ai-stairs-twins-stable diffusion genariq

ᴀɪ ɪɴ ᴛʜᴇ ៳𝑦𝑠ₑԼ𝑓 ᴍᴜꜱɪᴄ ᴠɪᴅᴇᴏ

Gen Xin

 

𝑓ᵢԼ𝑚ᵢ𝑛𝑔 'ᴍʏꜱᴇʟꜰ'

filming of myself ceddy ariq ai
While shooting the video for @killceddyy titled "MYSELF" we could already envision the many ways to go about bringing it all together. The decision was to revolve the video around the use of light AI effects scattered throughout the video.
Creating a dream like drug induced adventure around the city.

   The team at Genariq figured the concept of using Stable Diffusion to harden the feeling of theme overall theme in the video worked perfectly. Within the music itself, it complimented the smooth dark vibes in the audio as well as nighttime atmosphere.  

Beginning

on April 7th 2023 in Baltimore, MD

   We began with shooting the first few scenes in the hotel room of Baltimore with Ceddy waking up; we thought of him in bed after a fun night out with his girl would work well.

   In the middle of the night something awakens him and he then try's suppressing these unwarranted thoughts or visions with a blunt to the face.

 

wake up face za music video

 

 

   After the first hit he gets up, seeing himself in the mirror. Beginning to remember who he really is.

  

 

   Not long after getting up and walking to the mirror there's a quick edit of Ceddy smoking then morphing into a sort of oil painted anime. This is personally one of my favorite parts in the video. With it being so simple yet it being out of the ordinary making it much more engaging.

   We made this happen by going frame by frame in Premiere Pro and running them through Stable Diffusion (Automatic1111) individually. 

oil paint ai stable diffusion in mirror music video

 

   Once we've saved all the frames we need we take them into the SD WebUI with a positive prompt of something such as: "Smoke oil painting anime". While we don't usually use negative prompts, depending on the model you're using your output will become more sensitive towards the prompt. Especially with a keyword such as anime.

 

   ceddy myself mirror smoke ai boomerang

   The last shots we got while in the hotel were the ones in the rooms dark hallway giving the creepy feeling that there was another Ceddy unknowingly wondering the room. My favorite part here is when he throws the pills, something about makes me feel fully immersed into the video... Like I could just grab one (or another) and join him in the fun.

The City Shots

   Consisted of Ceddy simply walking around and basically finding he way around what seems to be a maze. I don't think we did much of these part with any music or him performing. We kept it simple. Him smoking and venturing Baltimore.

   Everything worked out perfectly. As soon as we stepped outside the hotel looked right and left, then started walking. We got some scenes of him walking from behind, side, front, close and far because why not? 

   After about 5 minutes of exploring around we stumbled across this parking garage with those fish eye mirrors they use to see better when exiting the garage. Once we saw the mirror I instantly thought about the scenes with the bathroom mirror on the counter and figured that would be the connection between the two scenes.

   By using Stable Diffusion I was able to use what looked to be the POV of Ceddy to convey the feeling of 'tripping' or some what hallucinating. We took some screen grabs of the scene walking up/past the mirrors and used a simple prompt such as 'destructed city' which helped generate fires on buildings and crashed cars.

   Also using keywords in the prompt such as 'cracked concrete' which did exactly what you think it would.. Cracked anything it suspected to be concrete!

 

The Building

   Somewhere in middle Baltimore City there was this very classical building which looked to be used in some sort of politics and after further research (just now) I've learned it was the Baltimore City Hall.

   We used the symmetric aspect of the building to great a great shot of Ceddy walking up one side (with a camera following behind) then shot of him walking down (with a camera following in front) and combined them using Adobe After Effects for the rotoscoping then stable diffusion to smooth what the AI felt like.

  We walked around this building using different sides of it to get the absolute most out of the details. Such as one side with the white metal barriers on it.

   While we could have went extreme on each of theses scenes with AI, we figured it would be better to take as little as we needed to and make it work which helped sell much more authentic feeling to the audience. 

   Throughout the entire video the viewer may notice Ceddy is unwittingly seeing double and even triple. From himself in the mirror.. the white trucks by the parking garage or the multiple busses that seem to pop up out of no where. This all helped tie the entire video together. 

   Being that the song is explaining how he takes drugs to feel like himself while still attempting to find him through what seems to be a maze in Baltimore without entirely noticing much at all.

 

   By the end Ceddy repeats himself multiple times "think I'm up to something". After walking from the hallway to the main room looking at the bed then 3 of him staring in the mirrors. His girl has no idea where he's or what he maybe up to. The video ends with him back where he began in bed with the same paranoia he began with.

   Zooming out the window someone or something is outside waiting and observing just as they were in the beginning.

   Though now it seems Ceddy may have realized he can't stop these thoughts and feeding into them has only hindered his life more than he anticipated. Therefor his only option is the let things be as they are.... come back to reality... and remain.. Himself!

 

 

 

 

Back to blog

Leave a comment