Main “problem” with DAWs imo is that most of them try to model an old taperecording paradigm - even Renoise. Discussing what data structures are best suited to implement this isn’t very interesting to me.
There is this awesome project - that I don’t remember the name of now - where the dev shows how to “freely” add transformer blocks on a timeline. Essentially adding another dimension to the 2D-space of a timeline (tracks+time) in a user friendly way. This is the kind of thinking i like, instead of just another software modelling 1920 technology (?).
I’ve dabbled with the idea of making a simple chiptune tracker myself. For me, the composite pattern seems very interesting, where objects are relatively agnostic to what they are. Only a few special cases would exist (e g when buffers are needed), but otherwise the canvas is rather free - and most special things are implemented more on the gui level than the “sound object” level. I envision a song, including song data and sound patches, theoretically definable in ONE single nested expression.
I’m not a data scientist with lots of knowledge about data structures though.
AI isen’t bad in generally. It can give you very good results in Voice Isolation for instants. As a remixer i very like this and appreciated for that. In past i used heavy filtering in Audacity for that with average results under much time effort. With AI i get good till very good results in a fraction of the time. I be in generally a late adopter and very sceptic on new technologys, but if i was false, i can admit. So i came to the decission, using AI as a Tool for special tasks is leggit and efficient.
If you follow the entire thread (and the comment immediately before yours), you’ll see I’m always approaching this from a general perspective. Renoise is just one example. Perhaps I haven’t realized that, since this is a Renoise forum, and the word appears here, some of you might be interpreting it as if we’re only talking about that DAW. But that’s not the case at all. I haven’t said how things should be, but ask what they could be like (to pose, to ask, to learn). That distinction is very important. I’m fully aware of how difficult it is to change something complex. I haven’t asked for anything to be changed. I just wanted to expand my knowledge. Perhaps sharing data to define things in a thread was a mistake, it seems.
What I have noticed here is that I get the feeling that things are happening outside of this thread that have nothing to do with it, and some people are reacting by bringing those things up here. But it’s okay. We can talk about anything.
I think there’s an underlying problem here. AI is the current revolution. This will soon lead to massive job losses, not just layoffs. Even ethically, the idea that a company can take data from the internet and use it as its own to train programs is a real issue. Wow, that’s a delicate matter. It’s also complex to manage, even though it might seem simple to use by typing prompts. It seems there’s a small battle going on between humans and AI right now. We’ll see where we all end up.
But discussing how a type of program could be different, understanding the evolution in that sense, isn’t a bad thing at all. In fact, it’s very healthy.
I think there are two issues here with AI (which is just a small part of this thread). Using it to research and learn, or using it to create (like in the case of a DAW, a tool that generates voices, or creates melodies and things like that).
About a week ago, I found a YouTube channel that shares songs of about 5 or 6 minutes with a large percentage created using AI. It’s scary and amazing at the same time. And that’s already happening. And we won’t be able to stop it unless all these AI companies stop.
There’s a particular case. A user in a specific profession might keep all their knowledge hidden. However, they use a YouTube channel to let everyone know how to do everything they know. They might seem like a traitor to their profession. That’s going to happen (that someone, if not many, will do that). It’s inevitable. The first ones to do it will reap the rewards. That’s how the world works. Now think about AI.
Anyway, I’d prefer to redirect the thread and leave AI aside (it already helped me before to start the thread). We all already know a little about the repercussions and impressions of using it. I’m learning about this right now too. We’re all in the same boat.
And how difficult it is. To take a different approach to what already exists and try to develop it. I wish I had those skills.
Sounds good. Do it. Take advantage of everything we have now. The satisfaction of creating something new is priceless. It will become increasingly difficult to be original in the creation of software programs or tools.
Because you already understand everything. I’m at an earlier stage and I still don’t know why some things are the way they are and why certain developments haven’t happened. Maybe they’re already happening, and I just haven’t heard about it. But for me, organizing things is essential, otherwise I don’t know where to start.
Look, I’m sharing the channel. Listen to the songs. Let me know what you think. This is already happening:
If you analyze the voices, you can hear something like the absence of human breath. They’re subtle details. But I suppose this will improve with time. The singing style, everything is too perfect. This will also improve with a more humanized approach, I suppose. It seems like that’s the direction things are heading.
Another example that is NOT AI. Let’s compare?
We should start requiring that any AI-generated content clearly identify it as AI; otherwise, it should be removed. It seems very unfair and constitutes unfair competition.
Ok, but you risk analysis-paralysis situation. There is stuff you only learn when you start doing stuff - there is known unknowns and unknown unknowns - later ones usually come up during implementation phase.
I’m annoyed about your super long screeds clearly written by/with AI. It’s so obvious as to be painful. AI is a tool yes, whatever, but you’re clearly copying responses here into “AI” then pasting the responses verbatim, or lightly edited. The formatting is a dead giveaway. I came back to the thread then I see you’re sharing “AI music" which is “impressive” insofar as thousands of dollars of GPUs have spat out an approximation of music and it is shared. Now what? Am I supposed to listen to this seriously? “Music” prompted into existence and for what, exactly? I don’t find it impressive I call it a waste of capex and a fucking sin with the desperate levels of inequality in the world.
The pretentious “debate” about the future of a very good piece of software based on what a server rack in the American Midwest “thinks” means fuck all.
I am hostile to AI generally, but especially boosters absolutely drunk on the “power” of their sycophantic little bots, who have decided that they have some unique fucking insight given to them by the machine gods.
Enjoy it while it lasts, the AI companies can’t subsidise tokens forever and very few people will be able to afford what it actually costs to use these things at a price point reflective of cost plus profit. Don’t get too addicted to your silicon friend.
My problem with AI is the environment cost, energy, resources, minerals, labour - and it’s being siphoned from public in general as increasing energy bills, and eg the rising costs of RAMs.
We all are paying for some silicon boys sandbox dreams - with questionable added value, potentially causing massive problems if it goes full psycho mode, like recently.
@remsky@BriocheBaps Ugh, this whole AI-generated music thing might not interest a lot of people; it’s very subjective. I mean, these days we’re consuming “human music” that’s probably not the same quality as it was years ago. The music itself, the lyrics, the messages, etc. And people are consuming it massively. There are some really good songs these days, but the ones that are consumed the most aren’t necessarily such good songs. Will AI be a problem for people to consume it? There are “conscientious people” who understand the repercussions, but many people are not.
So, how do people react en masse to this stuff? I am referring to the origin of what is consumed. You can see what people think by looking at the comments on AI-generated music videos. Some don’t know if it’s AI. Others accept it. Others are surprised. But there they are, racking up views.
This AI thing is going to “eat us all up.” Just look at how in China (1.4 million inhabitants), at a very early stage, all the students are already learning to use it on a massive scale, young people are the future of the country. This is just the beginning.
But understand this: if someone writes about AI, or shares some snippets of AI to illustrate concepts here, there’s no problem. I suppose you’ll have to get used to it, not just here, but in many places. It’s what’s going to be used, even if we’re against it and there are many negative aspects. It’s clear that powerful companies will do whatever they want, and with the politicians we have and all that, who’s going to stop this? Hate won’t stop it.
It can be fought if humans do things better, whether it’s human-made software, heartfelt songs with soul, art, or whatever.
Don’t use it. Most people with minimal resources will do it. Eventually, you’ll be left behind. That’s how things are. For me, it’s just a tool for now. I don’t attach much importance to the matter of using it or not. It’s like using an image editor. You draw things. I don’t know. But I’m a big advocate for human creations, for feeling and creating beautiful things. I’m aware of everything @remsky has said before. I am also aware that we won’t be able to change anything. Masses of people are going to come and use and consume it. This is already happening.
They’re already holding races with humanoid robots in China. I haven’t seen that in the EU or the US. We’re falling behind!
Having said all that, this thread is now lost (and I have partly participated in that when entering into the subject of AI). If anyone else wants to talk about audio projects (what is raised in the first comment of this thread), developments, improvements, organization, human resources, anything related to how these projects are done, their limitations, and so on, or even interesting ideas like @joule mentioned earlier, I’d like to read them. And let’s leave this AI stuff aside, as it’s a completely separate topic and apparently delicate.
You brought AI into it with your very first AI vomit post, replete with little emojis and bullet points. Reams of absolute nonsense presented like a high schooler PowerPoint. Genuinely gross to read.
There are more than 1.4 million people in China, FYI. Pointing to the Chinese beating the rest of the world is a very well known talking point of AI boosters interested in the perpetuation of the grift. Can’t let the Chinese win, so we’d all better be good and say hello to Copilot and pay M$ plenty of money. The new Red Scare but even more pathetic.
AI music is significantly less impressive than people think. For BILLIONS of dollars of CapEx, Training and Inference the facts that hundreds of thousands of dollars worth of silicon can make a passing resemblance of Taylor Swift is absolutely the least it should be capable of and it’s embarrassing that even with BILLIONS of dollars of investment that it’s still shit.
Do you seek out and listen to that nonsense? Genuinely curious. If you do, why? An amalgam of the entirety of recorded sound, smashed together by machines. Sounds wonderful. All it can ever do is average out it’s inputs. A race to the most average. How utterly thrilling.
Sounds like a “modern DAW” is well within the scope of what these machine gods can quickly vibe code in an afternoon. I look forward to seeing what you can cook up, please, keep the logs and share them so I can witness the AI absolutely fucking the whole thing up over and over and over. Also make sure you’re not on a metered plan lol, otherwise you’d better be able to sell a lot of copies of your revolutionary new DAW.
It seems you haven’t read much of this thread, just stuff about AI and how everyone’s working for it. The 1.4 (1.400) million figure is obviously a typo on my part. The message was clear. I already understand your opinion. What other people do is not my fault.
AI audio content (semantically it’s erroneous to call this music) cost are in 100M$/1B$ ballpark, yet I see no one actually listening to it. Many cool musician could feed are grow on that money.
A forum is for human discussion and debates. Bot usage is off limit, rules ask you to be civil. You starting the discussion with a generated wall of text and then asking for respectful whatevers is rich in itself. As you bias the playground from the start.
What good cake are you expecting if step one of the recipe is pooping on the plate. Seriously.
and if i had more wheels i would be train… scale might be minimal but there’s an impact. you’re normalizing a very rude behavior, and roleplay the neutral/victim afterward. This comes with consequences. No biggie indeed, but still.
Also shifting the blame on Chinese people and least computer-educated demographics is pretty crappy elitism.
All this won’t enhance your latency, neither make your video game GPU less idle.
Some of you are going too far. Disrespecting and implying that I am saying things that I have not said or thought at all. If youselves don’t like what’s happening out there, don’t blame other people or personalize it (people out there will continue to do what they want).
The world is the way it is. You don’t want to talk about the topic of this thread (modern vision of the development of a new DAW)?. Ok. But this isn’t pleasant at all.
I’m not going to answer any more about this matter.
You’re not saying things, you make some computer write things for you. I think it’s rude and toxic, and I’m not alone.
You’re writing english fluently, and smart enough to have thought of your own. Don’t use a forum as your notepad. Forum users deserve better from you. Do better.
The world is the way it is indeed, and as you can read : the world is actively disliking AI-fueled exchanges. Sorry if it’s not what you expected, but it’s « the way the world is » and it’s rejecting this way of exchanging.
What doas it play for a role HOW he organised his text/post for the thread topic/thematics, as long it are his own thoughs?! What we here see is the well known mechanism of a part of the renoise community. Instead of writing some senseful or argumentative to the threads topic, the thread will overtake from a few people which are triggered from a minor detail and feel compelled to give here personal worldview about it. And always after a few posts it turns on accusations on a personal level.
Raul did/doas much for the Renoise community, the Renoise Tools enviroment and the Gui/Coding Eco system. He inspired many people and gave/give his time, knowhow, and experience in coding and workflow to make Renoise a better DAW. And for this we should respect him.
Everybody can think over AI like he want (I dont like AI very much too in much areas), but using it as a tool for textorganisation dosen’t make his expression of opinion in this forum here illegitim. He do a few post in addition to specify his questions and targets and what his intentions are with this thread. And this should not be classified and dequalify as “KI bot actions” or something similar.
Hey, don’t worry. Everything’s fine. I also understand that some people are upset about certain situations. I know that for some people, this particular topic is a bit sensitive.
I’ll just leave a thought on this matter, and everyone can take it as they wish…
As I said before, I’m a huge supporter of human creations. I appreciate human art, and I base my life on feelings and overcoming challenges. I’m not a programmer (it’s not how I make a living); I work in something completely different. I enjoy programming as a hobby, and it’s a passion of mine, combined with composing music, learning about it, and feeling it. But I also do my own research, and I like discovering and understanding new things.
But I’m very aware of my place in the world, and regarding AI, it’s an issue beyond anyone’s control here. Some participants might shout, get angry, and all that, thinking they’re going to change something external. I understand that. Very unfair things are going to happen. But being disrespectful or destroying threads with unfortunate comments is out of line.
Relating this issue of using AI assistants to the main topic of this thread, I’m thinking about one thing: trends. Everything points to AI increasing dramatically in the next 10 years. This is unstoppable and will require legislation and regulation. The norm is that everything will become much more efficient (that’s what research is for, both in creating cheaper energy and in developing specific chips, distinct from GPUs). While increasing efficiency and solving many current problems in that regard, advanced AI will be used much more (personal assistants, robotics, biology, saving lives, research in many fields, to extend the life of the human being, and much more). How do we weigh the positives against the negatives? Well, right now many people who complain travel by car and pollute. That was also a revolution in its time, and we’re quite happy now to be able to get everywhere in a short amount of time. Hypocrisy. Projecting trends is speculating about the future while also considering the past. Today, many of those who complain in the world will end up using AI. And then they won’t complain.
That’s my reflection. I understand the reactions, but we shouldn’t go overboard. Just because someone comments on a specific topic here doesn’t mean they’re endorsing it. They’re simply stating that that is happening.
Okay, having said that, relating it to a software project, I find it hard to believe that you have two very powerful components in your computer (component 1: CPU, component 2: GPU) and only one is being used to its full potential, simply because it’s so powerful, while the other isn’t, because it’s geared towards other things (graphics). Well, just as in AI, specific chips for AI are going to be manufactured (if they aren’t already), much more efficient (more operations with less energy), the same could happen with GPU chips. But if we never talk about these things, if we don’t plan ahead, if we don’t rethink things, this won’t happen. I’d like to think that audio software will evolve in the coming years. If I don’t understand how things work, all these questions remain unanswered.
Right now, audio programs use the CPU for audio and the GPU for graphics, each for its strengths, and they work in sync. But hey, current GPUs are very powerful. If this could be adapted to process more audio-related tasks, since it’s more than capable of handling graphics as well, everyone here would be amazed. We’re already seeing this with Bitwig (CPU audio + GPU graphics), aren’t we?
Regarding Renoise (because some previous comments don’t seem to reflect my opinion), for me, it’s a magnificent piece of software. Looking at its history and development over a lifetime, if any program deserves more recognition, it’s Renoise. And that doesn’t mean anyone is asking for changes (there are thousands of requests for that in the forums), but rather that it’s a full acknowledgment of software that’s currently undervalued, and which would be a real treat if it were perfected at some point in its history.
I want to understand all of this, and know why things work the way they do and what the limitations are. That’s why I started this thread. I’m someone who left university a long time ago. If I want to learn something now, I’m self-taught. I use forums, search engines, AI assistants, and whatever else is necessary.
There are already a few VST’s that are using GPU Calculation for Audiocalculations like some Reverb effects on the market. FFT could be much faster & much more efficient on GPU Calculations. And there are some VST Synths using modern CPU Coprocessing like AVX2 and AVX512 like Avenger 2. Sadly my i7-2600K CPU (2. Generation Sandy Brigde) only has AVX and cannot run Avenger 2. The setup routine exit with AVX 2 on my PC.
With low level GFX API like DX12 or Vulkan which have higher drawcall amounts should low latency operations on GFX Hardware for Audio Data possible. You have to write your own Memory Allocator maybe and some Data Streaming Wrapper to load Audio Data Chunks in GPU memory. Naturally not all things will be
possible on this way. But some for sure. There are many ways for high parallel calculations then HLSL, CUDA or low level APi access.
Hauptmerkmale und Vorteile von CUDA:
Parallele Verarbeitung: CUDA nutzt tausende Rechenkerne in NVIDIA-GPUs, um Aufgaben massiv parallel auszuführen, was bei datenintensiven Anwendungen zu enormen Geschwindigkeitsvorteilen gegenüber CPUs führt.
CUDA-X Libraries: Umfasst eine breite Palette von Bibliotheken (z. B. für Deep Learning, lineare Algebra, Bildverarbeitung) sowie Tools, die Entwicklern die Arbeit erleichtern.
Breite Unterstützung: Viele Anwendungen, insbesondere im Bereich Deep Learning (PyTorch, TensorFlow), nutzen CUDA im Hintergrund.
Programmiersprachen: Basiert primär auf C/C++, bietet aber auch Wrapper für Python, Fortran, Java und andere Sprachen.
