Notorious Optimist. Designing @MireloAI Previously @Microsoft, @Pitch @Wunderlist
Berlin
Joined January 2011
- Tweets2.3K
- Following623
- Followers1.8K
- Likes12.8K
TOM B✪GNER retweeted
Our best Audio-to-MIDI model just got an API.
Last week we open-sourced 3 MuScriptor models for multi-instrument transcription, each with different speed vs quality tradeoffs.
Today we're putting our best version behind an API, so you can build it straight into your own apps.
The API gives you the highest accuracy with nothing to host or maintain.
And if you'd rather self-host or run things locally, the open-source models aren't going anywhere.
I look at portfolios day in, day out right now and @juliusgehrig s is a breath of fresh air.
Simple. Modern. Exceptional attention to detail and craft.
If you're looking to hire an outstanding designer AND someone who's one of the kindest people I know: Reach out!
TOM B✪GNER retweeted
Today, together with @kyutai_labs, we’re introducing our new Audio-to-MIDI model.
It takes a finished recording, identifies the instruments playing, and returns separate MIDI tracks for each — voice, drums, bass, keys, and more.
Unlike most existing solutions, our model works directly from the full mix rather than requiring separate stems.
It also detects chords, key, and tempo, giving producers broader musical context.
We’ve written more about the model, the problem, and how it works here: mirelo.ai/blog/turning-audio…
I'm hiring!
We're looking for a Product Designer to join @MireloAI — Berlin, London, Tübingen, or EU remote.
This is a rare opportunity to join a European team building both its own foundational model and applications on top of it.
You'll shape how video editors, sound designers, and creators generate audio — from SFX to scoring.
Product experience and a strong sense of craft are required. Interest in generative AI, and vibe coding is a huge plus.
Sounds interesting? CV & Portfolio to [email protected]
The new Apple SF Symbols Beta features animations now.
Buckle up... User interfaces are about to become a lot more dynamic and expressive:
developer.apple.com/sf-symbo…
Apple is finally moving away from the rather low FPS .png-sequences to render motion on their websites:
"One of our team members broke all the rules and it was dope..."
Big tech got so obsessed with rigid design guidelines over the years that we went full circle and have to create videos to remind designers that design is in fact, about creativity:
youtu.be/vnDhq8W98O4?t=1127
Over the years, I've created dozens of product videos. Typically, I start with a computer-generated voice-over before recording a professional version.
But with this latest video, the #AI-generated audio was so impressive that I didn't feel the need to use a human voice-over:
Just discovered the power of combining Midjourney images with AI-generated depth maps to create surprisingly stunning animations:
(I added some rather cheap visual effects to hide some flaws you'll notice at closer inspection)
Midjourney introduced a new 'describe' feature, allowing you to convert images into words.
The generated descriptions (or prompts) enable easy creation of new variations.
I tested it with visuals used in our latest designs, and here are some results:
TOM B✪GNER retweeted
From the first time I saw this work by @dasTom I was fascinated. I wanted to show that creating it using #threejs is not as difficult as it looks. It has only 240 lines of code. The live link and the source code will be published this week. #creativecoding #webgl #threejsJourney