My journey into the world of digital art, from manipulation to AI

Update (May 2026): This article was originally published on February 20, 2025, and reviews the original Riffusion tool. Since then, the Riffusion technology and team (Producer.ai) were acquired by Google, and the service is now known as Google Flow Music . This post has been preserved in its original form as a historical overview of the early days of generating AI music using spectrograms. Riffusion is a powerful tool for generating music from text prompts. It leverages Stable Diffusion, a machine-learning model originally designed for image generation. Riffusion works by creating spectrograms, which are visual representations of sound, and then converting these spectrograms into audio clips. This innovative tool allows you to generate music in a variety of styles , even creating pieces inspired by specific artists or songs .  Riffusion is a free and open-source application , accessible online or downloadable for use on your own computer. With Riffusion, generating music is as...

 In recent years, a new form of digital art has emerged that combines deep learning algorithms with creative expression.  The process of creating these images involves inputting a prompt into a deep learning algorithm, which then generates a unique image based on that prompt. The artist can then choose to leave the image as-is or enhance it by adding details. Here's my experiment with the Midjourney program.  Viime vuosina on syntynyt uusi digitaalinen taiteen muoto, joka yhdistää syvän oppimisen algoritmit luovaan ilmaisuun. Näiden kuvien luomisprosessiin sisältyy kehotuksen syöttäminen syvän oppimisen algoritmiin, joka sitten luo ainutlaatuisen kuvan kyseisen kehotuksen perusteella. Taiteilija voi sitten päättää jättää kuvan sellaisenaan tai parantaa sitä lisäämällä yksityiskohtia. Tässä on kokeiluni Midjourney -ohjelmalla.