"Try LALAL.AI before deciding it's a hopeless cause." Interview with Gaëtan Schurrer
At first, Gaëtan didn't want to employ AI to save a live recording he was working on. Now though, he and his entire team is singing praise to LALAL.AI for rescuing what turned out to be the recording of an awesome performance.

North Street Group was created by Bourby Webster in the mid noughties originally to empower and support Western Australian musicians by showcasing the incredible talent.
It became the management company behind the Perth Symphony Orchestra at its inception. It has now evolved into a production company, The RARE Company, creating large scale orchestra productions, encompassing the North Street Music (which is the academic arm only) with consulting, artist development and specialized courses, festival and cultural event programming, and orchestra fixing and musicians for hire.
As its technical director, with a long music recording and production background, Gaëtan's involvement has been all things production and technical, with a special focus on Audio.
We sat down with Gaëtan Schurrer to talk about his work and how LALAL.AI has become a go-to tool in it, even though he was reluctant to try it at first.
"If It Sounds Good, Don't Change It."
Throughout the 90s and 00s, I worked on myriad albums and singles, mostly in the dance and pop world, but since moving to Australia in 2011, I have shifted to live sound engineering rather than the studio, and have found many new challenges in the recording and mixing of live recordings.
Recording studio work has still been present, but has taken a backseat. I have also collaborated in film scores as well as composed for advertising, and had a stint in theatre where I got more involved in sound design. With NSG, there has also been a need for sound design on various productions we either created or curated. I really enjoy creating ambiences or setting a mood for the pre, mid or post-show moments, keeping the audience engaged and on point. I have also been involved in audio post-production, teaching it in Florida at Full Sail University, and applying my skills to local Australian films and documentaries.
In my early years, I made a lot of dance music, and was involved with many world class DJs and producers at the time dance music crossed over into the pop world. While I did also work on some more esoteric projects of various genres, I think that my early experience in listening to many vinyl records as a teenager in the 80s, then growing through the more digital sound of the 90s, has given me an understanding of production and mixing sensibilities that enable me to deliver results much appreciated by my clients.
I have always had a "if it sounds good, don't change it” approach, where it's all about balancing things rather than attempting to enhance or over beautify any elements. I would always rather EQ out a few offending frequencies to help things gel together and create space for everything, over adding frequencies to enhance elements, dwarfing others in the process. I am also a staunch advocate of quieter mastering in order to allow the dynamics of a piece to come through. I think the loudness wars have created so many records that never grow to be massive at any one point, because they're just loud all the time. I so much prefer a piece that starts quietly, so you turn your music system up, and then when it goes big, it's really big!
Mastering has changed enormously over the years. When I started in an analog recording studio environment, where the master mix went down to a quarter inch or half inch tape, the mastering process was still mostly geared towards vinyl. This was an incredible science that only few mastering engineers truly understood and perfected, and for those the results were otherworldly. You always knew when you put your needle on one of those records as they sounded so, so much better. Even getting a decent master for cassette tapes was a bit of a feat, and an entirely different process.
As we switched to the CD format, there no longer was a need to translate the studio master into a mastered version that fit the medium. It became more about getting the most level possible within the digital range, at the expense of dynamics.
"I do think we have gone full circle now, as mastering in my view has become more musical again, and such medium as vinyl is actually making a bit of a comeback."
But to get on point, when I work on any production, as I reach a point where I am happy with my mix, I will then put on my mastering hat, and process the finished master mix through what has become a few go-to plug-ins that give me just the right amount of EQ and compression controls to put a nice sheen on the finished product, and ensuring it is loud enough (compression is wonderful when used judiciously) while retaining the piece’s dynamics.

My best dance music work was collaborating with DJ Sasha, originally remixing every one from Michael Jackson to Madonna, and later recording and producing such seminal albums as Northern Exposure and Airdrawn Dagger. In the pop world, I made three albums with M-People in the UK, and applied my sound programming and production skills to others such as Faithless and Simply Red.
I’ve also had a long-standing relationship with Jeff Wayne and his Musical Version of the War of the Worlds, which I’ve remixed, rebuilt from its original recordings and mixed in surround sound, and created all the audio for the initial live touring production of 2006.
The War of the Worlds is a concept album still popular after almost 50 years, and it’s been an honour to be able to contribute all I have over the years. In fact, NSG is working on a new orchestral production which will take it to place it hasn’t been.
"Originally remixing the hottest records, and then feeling the vibe of thousands dancing to them in clubs all over the world was very rewarding."
Sasha was and is one of the most pioneering and avant-garde DJs, always pushing the boundaries which for me sonically was all I ever wanted to do. Our Northern Exposure I and II double albums took the genre to another level unmatched since, and DJs all over the world still scratch their heads in denial attempting to do anything like it.
And while it originally felt like selling my soul to the (commercial) devil, the work I did on M-People’s albums I am very proud of and remember fondly. And then, there are so many smaller projects which, while less successful, broke ground in their own way, and take me back to a world I’d go back to in the blink of an eye.
"In this era of earpods, phone, laptop & bluetooth speakers, I feel audio quality has become a bit lost, but still people appreciate a better sounding record or show."
Live recording and mixing has a unique set of challenges: how to get the best results in record time, with the inability to go back, often working with sub-par input and battling with bleed issues. In my production years with NSG, I’ve learned to truly appreciate how easy it can be to record an orchestra in a controlled environment, and conversely how hard it is to amplify it in a live environment.
Then add some loud electric instruments and amplified vocals, and you’ve got a serious job juggling it all to deliver a pleasing result. An incredible challenge, and actually a relief when it really works!
In this era of earpods, phone, laptop, and bluetooth speakers, I feel audio quality has become a bit lost, but still people appreciate a better sounding record or show.
In the live world there never is the time (or budget) to do things right. So you just do the best you can with what you have.. I’m sure this will resonate with many. And don’t even get me started on the audio post world—always the last priority in the budget, always has to be done yesterday, and then you’re lucky if or when you get paid!
"Before using LALAL.AI, I used rather crude filtering using EQ, compression and manual levelling, to be honest I didn’t spend an enormous amount of time trying, as the results didn’t justify the means."
"The guitarist suggested AI to attempt to remove the noise. I was initially reluctant to go down that rabbit hole."
As I was mixing a live drum and bass band recording I’d captured, the guitarist complained his guitar was not loud enough (familiar?). But then, his guitar came through a small amp into a dynamic mic that captured more band and room sound than anything else. When told, he suggested AI to attempt to remove the noise. I was initially reluctant to go down that rabbit hole, so he offered to do just his track in the offending song and sent me the result.
I was rather impressed and asked what he'd used, the rest of history.
I went back to multiple stems in that live band recording. Not only the guitar, but some of the main vocals, and especially the freestyling MC performance had more noise than material in the recordings.
"The vocal isolation helped hugely in being able to mix these to obtain a decent result."
We had a recording made of one of our shows in Canada this January that was produced remotely. The orchestral performance was outstanding, but in our absence the local producers and sound crew did not understand our intent in mixing amplified vocals and drums with an orchestra. The results were all but unusable for our purposes, which was to showcase the orchestrations of this incredible re-imagination of a seminal grunge band.
Having experienced LALAL.AI’s abilities in my own live recordings, I experimented on a short sample to try and first remove the vocal, and then remove the drums to enable me to rebalance the recordings.
"I was able to mix it again in a way that had our composer shedding tears of happiness for finally being able to hear his arrangements performed to perfection."
I did not believe that first removing the vocals from a poorly recorded performance, and then removing the drums from the no-vocals stem, would result in a usable stem that would enable me to re-balance the mix.
I even experimented in trying to remove the drums first, and then the vocals to compare the result in sound quality, and found a small difference. My order choice was confirmed by an email from LALAL.AI after I explained my query. Having separated my stems, I was able to reassemble and remix to my liking, creating a final mix that is actually what we wanted from a recording that was all but ruined.
We now have a video of our show we can be proud to promote it with.
But wasn't it for LALAL.AI, I do not believe I would have been able to rescue this recording, as the orchestra was barely audible behind the vocals and drums, and no tools that I know of would have been able to do what LALAL.AI did.
I found LALAL.AI to be providing a better sound quality than other similar tools I tried. While I could hear some unwanted artefacts in the solo stems, when combining them these are all but inaudible.
"Without such an AI tool our recordings were simply trash. Try LALAL.AI before deciding it's a hopeless cause."
I can think of many creative uses employing LALAL.AI in ways it is most definitely not designed to do, and cannot wait to give these a go. I have an experimental sound and video lab in which we create soundscapes that I am interested in mangling by using LALAL.AI.
Early days, but I can see that all my live recordings mixing work is going to improve drastically, enabling me to get fast results from recordings I would otherwise have thrown in the bin.
"The band I first mixed using LALAL.AI was very complimentary of the mixes. We have recorded before, but I think these were the best results we were able to achieve to date."
North Street Music itself is singing praise to LALAL.AI for rescuing what turned out to be the recording of an awesome performance.
I'd most definitely recommend LALAL.AI to other sound engineers. I think live sound recording engineers will quickly learn to appreciate what can now be done in literally a few minutes to improve the recordings enormously in order to obtain much better finished mixes of the live shows.
I can't wait to use LALAL.AI in ways it was not designed for, as it's my firm belief that most creativity results from accidental experimentation. While I do precious little post-production these days, I recently had one very poorly recorded voiceover that would have massively benefited from a LALAL.AI treatment. The tool I used at the time was vastly inferior.
I have also often had the need to isolate certain elements of field recordings, which LALAL.AI would excel at. Even the creation of simple karaoke versions of any songs to help vocalists hone their performance or even just to learn lyrics, is an invaluable use of LALAL.AI.
Without such an AI tool our recordings were simply trash. Try LALAL.AI before deciding it's a hopeless cause 🙂
Follow LALAL.AI on Instagram, Facebook, Twitter, TikTok, Reddit, LinkedIn, and YouTube to keep up with all our updates and special offers.