what this research entailed?

The core research behind LANDR came out of a group of Masters and Ph.D students from Queen Mary University of London’s C4DM (Centre for Digital Music). These researchers were investigating how things like machine learning and MIR could do some of the fundamental things sound engineers do, or closely approximate that.

With the help of a local incubator who specialized in technology transfer, we acquired the exclusive rights to the body of research these students had developed and co-founded our company with one of the top students there. A few others from the program have since joined us.

What is a typical day of work like at the LANDR/MixGenius office?

A lot of excitement. We are constantly refining the process. So there are our data science/machine learning team who are very deep in refining the algorithms, our team of mastering and sound engineers who are validating what’s coming in, the cloud team who are building the backbone on Amazon that the system runs on, our UI/UX and design team who are building the next generation of the website, and our sales and marketing team who build partnerships and take care of our end users.

We’re mastering hundreds of thousands of songs a month, so there’s a lot of customer support. Beyond this, we really care about the success of our users, so we’re trying to promote LANDR users, and also manage a customer community where our engineers help people with their mixes and the artists that use LANDR can communicate with each other and give each other tips, etc.

On some days local musicians drop by and give us feedback on how the engine is progressing … it’s pretty crazy but pretty fun.


LANDR seems very open to receiving feedback from users, why? How is that feedback applied?

We are supervised machine learning, so the more input we can get the better the machine can learn. The more we can take meaningful feedback and feed it to the algorithm team, the quicker the system gets better. LANDR, like mastering, is an evolving system.

When we spoke at NAMM, you mentioned a statistic about the percentage of music available on the internet that has actually been mastered, and that the percentage is actually incredibly small, can you recall that statistic?

From a bunch of research that we’ve done, we believe that less than 1% of songs that are recorded are ever mastered.

Has the “Loudness War” been won yet?

From my perspective it’s not the right question. There are too many different needs and applications and genre requirements for there to really be a “Loudness War.” I think it’s probably more of which war are you fighting.

We know our users have very strong opinions on both sides of this conversation, and often it has to do with genre or geolocation.

Or from a smart understanding of how a musician’s fan is most likely to be listening, on what kind of device, etc.

So I think LANDR is agnostic on this question, and it’s why the only control we give to users is the ability to choose how loud their end output is.

If someone wants it slammed, then they can get it slammed, if they want a lot of dynamic range, they can choose our LOW intensity feature and they will get mastering that is probably closer to what I’d prefer.

Are there any advancements in technology (audio or otherwise) that have impressed or inspired you recently?

Tons, really. It’s hard to name them all.

I’ve seen some crazy stuff that links bio-feedback from wearables to music composition or instruments … that’s pretty out there and fun.

I think a lot in the music composition and sound engineering space is going to change radically in the next few years.

Driverless cars are pretty rad. How they sync with services like Uber is pretty mindwarping. I think we’re going to see a very complex relationship emerge between cloud business and services that are automated by machine intelligence in so many places in our lives that it’s going to be almost invisible to us in 20 years. But that world will be astoundingly different than the one we currently inhabit.

Ian Vargo

Ian Vargo is a Producer, Mixer and Audio Professor based in Los Angeles. He has worked on numerous major label and independent records. Get in touch on his website or learn more from him in Mastering in the Box and Mixing Pop.