Some coding fun on music & an NFT
Music moves humanity and so does coding. Let’s combine the two and ask some questions given Grammy awards just concluded. Brace yourself for a journey across the senses where we use text, audio, image & video processing using heuristic code & AI to satiate the curiosity.
Note: All data is from Spotify (API) & Wikipedia. OpenAI’s CLIP has been used for generative techniques
Let’s get started.
Who has won most in the history of the Heavy Metal genre?
Oh wow, Metallica truly rocks. The biggies Slayer & Black Sabbath are right up there. Interesting to see Trent Reznor’s Nine Inch Nails up amongst the top.
But as fans, we are not bothered about the winner. Music has no one winner!!!
Who got nominated most would be interesting?
That changes the complexion. Megadeth, Slipknot, Ministry, Korn, Pantera… the bands with cult following get to the list a lot of years.
If we go by popularity who should have won the Grammy?
Green row means the most popular (as per Spotify) amongst the nominees actually won.
Roughly 25% of the time the award went to the most popular one. That’s interestingly low, isn’t it? So what matters… is it the music attributes?
Has the energy & tempo of songs increased or decreased?
Energy & tempo seem to have sustained with ups and downs in between. Loudness in the time period has increased which is awesome. Dance-ability? If headbanging is a form of dance then yes it is pretty much on the right trajectory.
Heavy metal is loud with high energy & tempo so basis that who should have won?
Basis energy… (column with popular shows the results)
When does the winner match the attribute… It always hovers in the 25% range. We were expecting it to be higher. Let’s understand these bands a bit more.
How are the bands related basis the band’s life journey?
No clear patterns here. Lots of metal bands cluster together reflecting similar genres of music and in a lot of cases similar life journeys.
So are all bands the same and not much difference? Are there sub-genres?
What sub-genres have been nominated across the years?
Oh wow, that’s a lot of subgenres… Metal is nuanced music. Basic rock, alternative metal & Nu-metal seem to dominate nominations across the years.
What nominated bands do fans read about most (since 2015)?
Ozzy is the boss and most fans read a lot about him. Interesting to see some of the most nominated bands are also what fans highly read about
Has the fan reading evolved overtime?
Some bands are just evergreen like Metallica, Black Sabbath, Ozzy, Slayer, Iron Maiden, etc… guess that’s why they are considered icons.
What would the audio spectrum of these songs look like?
More fun. In the age of Non Fungible Tokens (NFTs), collage of images is getting sold at crazy valuations. How would a collage of all Grammy winner song spectrums look like? So here you go a collage of Mel Spectrograms of all Grammy Winners in the Heavy Metal category.
We couldn't resist and did create a Non Fungible Token (NFT) of this and listed it on the OpenSea platform (https://tinyurl.com/4r6zuuzc).
Seems like the internet has enabled the crowd to determine every attribute of the song? Why aren’t awards machine-generated? Maybe even the award itself can be AI-generated?
We used OpenAI’s CLIP to generate a “Grammy Award for Heavy Metal Music” and here is what we got
An animation of its creation is below
AI rocks, Generative networks rocks, APIs rock, coding rocks, music rocks as long as it is metal.