Sony CSL Paris is a fundamental research laboratory founded in 1996, using tools from complexity science, data science, and artificial intelligence to investigate fundamental questions in areas as diverse as understanding and creating music, language, sustainability, innovation dynamics, and creativity.
Пікірлер
This sounds like Muzak with words. No one with at least an ounce of musical perception would confuse it with a Beatles song.
Amazing. I didn't even know that Bach wrote parallel fifths and octaves so often. Look at the repeat of the first verse. In all seriousness, if you know anything about music, you can tell that this is utter garbage.
Guys are you ever going to release a public tool? It’s depressing you saying you’re not releasing any of this because of training data. You’ve made some amazing tools I’ve never seen from others, even better than DeepMind’s Lyria
ALV 2016 🤯🤯🤯
Un groupe incroyable et un chanteur si talentieux
been obsessed with this song since 2016, so i covered it! kzread.info/dash/bejne/dGSGsKqqgcu1lKQ.htmlsi=Bh6f3RVodwDD4hcA
When is this plugin releasing for offline use?
It's more modal than Bach. Sounds more like Tallis, or Byrd.
Very interesting! And looks like it could be insanely useful for the home producer. Since you're not releasing the current model, are you planning some sort of eventual DAW integration?
Discovered him with Devilman, and started to learn more about his oeuvre, this man is a master
Time to saturate the music industry. Pump out as much shit as you can.
music is dead
This sounds more like Sean Lennon than The Beatles.
Very cool! Thanks for sharing. Output quality sounds great! Been waiting for a 48kHz model like this. Is is true stereo btw or similar to how Musicgen "stereo" works?
It’s pseudo-stereo indeed but it does quite the trick for low widths. For high width you start getting misaligned stereo images, i.e., sources erroneously fluctuating left and right. We’re now working on a native stereo model :)
Sadly they say they aren’t releasing this because of the training data used. So this is currently just a research project. Hopefully when the law becomes clearer they’ll change their mind, or this will inspire other devs to do it. Like an open source project like Stable Audio. Im surprised we haven’t seen much development for Stable Audio yet like we have with Stable Diffusion
quem mais ta aqui pelo curso?
quem mais aqui ta pelo curso?
quem mais ta aqui pelo curso?
WTF 7 years ago!!!
Stop! Does this really sound like anything but bad computer generated "music"? EDM still involves humans. This is atrocious.
It's said to mimic the style of The Beatles. Certainly seems that way.
❤❤❤❤❤❤❤❤❤
machines made far better music in the seventies. this is disappointing. the problem is that machines are learning from humans right now, and they have sunk into deep darkness. just look at the sky
The perfect corporate song . Nobody to pay .
Le présentateur essaie d'analyser un truc purement cosmique. Mac Carney s'est réveillé un matin et il avait en tête toute l' harmonie de " Yesterday" , qui s'appellait au départ " scrumbled eggs" puisque c'était au petit dej lol. Aucune explication à donner même si Hubert est plus sur la recherche d'un meilleur accord, ça tombe du ciel en grande partie. On peut toujours analyser par rapport à la mélodie mais l'interet d'Hubert c'est qu' on sent qu'il ne se réfère pas à des cadences de jazz ou il se serait enfermé, c'est de l'oreille qu'il puise dans son modeste répertoire d'accords guidé par une petite voix. Le début n'est pas banal du tout, il y a une ligne de basse descendante due au renversement du Sol: ça pourrait très bien être jazzy.
ソニーCSLとチェズマボ！？
Qui est le con qui pose ces questions connes ?
Still long way to go. (Wait a minute. Is this video 8 years ago?)
when is it releasing? Its crazy awesome!
Que Merda
J'aime bien moi la chanson "Dans ce train" ! 😊
Quand va t'on cesser de s'émerveiller qu'une machine soit capable de faire ce que n'importe quel musicien un peu cultivé sait faire également ?
ça reste surprenant que de simple calcul sur une machine puisse produire un résultat convaincant, et c'était en 2016 imagines actuellement ce que ça pourrait donner
ahead of its time for this AI song amazing productions!!
Génie et simplicité.
Someone please put lennon's voice on this 😃
Parti bien trop tot...
Back then, this was awesome to me to think AI can do this Now it's terrifying
Genius at work
Lol. The 7th Guest music.
Not Bach style. Shound like a bootleg version of Frescobaldi
Great job, Sony CSL ! I'm a big fan of your work. Be sure to explore my latest AI music and video creations with A.I. Death Pinch. I'm pioneering the first AI music channel, and your support means everything. Everyone is welcome to join in! Currently, I'm hard at work on six new music videos. FOLLOW THE KZread LINK kzread.info/dash/bejne/hoSg1I-ycb2WYso.html
Wow! It’s April 14th,2024. Can’t believe I’m just hearing this now. This was years before the open AI models of today became “mainstream.” With human and AI collaboration, the possibilities seem infinite!
la fadeur du début de Hey Jude....comme quoi on a pas tous les mêmes oreilles. J'ai souvent l'impression que le plus dur en musique c'est d'épurer, simplifier, garder l'essentiel. tous les accords sont riches harmoniquement puisque c'est leur propos, ce pour quoi ils ont été créés. Certains plus riches que d'autres mais tout dépend de la mélodie, en fait, et de la tonalité...A part ça, je n'aime pas trop le refrain de cette chanson, je préfère de loin Le Capitaine, à base d'accords simples, mais la lumière vient du changement de tonalité entre le couplet et le refrain. Et la mélodie.
Un génie du son ❤
😢
why does this this lowkey go hard as fuck?
Fr! ❤
it is chaotic. my brain can't handle it. it is weird very much. Not very good 😞
This song sorta just reminds me of Tears For Fears' 2004 album, which was definitely Beatles-influenced in parts. So it kinda makes it go full circle :P
better than anything being released today.
Ce mec maîtrisait la grammaire musicale comme personne.