AI-Powered Creative Platform for Musicians
An older article but never-the-less, an interesting article on AI Mastering. I remember doing shootouts with @landr and humbly thought a novice attempt to master a song sounded better to their AI process.
For years we've been discussing the future of the master engineer so I thought it was interesting to see how large Landr has gotten now.
When RO started back in 1999, we had the largest Mastering Engineers forum on the planet. It had some of the worlds best mastering engineers as moderators to manage all the posts. It was pretty exciting. Year by year that forum has been dwindling away to a point its pretty much dead now.
Without going on... Check out this article published in 2019.
MONTREAL and LOS ANGELES, CA, July 16, 2019 /CNW Telbec/ - LANDR, the AI-powered creative platform for musicians, announced the successful closing of its Series B financing round today, raising $26 million. The round was led by Sony Innovation Fund, microphone manufacturer Shure, state-owned financing corporation Investissement Québec and Fonds de solidarité FTQ, with investments from Warner Music, Plus Eight Equity Partners, Slaight Communications, YUL Ventures and PEAK Capital Partners
I naturally object a bit to anything that takes away from income of audio engineers. Despite that i think the tech behind it is interesting, AI in general. Ive noticed Ozone including more and more AI based features which i generally haven't used much. That said if something is useful, im ok with it.
It would not surprise me if landr eventually exceeds a novice or even intermediate mastering in a fair amount of situations. Between highly sophisticated dsp and further advances in brain science i could see where it could be better "on average". They could also extend into analog processing which would be very interesting to me, using digitally controlled analog. I would be alot more interested in shelling out a few bucks if my song ran thru half a million in analog gear. That would be a good value imho.
I still have a hard time see where the AI would make "artistic" choices, especially when its a non-average type move, or technically "incorrect". This could be aided if the artist had some input on things like "wants alot of bass" or "vocals upfront" or "dark overall", but i know i often want outside opinions because i don't know what i want or where to go.
Hope a service or plugging becomes acessable to the YouTube broadcasting arena, since i find even on big budget companies like ESPN (Disney) their podcast/ broadcasts often have ameture mistakes like low volume, or mismatched voice levels in interviews. Embarrassing mistakes youd get fired for doing sound at a coffee shop.
I still like to control stuff, but the more a evolve the more a trust my ears and the less need i have for ultimate control. I used to dislike switches, fixed frequency bands, stepped knobs, for lack of tweakability. I now find myself gravitating towards this type of stuff for ease of use and immediacy. Ive grown to like simple control sets more and more.
That said i hope mastering engineers can find a niche in the world of AI. It won't be too long before mixers and tracking engineers are in the firing line.
To me this is the equivalent to the super computer tech's that said "one day this will beat humans at chess" and that day eventually came many years after the challenge was posed. This challenge is much harder because there is art involved in mastering. I do see that it could be used in conjunction with humans very soon.
I've used this before to master finished tracks - it gives pretty good results in my opinion without any expertise in mastering, which is really helpful for musicians with limited experience in my opinion. The tool has a few knobs that you can adjust the output with, but with a real engineer you'll be able to make a much more precise / creative master. I don't think this will replace the work for all engineers by any stretch, but will be a good alternative for musicians trying to polish their work on a budget.