avaku

11 ms·

1. 1.▲
   by avaku 7mo ago

   Exactly
2. 2.▲
   by avaku 1y ago

   This looks very useful! I wonder how other AI coding tools manage merging work from different agents (I've never tried running multiple agents simultaneously yet, but I was thinking about that).
3. 3.▲
   by avaku 1y ago

   You're right, your comment is valid. I will make better blog posts in the future :)
4. 4.▲
   by avaku 1y ago

   I made this blog post, because I did very little fudging and was shocked by how good the results are. However, this only happened with the latest model (o3). That's why I wanted to describe my approach. First, give AI the model /
5. 5.▲
   by avaku 1y ago

   I am sorry for not making a long blog post with all the details, I just wanted to highlight how the cooperation with AI went to the next level (model design, then implementation). At least I posted the code :) It's just a mock up, and
6. 6.▲
   by avaku 1y ago

   There is definitely a ceiling problem. However, I found that with smarter models (such as o3) you could first discuss the mathematical model (or system architecture), so that it knows the context. After that you can ask it to write code, an
7. 7.▲
   by avaku 1y ago

   I thought that the full details might not be interesting for people, since the algorithm is just a mock up of the idea, but which already works! If I wrote the whole post about all the details, it would be very long. I can see how it's
8. 8.▲
   Co-designing a sparse music codec with ChatGPT o3(akuz.me)

   52 points by avaku 1y ago|22 comments
9. 9.▲
   by avaku 4y ago

   Great idea. I've read some people's comments. There are a lot of companies in the UK that sell their equipment, and use contractors to install them (Worcester Bosch for boilers, even Tesla uses external contractors to install the
10. 10.▲
    by avaku 5y ago

    I think they want everyone to move to Metal Performance Shaders. I've done some stuff on them, but not nearly as developed as CUDA.
11. 11.▲
    by avaku 5y ago

    I am still using this now, after 3 days, to read hacker news. I hope this project stays alive.
12. 12.▲
    by avaku 5y ago

    Reading this post on Haxplore itself is even better :)
13. 13.▲
    by avaku 6y ago

    This seems certainly possible. But I don't know how would you display all of the frequencies that are present in music at the same time . Usually, they only play one or two frequencies and look at the result. Mash them all together? M
14. 14.▲
    by avaku 6y ago

    One could imagine making a “fake” cymatic visualisation while still looking interesting!
15. 15.▲
    by avaku 6y ago

    Hi! Thanks for your comment! It’s interesting. I don’t think there is any magic about cymatics per se: it depends on how sound interacts with physical objects, and these objects have certain resonant frequencies, which is when the start to
16. 16.▲
    by avaku 6y ago

    These are very interesting thoughts, they remind me of when I’m deeply thinking about how a track/song is represented in my mind and trying to visualise it. (With the exception of red being a “do” note, I think because I’m red-green co
17. 17.▲
    by avaku 6y ago

    I’ve actually wanted to make this in Unity! But they don’t provide (reasonably useful) access to real-time data from microphone! I’ve also been contacted by a recruiter from them by accident, so I’ve sent them my feedback about this :) I’m
18. 18.▲
    by avaku 6y ago

    Thanks for your suggestions!
19. 19.▲
    by avaku 6y ago

    In the “Zoom” settings try changing the Equalizer to “Flat” setting (on the right), then low frequencies will be more visible!
20. 20.▲
    by avaku 6y ago

    Got it! Thanks! I will reply soon! (On the phone long time today, wanted to write a proper reply.)
21. 21.▲
    by avaku 6y ago

    Thanks!
22. 22.▲
    by avaku 6y ago

    Never heard that word before :) I'll google!
23. 23.▲
    by avaku 6y ago

    Thanks a lot for your reply!
24. 24.▲
    by avaku 6y ago

    Thanks for comment. I think this can be useful to music producers (in some way), but the market for plugins is so tough, basically hard to survive even for one developer on revenue from one plugin. I’ll be thinking about this.
25. 25.▲
    by avaku 6y ago

    I will add extra credits on the page as well.
26. 26.▲
    by avaku 6y ago

    Of course. They are definitely tagged below each Instagram posts, and I also asked them in person when I made those videos.
27. 27.▲
    by avaku 6y ago

    The pictures you have there look similar to what I see when I visualise the spectrum over time (some examples here: https://vsound.app/high-precision-in-frequency-domain.html ). I found that white noise was quite interesting
28. 28.▲
    by avaku 6y ago

    Note at the moment, but I will work on this. Thanks for feedback! P.S. It looks much better with actual audio rather than microphone, see Instagram videos on app website.
29. 29.▲
    by avaku 6y ago

    Thanks for suggestion! Will definitely add in the next release.
30. 30.▲
    by avaku 6y ago

    Ok, I try. We use Wavelet Transform instead of FFT because it provides suitable resolution across all frequencies. P.S. I've realised that I implemented Wavelet Transform after I've already done it :)