Special rythmics from Lord General Saelynh (under Doctor SYgr supervision, as usual)
Yet another Destructive Commons (as usual, join us, www.destructivecommons.org(*))
Original voice from Prince Ali
And yup, we loved the original one so we tried being gentle (did we succeed? ask the original song)
http://www.no-human.com as usual


(*) yep we cut the service, back online beginning of february 2016. Shame shame shame

DC | Aladdin, Prince ALI (destructive commons)

Trinidad

Breakfast in America + Medley 2 (covers) - Live Drancy 10.02.2017

New Destructive Commons & Neural cover at the same time. Featuring Benjamin Clementine meeting our dear old IQ which, by the way, tries real hard so don't be too harsh with her.

** For Curious & Scientists :
Neural networks were used to generate the final picture, Texture Nets exactly which still are a wonderful tool. Neural networks were used too in order to extract the voice, please look at old videos credits for paper links. Original style is a picture of a mario bross level

** For Normal People
You do not exist, you just want real hard to. This is definitely not the same. We hug you in the brain. Have a drink

** For Not Normal People
You're not so original, but Quema loves you too. Try keeping this reality exactly as you'd have liked to find it. Good ol' reality

** For Mr B and Annoyed people
Actually, we invited Mrs Q as we had a crappy version of the voice. With a good version, we could do something great.

** For Mrs Q and Other Annoyed people
Hey, she looks like she enjoys it, no ? Does loving her imply forcing her to do only what YOU want her to do. Get a vegetable peeler and love him for what he is.

DC | Benjamin Clementine meets da Ice Queen (yep. proud of it we are)

Destructive Commons - Rihanna - Rude Boy

[Not so sure you need to listen to the lyrics. Waitin for OpenAI or any other one to give us a way to change that too ;/]

Voice extracted by neural networks, video clip with neural style, always a huge pleasure to remix such a great lady/

Ok so please bear with me, this is a combination of using Deep Learning (aka Artificial Intelligence) methods for artistic creation. Tools used there are :

* Open Unmix for the source separation making it possible to extract the original voice (github.com/sigsep/open-unmix-pytorch). There are three different github projects doing source separation, I have a preference for openUnmix over the Deezer project but did not give the recent Facebook Research project a try.

* Neural Style Transfer (good old) for generating the final video from the original video clip. Implementation here is github.com/lengstrom/fast-style-transfer which is great as long as you get used to the parameters (style weight VS content weight)

Enjoy & feel free to make any critics :)

D.C. Rihanna | Rude Boy

Join the fight agains humanity, bamboo thieves, not-salmon music, frigid lady teddy bears, ignorants, meat and banana communist haters ...
http://www.no-human.com

DC | Aladdin, Prince ALI (destructive commons)

NoHumanOrchestra

Recent comments

Too Much

Related tracks