I really had fun creating this track and experimented with several new ideas. This was the first time I’ve tried using the Vocaloid software for speech synthesis, and I think it turned out really well. I had to misspell “senpai” as “sempai” however to get it to sound correct.
Why not just hire a vocalist? Japanese audiences have been fascinated with synthesized vocals for some time, so in my quest to create Asian-inspired house music this just seemed like a good fit. There are other sounds you may recognized. I have a custom koto synthesizer which you heard in Arigato. I reused that same patch for this piece.
Since I used synthetic speech, it seemed appropriate to also try using AI to generate the album art. I tried a couple of different platforms, including DALL E 2 and Stable Diffusion. Some of the images I also used Topaz AI to increase the image size. Some of the results were stunning:
Not all of my efforts turned out so well. For some reason the AI likes to add extra limbs where they don’t belong. Here were a few that didn’t turn out so well:
What do you think? Did I pick the right photo for the cover? Have you tried any of the AI art generating sites?
Here’s the full track on YouTube. It’s also available on all the major streaming platforms including Spotify, Amazon, and iTunes.
