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.
Comments
Yeah, not better to me, and I have reasons for that. The original composition gets “the AI treatment” (3x) and the dynamics get squished, but the colors and some imagery (the skull) are retained (sort of). Gone is the openness of the original. The AI versions just seem cluttered to me.
On another note and maybe not worth saying because it’s obvious: the original painting is a real concrete object, made with actual physical materials that have a particular heft at a particular scale. Sadly, most never go to actual museums or galleries to see the real thing. And, I guess, some might prefer looking at famous paintings (many quite large), on their iPhones. So everything gets flattened and the artists efforts are represented in such a way that is deceptive to say the least. Hey, it can work both ways - I’ve seen painting that look pretty good on an iPad but utter crap in person.
Automatic video maker thingy...
Presumably that service is connected to a stock footage company (?), because the correct use of stock footage during a video or presentation is everything. And it certainly didn't "create" any of that footage.
It also didn't create a unique voice, it's a deepfake of David Attenborough. I'm sure he wouldn't endorse the use of it like that.
Yah it said that it selected and edited the footage, added titles etc. Soon enough this sort of thing will use image generators, dump it all to an editable timeline where you can edit content, tune transitions etc.
I don't know anything about him but just assumed his voice is licensable, if not that is a dumb move by them,
Not a chance he or the BBC would ever license his voice to a crappy video generator
Haha, well I didn't think specifically to these chodes so much as maybe in general it was available through some third party something something. But nah, didn't give it much thought, of course
Didn't spend much time with it but I think it could be fun for piss taking.
The prompt...
Create an absurdist video about how monkeys eating psychedelic mushrooms will rise up and spread peace through the planet by imprisoning humans to protect them from themselves by forcibly administering "nature's medicine" to humans for their own good. During the video make reference to Terrence McKenna being a revolutionary crusader patron saint primate to the monkeys. Also make fun of organized sports and business as being the repressed denial of therapeutic primal feces throwing. At the end of the video describe how over time humans will devolve to be much happier and simpler once again in tune with nature, but in quiet moments they mourn the loss of the dream of the stars.
The tone should be epic and dark. Do not show only footage of monkeys. Show footage that is mysterious.
@ALB , I'm totally with you on that. Yet, I can't help being a little fascinated in how the AI thing is going. I used to be a painter of both abstraction and representational stuff and I have fed DALL-E my own images to see what it can do; surprising results sometimes. My wife who is still working happens to be a great artist (my subjective opinion) and she works with egg tempera. Sure, you can make prints of egg tempera art but some of the beauty is in the actual translucent layers of paint on the panels that is un-reproduceable. btw, "Alb" used to be a nickname I had as a kid.
Personally, I think any working artist or illustrator should be feeding a desktop version of a good quality machine learning art program with their own artwork to train it. Think of these as computer assistants who are intimately familiar with your style, capable of filling in for you when and where you need them. That's what I'd be doing.
It is pretty darn handy.
Your prompt was well thought out but how it was elaborated by the AI is like you were working with an assistant that seemed to know what you had in mind, filling in where needed.
Thanks for your thoughtful reply. Yes, it is fascinating, but I am concerned that AI (as it is currently) just reconfigures what is already out there. And, in so doing, doesn’t have the ability to make real conceptual leaps. If you follow, it means that it is essentially stuck in what’s already out there. And, if you’re just feeding it your own stuff, then I assume that it’s going to spit out your own tropes and habits more or less, something that many artists want to avoid. I’ve never used it, but from what I’ve seen on this thread, the results are very cliche-ridden. If Impressionism or cubism never existed, would AI ever be able to get to those things on its own? I guess maybe that’s the next step, but I’m not seeing it yet. For now, it just feels like another bitcoin-like hype.
Meanwhile, DALL-E3 is coming soon: https://openai.com/dall-e-3
Some quotes:
That sounds like they're making it worse. In the real world artists copy the style of other artists all the time. Software copying a STYLE and not the exact same imagery should not be disallowed. Seems like a big unforced error on their part. I bet other generated art software to come won't be similarly crippled.
They will certainly have the advantage of not having to figure out system specific esoteric prompting. Their iterative conversational approach will have way more resonance with the masses than things like custom training, control net etc.
Fantastic. And I'm overjoyed I'm no longer doing any of this kind of stuff for a living. Having an idea and having a sophisticated visual taste is more important than agonizing over the skill to create these kinds of illustrations and logos. Let these machine learning systems do the grunt work.
And as they continue to become more sophisticated, I assume they'll be able to maintain complete image continuity while allowing the user to specify specific angles and framing, lens style, etc. Nothing will be impossible to visualize soon.
It looks like it's evolving from real world feedback to becoming more of a legitimate tool rather than mere plagiarism algorithms. This is a good thing and using more creative language you can learn to have actual skill in the use of it.
Even before AI we've always had people who think they're art directors, who say things like: "do it in the style of Wes Anderson" crossed with the style of "{insert another artist name here}" etc
Yeah, that's not being an art director I'm afraid
All that matters is results. How those results are achieved becomes less and less important.
No, the process is important, unless you're happy with generating mediocrity. Art is all about process, it's more than half the story, even conceptual art. A unique process can create a unique result. AI is just another useful tool to add to this.
“Was it made by a person or by a generative art program?”… I guarantee you’re not going to be able to tell the difference in another several years or less.
I think the point that they’re trying to make is that “garbage in, garbage out” still applies to AI
You obviously don't understand what motivates people to create art in the first place, it's not money, it's the process which motivates them. It's the action. It's a practice. If you just want an AI to generate everything then you're not an artist, especially if you're main workflow is plagiarising. The process dictates the quality of the practice. That process could be how you live your life, there are no rules, but it will come out in your work. Yes you could use AI as a primary tool if you want, but how you use it will reveal you and who you are.
You think AI will stop people acting, performing, making music, playing synths, sculpting, taking photos, writing, etc etc just because you can generate a facsimile of something representing those things from a cloud of data? No, most of us will use AI to help us do those things with less friction from the technical hurdles and barriers.
Yah this is why the words art and artist have no significant meaning for me anymore. I pretty much avoid using it, aside from on my resume, purely for survival, it seldom escapes my lips.
I have been an artist and musician my entire life. I think I know a little about my own motivations. LOL.
Self-directed artists and musicians will always exist. This isn’t what this is about. Machine learning systems are important for production and commercial purposes, but they also turn inexperienced people into “pros” because their results will be as good as or better than the person who committed their entire life to their craft.
Do you believe art and music are the product of one person’s experience(s)? I don’t.
When an artist is supplemented (and in some cases replaced) by these systems, the experience of the viewer/listener/audience may effectively remain unchanged.
If the artist creates only to serve their own amusement or alleviate their idleness, then how something was created is also of no consequence.
Incidentally, regarding sculptures... https://www.nytimes.com/2021/07/11/world/europe/carrara-italy-robot-sculptures.html
My brain hurts every time (in a good way sort of):😳
Here’s a site which will iterate different generated images as fast as you can type a prompt: https://fastsdxl.ai/
I'm very new to running SD locally, I use Stability Matrix which is a collection of tools all in one app and location on disk (if you install it in portable mode) which makes it easy to move to an external drive for more space for models. The app basically streamlines and manages a bunch of commandline and browser tools. It's a bit like running a development server on your machine because the web UI connects to it via a localhost address through your browser.
This video helped me to set it up: YT Link
It can get pretty complex but there are simpler options too I think. Depends on how much you want to get into it and what you want out of it. I use Forge Web UI for Stable Diffusion which is open source and runs on my 3090. Some web sites do use Stable Diffusion such as Leonardo.ai which seemed pretty good last time I looked for a co-worker a few months ago.
Depending how close you want to get to the source, or how specific your ideas are, when I tried Dall-E it was pretty frustrating compared to Stable Diffusion running Control Net (with Photoshop in the mix). I think Leonardo.ai supports Control Net still. Control Net may take a bit to get used too though. I know people who have worked in games with me for 20+ years and they just dont click with it, so who knows, heh.
Hmm, I work on the 3D side but have a co-worker who used https://runwayml.com/ to feed in stills that then animated, seemed pretty good at the time, but recently have seen crazy good posts on reddit about others, so hard to say who is the current bees knees. This stuff is evolving pretty quick and there is a ton of noise now.