Embark on a mesmerizing journey as Yumi delves into the captivating realm of AI-driven visuals with the creative mind behind her latest music video for “Desdemona.”
In an engaging conversation that illuminates the intersection of artistry and technological innovation, Yumi, the enigmatic artist behind the enchanting track “Desdemona” from her upcoming EP “Better Without You,” sits down with Michael Bonitz, the visionary behind RabbitHole Romance and AI Music Videos in Sydney.
As Yumi’s insatiable curiosity takes center stage, she embarks on an exhilarating exploration of the video’s creation. In a harmonious exchange of infectious enthusiasm, Michael invites us into his creative realm, where a fusion of AI, the Adobe Suite, and boundless experimentation seamlessly intertwine, giving birth to a mesmerising visual narrative that brings Yumi’s sonic odyssey to vivid life.
The story takes an unexpected turn as Yumi delves into the artistic inspirations that breathed life into the video. Michael’s chance encounter with an Alexander McQueen exhibition and his fascination with Jeremy Mann’s distinctive style become the unlikely muses for Desdemona, the mesmerizing visual realm he envisioned for Yumi’s sonic masterpiece. Blending elements of gothic grandeur and Mann’s evocative oil paintings, the video becomes a testament to the boundless creativity unlocked through the fusion of AI and human imagination.
As the interview progresses, Yumi’s inquisitive nature leads her to explore the delicate balance between human vision and AI execution. Michael shares insights into his collaborative dance with AI, where he sets the parameters and crafts prompts, allowing the technology to weave its magic and give birth to a visual odyssey that mirrors the intricate emotions and themes of Yumi’s lyrics. The result is a harmonious synergy between human intention and AI creation, creating a visual tapestry that captivates the senses.
In this captivating conversation, Yumi and Michael unravel the mysteries behind the creation of “Desdemona” music video, showcasing the transformative power of AI when harnessed by human creativity. Together, they delve into the enchanting world where art and technology intertwine, leaving us with a sense of awe and anticipation for the future of music visuals.
Yumi: How did you make the video?
Michael: This answer could get quite boring, so I’ll be as general as possible. Ai, Adobe Suite and a bunch of trial and error.
Yumi: What inspired the artworks during the video?
Michael: On the day I started working on the video, you happened to be at an Alexander McQueen exhibition. I had no idea who he was, so you sent me some pics of the exhibit. His crazy gothic style immediately resonated with me, and I knew I wanted to bring something like that to Desdemona.
Around that time, I had also been exploring the work of an artist called ‘Jeremy Mann’. He has a very unique oil painting style that uses stains and dripping paint to create really interesting art. With the power of Ai, I was able to combine their two styles to make the trippy mess that is Desdemona.
Yumi: How much did you have to tell the AI to create, and how much did you design yourself?
Michael: Great question! I wish every AI artist had to answer this publicly. So my role in this video was the following:
– establishing a clear style for the AI to follow
– Creating prompts that correspond with the lyrics
– Manually inputting all the camera movement/prompts to make the video move with the music
– Creating a custom AI model that recognises Yumi’s face
– upscaling and interpolating each frame
– Fine tuning the color and coherence with Adobe
But with that being said. The entire video is generated by AI. I didn’t draw a single frame. Sorry animators.
Yumi: How long did the video take to make?
Michael: I think it took me about 3-4 full days of work to finish Desdemona. If you had asked me 6 months earlier it would have taken far longer, but I was at the perfect point in my AI career to make Desdemona when you asked me.
Also you loved it, so there were no changes needed once I sent off the first cut.
Yumi: Discuss how you managed to mirror the lyrics with the artworks.
Michael: Essentially, I took the imagery that you used in your lyrics, and I translated that into Prompts that the AI would understand. Then I timed those prompts in the video to morph and evolve with the song.
Making this happen was the hardest part of the video, as it requires a large amount of trial and error. The AI will often misinterpret the lyrics and give you results you don’t want, so you have to adjust the prompts and run the AI again from the start.
Yumi: The visuals are flowing and like a trance, are you inspired by any directors/ filmmakers/ artists?
Michael: I could write you a full essay here, and it still probably wouldn’t cover all the people who have inspired me. My favourite aspect of filmmaking is VFX and animation used in music media. So my biggest inspirations are people/studios in those fields.
I’ve always wanted to bring VFX to live music footage, and no one does it better (imo) than ‘Gibson Hazard’. Just watch his videos and you’ll understand. Another awesome person that integrates trippy visuals into live music performance is ‘Guerilla McGavin’.
There are a lot of amazing music video creators out there that inspire me, but I feel like I have to mention the work that ‘Cole Bennet’ has done with Lyrical Lemonade. His videos have been pivotal in many artists careers, and are always exploring new uses of VFX in music videos.
Another music video creator that I have been following is called ‘Nitetive’. They push VFX to the absolute limit in their music videos. There are countless animators that I could mention, so I’ll briefly list some of my favourites:
– Felix Colgrave
– Minelauvart
– Dogceo
– HelloPersonality
– Remi_Molettee
– Eduard_ov
– LuckWhacky
Yumi: What interests you with using AI?
Michael: This is another question that I could write a short novel on. To say that AI is an ‘interest’ would be an understatement. AI is a necessity in my field, and without a good understanding of it would be a detriment to my growth.
When I first started making videos seriously, I had this grand idea of integrating amazing animations with music videos and live performances. As I developed my skills in editing this dream seemed less and less realistic, as my skills and hardware could only get me so far.
AI has given me the ability to create these unique animations and VFX that I would never have had the time or patience to make before. It is an absolute cheat code that, even if you hate it, needs to be recognised, because it is only growing stronger and stronger.
Yumi: What projects are you working on next?
Michael: I’m actually on holiday while chatting with you, so I’m not working on anything right now, but I have so many videos I want to make when I get back home. In fact, you and I have been discussing a music/fashion video that I’m super excited to make.
There are about 3 music videos that I have worked on recently that are yet to be released. I wish I could tease when they will be coming out, but I actually have no idea when/if that will be. If anyone likes Desdemona and wants a video like it, I’ll be ready for you in July :)
Yumi: When did your creative journey start – what medium?
Michael: When I was younger, I used to be in a breakdance crew with 4 of my friends. We would spend almost all of our free time practicing our moves and making videos of us breaking. I would be in charge of filming and editing, and I was always super proud of all the videos I made. They are actually still on YouTube if anyone is interested. Look up ‘WeFlipItOut’ (I didn’t come up with the name).
After high school we didn’t have enough time to keep breakdancing, so our crew broke up. I tried a business degree, but it was not for me. So I dropped out and started working full time for a meal delivery service catering to athletes. This is where I was first introduced to Adobe.
I spent so many hours cutting food and bodybuilders out of pictures with the pen tool and making motion graphics from them. I am super grateful for this time, as I was given a lot of creative freedom to make content the way I wanted.After a year working for the meal delivery service I decided I wanted to develop my skills in content creation, so I moved to Melbourne and got a Diploma in design. I learned a lot about all the different design disciplines, and honed my interests on editing and VFX.
Since then I’ve worked in many design roles for many different people, but my ultimate goal is to make music videos and music visualisations for live performances.
Yumi: What part of the video are you most proud of?
Michael: I really like the section where you sing about a weeping willow and the animation becomes a tree trunk with trees in the background.
There is also a part later in the song where one of the Yumi’s turns and looks at the camera. It happens completely by accident and the AI would normally never do that, so I’m quite pleased that it did.
Yumi: How does AI work and are you excited to see ai come into the music/ film industry?
Michael: AI is a gigantic database of information that a hyper intelligent machine can scan for specific prompts or information that a person asks for.
The AI I use scans a database of artwork to create images and animations that correspond with the users promots. Chat GPT scans written words on the internet to answer your questions.
Imagine ‘google’ but you get an original answer or image/animation for whatever you ask for.
I am both afraid and excited by the potential of this technology. With its rapid development I have seen so much unique work be created, but at the same time I have seen people dedicate their lives to skills that are now becoming trivial. One day it is likely that everything I do will be trivial too.
So in the meantime I plan to use AI in the most original ways I can, and try my best to not become redundant as the technology continues to evolve.
Yumi: Who are you most inspired by artistically/ what’s your main source of creativity?
Michael: I’ve already listed all the people that motivate me on a surface level, but in truth, nothing gets me more keen to create than hearing someone say they like my work. Call it egotistical, but It is the reason Desdemona came out the way it did. You liked what I was doing and saw how it could bring your music to life. Your belief in my work inspired me more than any other creatives work ever could.
So if you see someone making art that speaks to you. Let them know. You have no idea how much it could mean to them.
Yumi: What was your ultimate take on the song?
Michael: I remember when you first showed me the EP I was like ‘Damn Yumi… who hurt you?’. I don’t want to make too many assumptions about her personal life, but after listening to Desdemona and Better Without You, it seems like you has gone through quite a lot.
While your heart may have been broken. You have taken that negativity and channeled it into some beautiful music that has ultimately empowered you, as you are clearly better without them.
Yumi: Where do you see your career in AI art going in the future?
Michael: My dream is to make music videos/visualisers for my favourite artists and to also see my AI animations projected behind artists as they perform music. I have been making music visualisers for a little less than a year now on my AI page ‘RabbitHoleRomance‘, hoping to get the attention of my favourite artists.
I’ve had some nibbles, but no bites from any big fish yet. I also have day dreamed about having a live exhibition with huge screens showing my work. I think that would be a super cool and trippy experience for people that see it.
Yumi’s highly anticipated EP, “Better Without You,” finds its home at the prestigious Damaged Goods Records UK, marking a momentous collaboration that sets the stage for an exhilarating musical journey. You can catch the launch at The Chippo on June 29 info here.
