Audiobus: Use your music apps together.
What is Audiobus? — Audiobus is an award-winning music app for iPhone and iPad which lets you use your other music apps together. Chain effects on your favourite synth, run the output of apps or Audio Units into an app like GarageBand or Loopy, or select a different audio interface output for each app. Route MIDI between apps — drive a synth from a MIDI sequencer, or add an arpeggiator to your MIDI keyboard — or sync with your external MIDI gear. And control your entire setup from a MIDI controller.
Download on the App StoreAudiobus is the app that makes the rest of your setup better.
Artificial Intelligence apps ? (A.I. in music composition ?)
Have you guys ever tried artificial intelligence apps such as Amadeus Code? Is it good ? Do you know other apps that can generate AI music?
Comments
Have heard of that one and will check out.
But to answer your question, Refraktions has some excellent AI implemented. I like that the user can interact with the AI algorithm and it learns from your choices. Prefer AI bring a two-way street.
Well, it was mentioned in another thread recently: the most successful AI approach is PG music's Band in a Box, since the mid 90s
It's a rule based system with quite a huge facts base about musical styles, artists and articulations.
I assume Amadeus code is based on Apple's deep learning SDK extensions, which is an alternate approach driven by the current buzzword 'AI'.
But I question the 'instant success' promise on their website.
My experience with BiaB isn't terribly deep, but the software definitely is a tool that CAN assist a lot... IF you're willing to learn (and accept that it's a tool, not a final one-click solution).
It doesn't need artificial intelligence to figure out what defines a catchy tune.
Contemporary taste changes over the years, but composers have done just that in musical history. Atm (afaik) 75% of the top scoring tunes are produced by about half a dozen teams of composers/writers, most of them working almost hidden from public focus.
Thanks for starting this topic, I've been using so called AI environements in database developement for ages - which is actually the main reason to watch BiaB over the years.
It's the only wide spread application featuring this programming paradigm as an user product.
I believe in AI. There are already some “concepts” that tend to demonstrate that it really works but I haven’t seen something that works well for music.
If you find some tools, not necessary about music, it would be nice to share it
Here is a AI picture generator: http://ganbreeder.app
I've developed an open sourced Haskell project for Machine Learning in Music Analysis. I use the codebase for writing and performing as a 'Live Coder'. Here is the project:
https://github.com/OscarSouth/theHarmonicAlgorithm
Not exactly the user friendly interface that is more popular around these parts, but it's something that I find very interesting personally.
Next performance with UDAGAN, here!:
Oh yeah @bleassapp, I also won a copy of your Bleass delay plugin in your recent promotion -- I've been working on repertoire for that upcoming performance today (triggering sounds on iOS devices from Haskell's interactive GHCi environment) and have been jamming out with delay that in the process! Great sounds!!
In my opinion, AI music will sound like nothing we have heard so far. It won’t generate sequences or “midi” files or sheetnotes but instead will mix sounds, noise, tracks and all kind of audio material you can imagine (or not) and create something...new
Can anyone do that? Yes I am sure someone will do it!
I feel that the real value is in augmentation of the human creative process. Can imagine lots of potential for soundtrack music though, where the composer is usually just commissioned to imitate other composers styles/pieces without breaking copyright -- sounds like a job for ML.
I frequently use the Zynaptiq VSTs UnVeil (room/distance processing) and UnFilter (detect and attentuate or increase resonances in signals).
Both applications apply neural networks on input data and define a 'neutral' reference point based on their own processing.
This applies to a single channel (not a mix), but they cannot even solve that 'simple' task on their own - you have to adjust it manually.
Ha! Based on the price and the reviews, I’d say all you’ll learn from that Amadeus Code app is how NOT to design and sell an AI app...
They should rename it Salieri Code!
My .02 is that the best place for AI/machine learning is more to provide the tools to 1) eliminate drudge work, 2) provide contextual reference (scales, chords, progression fit, key changes, etc) and 3) monitor and adapt to the user... The same app/tool, after x hours of use should work/feel different for the guy writing/performing techno and the guy writing film scores...
It should also be able to adapt and offer a different set of tools and abilities if the context is performing vs composing.
It should also be able to work additive as well as subtractive. So if I start with a blank sheet, it should provide tools to help build up something... If I bring it a composition or performance, it should be able to modify or subtract from it to make it better...
I think word processing is a good analog... provide spell checking, grammar and punctuation correction, thesaurus-like contextual reference capabilities, etc.
For any Ableton users, this is a great suite of AI tools:
http://cdm.link/2019/02/magenta-studio-ai-ableton-live/
Thank you all for your answers! really good feedback , I've learnt a lot!
To share a bit more, here are other projects around A.I in music that we've heard of :
https://www.jukedeck.com/
Sony https://www.flow-machines.com/
https://www.aiva.ai/
but so far the most convincing A.I plugin that I've tried so far is Google's magenta plugins for Ableton: quite easy to use!
100% agree
The Magenta build is great. Coord.fm from Jay Hardesty is in development and there are some good demos of it with Ableton Live: https://twitter.com/AppliedTonality/status/1018671907263209477
I was really taken aback when Deep Dream started broadcasting, mesmerized and some haunting images! Would like to see an AR iOS app that could use the iOS camera to capture a control surface and provide information overlays, and mapping with Capability-Inquiry.
This is an area of interest for me always also. In both audio and visual tools. Not to replace human creativity but to augment it, they are time-savers effectively. Machine learning tools are getting better and better. The tool I'm using most is Factor synth by JJ burred. A machine learning synth which can isolate and arrange sounds in layered samples. Very useful. I hope it comes to iOS, he said maybe one day, it's run in Max for Live.
https://isotonikstudios.com/product/factorsynth-machine-learning-in-live/
Hi @bleassapp .
Just stumbled on this post as I'm doing some research on this subject. From what I understand, you have worked with JMJ since this post? Would love to discuss more
Wrote one a while back to explore the field and it resulted in quite an exciting and creative period of exploration with the finished product. If anyone is handy with complaining things then it’s available here open source:
https://github.com/OscarSouth/theHarmonicAlgorithm
Haskell & R
Oscar