Auto Mastering - Is It Really Mastering?
The recording industry is constantly changing and adapting. In the latest craze, automated mastering services like LANDR, CloudBounce and eMastered are becoming more and more popular and they’re certainly a hot topic in the audio community. I’m all for new technology, but in my opinion, there are a few problems with these services that cause many to be mislead and could prevent you from getting the most out of your art...
Firstly, automated mastering algorithms can’t hear like we can. Everything they do is based off of a deep analysis of the frequency spectrum, meaning they can only identify frequencies, not instruments. They don't know a kick from a bass guitar or a hi hat from vocal sibilance. Auto mastering simply adjusts the track to try and match a predetermined EQ curve - if you’ve ever tried to do this in real life, you’ll know that even though it looks perfect visually, it just doesn’t work. That’s because an essential part of mastering (or any part of the production process) is understanding and manipulating the relationship between different instruments to give the song the right feeling, which brings me onto the next point...
Feeling. The purpose of music is to convey emotion and make the listener feel something, whether it’s heartbreak or an urge to dad dance in the kitchen. An algorithm can’t understand that, so it can’t manipulate your track in a way that best portrays it. As a mastering engineer, a crucial part of my job is understanding the feeling that the song is putting across and ensuring that it’s translated to the listener as best it possibly can be. No science or set of rules can help with that particular part of the process, it’s purely a human thing.
An important part of what I do is giving my client’s feedback on their mix. Whether it’s to address a few issues, make some changes that will help the master to translate or simply to give some reassurance that the mix sounds great and is ready to go, this part of the process is really important for artists and mix engineers. An algorithm can’t do this, it can only take what you feed it, meaning your mix could be affecting its ability to give you an amazing master and you would never know.
Further to this, you can’t give it feedback on its master. Revisions are an absolutely crucial part of the process. Most good mastering engineers will work closely with their clients to make sure they’re 100% happy with the final result, making tweaks and trying out different ideas to make sure the song is portrayed in it’s best possible light. You can change things with sites like LANDR, but it’s complete trial and error, the new master isn’t based off of your last version and you can’t bounce ideas back and forth with somebody.
Communication skills from both parties are the key to getting the perfect master. To find a great mastering engineer for your project you need to develop a relationship in which both of you have the freedom to discuss ideas, experiment and work towards your end goal collaboratively. That’s not possible without a human behind the desk.
Back to the musical side of things, if you’re doing an EP or album, automated services can’t understand the relationship between the songs on your track list. None of them do album sequencing, which is a really important part of the process if you want your album to flow smoothly from track to track, with no misplaced volume jumps or inconsistencies.
Lastly, (and this may seem petty, but it’s something I do everyday when I’m mastering) automated mastering services don’t make DDPs for CD duplication and they don’t embed ISRC codes or any other data (artist name, songwriter, genre, year etc) to your files. I’m not sure why none have started doing this yet, as it doesn’t seem like it would be a hard feature to incorporate, but it’s a big part of the process that’s just completely missing.
So if an automated service can’t do any of these things, is it really mastering? By definition, I don’t think it is. It’s just “mastering-style” processing that will make your mix sound different, but not necessarily better. If you’ve spent weeks, months or years writing, arranging, producing and mixing your music, it’s a bit of a disservice to your art at the crucial final stage. At best, you could make a slight improvement to the sound with no way of knowing if it’s the best it can be. At worst, you could undo all of your hard work.
I don’t want to sound negative or salty here. Automated mastering services could be useful for hearing what your song might sound like mastered. (Although you can do this really easily with a limiter in your DAW) They can be great for hobbyists or people putting out quick beats, demos or live videos between their full releases. It also introduces people who are just starting out to the idea of mastering, getting people over that barrier of assuming it’s some kind of unteachable dark art.
It’s genuinely exciting to see this technology being developed and I’d love to see where it goes in the future, but for now I think it just adds more value to real mastering engineers. The important thing to takeaway is that at the end of the day, your listeners do not care or even know how your music was mastered. If the song sounds right to you, it is right, but don’t be afraid to try different methods. Try an automated service, try mastering it yourself and try using a professional - the power’s in your hands and the process that gets you to perfection is the right one.