I am building a SaaS startup in stealth mode (probably will apply YC)—finding it challenging to balance time between the startup and open-source projects. Ironically, it felt easier at my previous job as employee in a startup: work during office hours, open-source on the side.
Now I spend my free time playing my Fender and exploring ways to combine music with programming, maybe using LLMs for copilot performances.
I'm having a blast playing with Glicol. Thanks for posting it.
Any aspirations to add more song-level concepts like arrangement, chord progression, and/or semantic meaning? Right now it excels at creating short loops but I'm not seeing a good way to get a sense of movement and purpose (short of hand-coding everything)
But it's not complete yet. I have indeed been considering chords, but the syntax needs more changes and design.
But according to your description, https://strudel.cc/ can bring you a lot of fun.
Compared with Strudel/Tidal, Glicol provides more on precise time control in audio engine, low-level sound synthesis and sampling capabilities (needs more development but completely feasible).
This is very cool. Unfortunately it's quite hard to use ChatGPT to generate new music, since i don't see good docs that it can use. It will be helpful to create a nice doc as prompt to use it to force LLMs to create custom music.
ChatGPT has at least been trained on a variety of textual musical notations. It will generate chord charts or drum patterns. In the past I've asked it to generate chord progressions demonstrating various music theory concepts.
It would be cool - 'haps - to use one of the many music/accompaniment algorithms available, adding "nuisance" or "humanness" through AI, creating a virtual musician.-
Hello, and thank you for creating Glicol—it’s an outstanding tool for live coding and audio synthesis!
I was wondering about the possibility of extending Glicol’s functionality to include features such as:
TTS Emotional Replication
Generating human-like emotional expressions in sound (e.g., sighs, coughs, gasps).
Sound Effects Simulation
Creating realistic environmental sounds, such as thunder, rain, wind, and other ambient effects.
Novel Sound Creation
Enabling the generation of entirely new and experimental sounds, potentially by training on a provided audio dataset.
Given Glicol’s modular architecture, Rust-based engine, and WebAssembly (WASM) support, I’m curious whether it would be theoretically possible to support such extensions within the current framework.
Questions
Are any of these features currently possible with Glicol?
If not, are there any plans or ongoing developments to enable functionality like this in the future?
I believe these additions could expand Glicol’s applications far beyond music generation, opening up exciting possibilities in areas like:
TTS systems with nuanced emotional expression.
Sound design for games, films, and interactive media.
Experimental audio research and creative sound exploration.
Before concluding, I’d like to add that Glicol’s potential use cases could extend well beyond music generation, offering new opportunities for innovation in audio-based applications.
I’d love to hear your thoughts on this and whether these ideas align with Glicol’s future vision. Thank you for your time and for developing such an inspiring platform!
To be honest, I am not an expert in TTS. However, the angle you mentioned does have many application prospects, but I feel that the project itself may be larger than Glicol :), and it may be provided as an independent rust package. The short-term goal of Glicol is to cooperate with other instruments and LLM.
Strudel [1] is similar (no affiliation), inspired by Glicol, is a bit more approachable in terms of syntax, has a great visualization in the editor (highlighting the sounds being played) and has a nice tutorial [2].
Csound is legend. You can find almost all the sound effects in csound, and the score of csound is an exclusive feature.
Compared with Csound, Glicol has a completely different logic in syntax, which is much more lightweight and focused on sound synthesis and live coding. The core logic of Glicol can actually use "organized numbers" to make music. Numbers are used between nodes, and you can even "listen" to an envelope. Controlling these nodes is like having countless sound modules. The original idea of Glicol was that after I became familiar with DAW and modular synthesizers, I wanted to design a language that could simulate this workflow.
Related:
Glicol: Next-generation computer music language - https://news.ycombinator.com/item?id=37360720 - Sept 2023 (2 comments)
Show HN: Glicol - A Graph-Oriented Live Coding Language/DSP Lib Written in Rust - https://news.ycombinator.com/item?id=30942074 - April 2022 (29 comments)
Show HN: Glicol: Music/DSP live coding language 'n audio library written in Rust - https://news.ycombinator.com/item?id=30811015 - March 2022 (1 comment)
Glicol – Next-generation computer music language - https://news.ycombinator.com/item?id=30746299 - March 2022 (1 comment)
Author here, happy to answer any questions.
I am building a SaaS startup in stealth mode (probably will apply YC)—finding it challenging to balance time between the startup and open-source projects. Ironically, it felt easier at my previous job as employee in a startup: work during office hours, open-source on the side.
Now I spend my free time playing my Fender and exploring ways to combine music with programming, maybe using LLMs for copilot performances.
I'm having a blast playing with Glicol. Thanks for posting it.
Any aspirations to add more song-level concepts like arrangement, chord progression, and/or semantic meaning? Right now it excels at creating short loops but I'm not seeing a good way to get a sense of movement and purpose (short of hand-coding everything)
There is an example of arrange node:
https://github.com/chaosprint/glicol/discussions/114
But it's not complete yet. I have indeed been considering chords, but the syntax needs more changes and design.
But according to your description, https://strudel.cc/ can bring you a lot of fun.
Compared with Strudel/Tidal, Glicol provides more on precise time control in audio engine, low-level sound synthesis and sampling capabilities (needs more development but completely feasible).
This is very cool. Unfortunately it's quite hard to use ChatGPT to generate new music, since i don't see good docs that it can use. It will be helpful to create a nice doc as prompt to use it to force LLMs to create custom music.
ChatGPT has at least been trained on a variety of textual musical notations. It will generate chord charts or drum patterns. In the past I've asked it to generate chord progressions demonstrating various music theory concepts.
It would be cool - 'haps - to use one of the many music/accompaniment algorithms available, adding "nuisance" or "humanness" through AI, creating a virtual musician.-
Hello, and thank you for creating Glicol—it’s an outstanding tool for live coding and audio synthesis!
I was wondering about the possibility of extending Glicol’s functionality to include features such as:
TTS Emotional Replication Generating human-like emotional expressions in sound (e.g., sighs, coughs, gasps). Sound Effects Simulation Creating realistic environmental sounds, such as thunder, rain, wind, and other ambient effects. Novel Sound Creation Enabling the generation of entirely new and experimental sounds, potentially by training on a provided audio dataset. Given Glicol’s modular architecture, Rust-based engine, and WebAssembly (WASM) support, I’m curious whether it would be theoretically possible to support such extensions within the current framework.
Questions Are any of these features currently possible with Glicol? If not, are there any plans or ongoing developments to enable functionality like this in the future? I believe these additions could expand Glicol’s applications far beyond music generation, opening up exciting possibilities in areas like:
TTS systems with nuanced emotional expression. Sound design for games, films, and interactive media. Experimental audio research and creative sound exploration. Before concluding, I’d like to add that Glicol’s potential use cases could extend well beyond music generation, offering new opportunities for innovation in audio-based applications.
I’d love to hear your thoughts on this and whether these ideas align with Glicol’s future vision. Thank you for your time and for developing such an inspiring platform!
To be honest, I am not an expert in TTS. However, the angle you mentioned does have many application prospects, but I feel that the project itself may be larger than Glicol :), and it may be provided as an independent rust package. The short-term goal of Glicol is to cooperate with other instruments and LLM.
> TTS Emotional Replication
What a tought.-
Velocity-driven audience coughs :)
PS. In jest. I do not mean to subtract from your substantive and interesting comment.-
In my ignorance, this puts me in mind of p5 and the like.-
Strudel [1] is similar (no affiliation), inspired by Glicol, is a bit more approachable in terms of syntax, has a great visualization in the editor (highlighting the sounds being played) and has a nice tutorial [2].
[1] https://strudel.cc
[2] https://strudel.cc/workshop/getting-started/
Afaik Strudel is a Tidal javascript port, so not inspired by Glicol since I think Tidal was available way before ? Both are awesome anyway.
Neat! I was wondering if it is possible to build something on top of this - like a synth sequencer? Or build a synth plugin as such?
yes see https://github.com/chaosprint/dattorro-vst-rs
also this one https://youtu.be/tmmBhBmIEW0?si=sRPDPsocDFfVNDyE
niceee! gonna try out it as soon as possible :)
How does this compare with csound?
Csound is legend. You can find almost all the sound effects in csound, and the score of csound is an exclusive feature.
Compared with Csound, Glicol has a completely different logic in syntax, which is much more lightweight and focused on sound synthesis and live coding. The core logic of Glicol can actually use "organized numbers" to make music. Numbers are used between nodes, and you can even "listen" to an envelope. Controlling these nodes is like having countless sound modules. The original idea of Glicol was that after I became familiar with DAW and modular synthesizers, I wanted to design a language that could simulate this workflow.
Cool project!
Note that ~ is a hassle to type in on many international keyboard layouts.