Posts

• social/facebook

  3 Feb 2021 [https://www.facebook.com/10111169199177213/posts/10108595171683943]

  ---

  Ooooh, did I hear that right? Is LAPOMPE playing a show at the The Oriental Theater? Go ahead, tell them about it, Dave!

  Dave: “Well, lemme tell ya… https://www.facebook.com/events/396331334767299”

  That’s right, Dave. It’s been a long time since we’ve played out and we’re starting to miss you all, so if you’re missing us too come hang out for a set or two!

  Dave: “Reminder this event will sell-out. Since its super limited seating. First Set: Django Classics, Second Set: Originals. Or get a discount for coming to both!”

  Thanks, Dave. Hope to see you all there!

  Dave: “Hope to see you there!”

  I just said that…

• social/hackernews

  16 Dec 2020 [https://news.ycombinator.com/item?id=25447671]

  ---

  Tactics trainers:

  https://lichess.org/training/

  https://chesstempo.com/chess-tactics/

  https://www.chess.com/puzzles

• social/reddit

  15 Dec 2020 [https://reddit.com/r/MLQuestions/comments/kdmeqq/project_acoustic_event_detection/gfyu8vp/]

  ---

  Your model will only be able to properly operate on input of the type it was trained on. You can’t throw full recordings into a model trained on short clips and expect it to perform the same.

  If you want to be able to dump a full-length recording in and have all possible sound sources labelled, then you need to train the model using full-length recordings as input, with all the sounds labelled. Something like that would likely be more difficult to train than your simpler model, but that’s a separate issue.

  If you want to use your model that was trained on short single-sound audio clips, then the input needs to be short (potentially) single-sound audio clips. Maybe you can do pre-processing on the full recording to make the input fit into your current model… For example, segment the full recording into overlapping 2-3 second chunks, predict the contents of the chunks, and then aggregate them into a composite prediction for the full recording. Or maybe you want to use some other technique to identify sound onsets to do the segmentation more intelligently.

  I think a pre- and post-processing pipeline in your case would be easier than trying to train a single model to do it all at once, but that’s up to you.

• social/facebook

  22 Nov 2020 [https://www.facebook.com/10111169199177213/posts/10108427804693733]

  ---

  happy bday dude! - check out this ultra throwback that fb auto suggested for me…

• social/reddit

  17 Nov 2020 [https://reddit.com/r/jazzguitar/comments/jvsng5/a_beginners_path/gcmd7ml/]

  ---

  I have a lot of thoughts but I’m going to try to limit myself to three big ideas:

  1. If you don’t listen to jazz regularly, you will never be able to learn to play it. Just like you can’t learn to speak French without listening to anyone else speaking French. Finding the jazz you like, and listening to it regularly enough to get it into your ear, is more important than everything else you’ve listed there. It’s foundation upon which everything else is built.

  2. The theory/harmony/melodic ideas/chord voicings that you’ll want to learn depend a lot on what kind of jazz you’ve decided that you like, and that you are listening to regularly. You can play entire 3-hour sets of jazz music using only “cowboy” chords and barre chords, and it can sound fine if you’re playing old time swing jazz. If what you want is a “hipper” sound with more extensions, then well.. how hip? Bebop hip? Post bop hip? Fusion hip? Free jazz hip? Each collection of phrases and voicings is a dialect in the language of jazz, and there’s no point in learning a bunch of stuff just because someone else said that’s what you need to know, if it doesn’t apply to the sound that you like. Older jazz might be a good place to start because it’s easier to pick up, harmonically simpler, and all the rest of jazz is rooted in it… but it doesn’t matter how easy it is to pick up if you don’t like how it sounds, so you don’t listen to it, so you never get it in your ear!

  3. Jazz is just a musical language, like any other, and in the big picture it’s no easier or more difficult to learn than any other musical language (metal, for example). Everything you know about learning metal applies directly to jazz. You learn it the way you learn any other language - by listening and imitation at first, and slowly developing your own voice over time. Theory and structure can help facilitate that process, especially when you start as an adult, but maybe like 10%, where the real 90% is listening and imitating. The difference with jazz is that it seems to be this holy grail musical language that everyone wants to be able to play even though they’re not actually interested in listening to it. When you come here and essentially ask “I want to play jazz, but I don’t really listen to it”, as many people are essentially saying when they start (and no hate from me, that’s totally understandable), I think what you/they really should be asking is “what kind of jazz do I like, and where can I find more of it to listen to? where can I find enough enjoyment in jazz that I’m inspired to just start learning those damn songs instead of over-analyzing what that learning should look like?”. And nobody can answer that question except you. I can give you a dozen resources on jazz standards, and chord voicings, and scales, and theory, and you could find a dozen resources by googling for yourself. But in my opinion, that is completely beside the point. What you need to do is just LISTEN TO MORE JAZZ! Find the jazz you like, and then find more, and listen to it more. When your question goes from “How do I learn jazz?” to “How do I learn ?” then you’re where you need to be. And honestly once you’re there, I assume that based on your experience with metal, you’ll be able answer that question yourself (with the help of the internet I’m sure).

  I don’t mean any of this disparagingly by the way, I mean it to be (I hope it is) encouraging. Learning jazz is really simple, and I think you’re just waaay overthinking it, as most people do with jazz (and some gatekeepers encourage). Most of that stuff you listed above comes later. The hard part is finding the stuff you like, and then just sitting down and figuring it out, piece by piece. It’s the enjoyment that keeps you going, figuring out the pieces. Without the enjoyment of the music, you’ll never get anywhere… but if you don’t enjoy it why would you want to learn to play it?

• social/facebook

  12 Sept 2020 [https://www.facebook.com/10111169199177213/posts/10108272067412493]

  ---

  Here’s the thing about Meadow Mountain… They just make beautiful music. If you’ve got some time, take a listen to their existing album and maybe help out with their next one by contributing to their kickstarter.

• social/facebook

  29 Aug 2020 [https://www.facebook.com/10111169199177213/posts/10108239567442693]

  ---

  mmmhmm, you two have always been my favorite vocal duo that lapompe has had

• social/reddit

  27 Aug 2020 [https://i.redd.it/gs5mei9xxlj51.png]

  ---

  Here’s a rough transcription of that “Clyde Stubblefield influenced drum groove” from earlier by xCerealKilla

• social/facebook

  5 Aug 2020 [https://www.facebook.com/10111169199177213/posts/10108183247248813]

  ---

  I’ve already heard the tune (woot) so I know it’s awesome (especially the bubbly bits), but I’m hype for the video because I’ve heard it involves a news person going crazy…? and you know how much I like crazy news people.

• social/reddit

  18 Apr 2020 [https://reddit.com/r/creativecoding/comments/g359fn/oc_vasarely_made_with_processing/fnspnxq/]

  ---

  As a general rule, almost everything that can be done in Processing can be done in P5. They are extremely similar. Here’s the code from above, converted to P5:

  https://editor.p5js.org/kwertyops/sketches/_ULHNH8Hc