Not exactly a trombone. In a trombone you decide what note do you want to play by slider and LIPS cause the higher "buzz" you make by mouth the higher will be the note on the trombone without moving the slider. I think it would be hard to make that "lips moving" by machine. In slide whistle you just blow. And yes. I know that a joke but i wanted to explain this
Tim: It's time to show off yet another project I did years ago and didn't tell anyone about. Me: It's time to show off yet another project I just started and will never finish.
I wonder if posting just the songs separately with links to the full video would be a good idea to draw in the fans of "[Song] Played With [Thing]" videos
Not gonna lie, that's precisely what made me think of it too That and I feel like it would simply be easier to show someone "this cool video of a song being played in a unique way" without the extra detail that more casual audiences wouldn't particularly enjoy. I mean, I personally love it to no end, but many people I've shared videos to haven't shared my enthusiasm
I was the same too, cruel angel thesis caused me to pause and immediately share the video, this does make me wonder how much of the audience had this exact reaction
Your demeanor is exactly the right brand of eccentrically maniacal that makes this whole thing top notch. You deserve more subs. Can't wait to see what you do next!
@@crides0 i know how to download and convert youtube videos but maybe you have heard of RU-vid compression and how it waters down the original sound quality.
@@duncathan_salt Well it depends if we're talking woodwind specifically or not. A trombone is a brass instrument and not a woodwind. They don't occupy the same niche.
I am in AWE at the linearity of servo rotation to pitch. At the beginning of the video I thought the servo pair with linkages was a strange actuator choice and I was thinking about it all the way up until you started talking about the linearization, and then when you showed the graph my jaw was on the floor. Brilliant!
"Keywinds"? Or well, "Brasskeys" or "Keybrass" in that case. Wind or Brass instrument's rigged up with a keyboard like a "Keytar." Would needa getup like marching with a bassdrum or Xylophone. Imagine the poor fella wacked by a motorized trombone slide though.... :d
I love the idea but I think trombone is nigh impossible since the slide doesn't cover all the notes. You need complex embouchure (lip tension manipulation) to switch registers (or even notes in the same register) and even the basic lip vibration technique would be really hard to recreate artificially. I doubt anyone could make a working midi trombone.
@@tomthepom98 I applaud your creativity but I'm afraid that the embouchure is different per note. I used to play a little (self taught, still a beginner) a while back so I picked it up today to feel how it worked again. Once I started playing, I remembered: you need to buzz your lips at the same pitch as what you're playing with your slide. It's kind of like how you can make a balloon squeal at different pitches if you deflate it while pulling the outlet tighter or looser. I hope someone proves me wrong and makes it work though. It would be so awesome.
@@TaylorTheOtter You don't need to replicate lips buzzing, you can just use a small, powerful speaker to pump a sound of the correct frequency into the tube where the mouthpiece goes. The trombone doesn't care exactly how you make the air inside it resonate, just that you do. Instead of a normal mouthpiece, you could fabricate a horn bell shape that attaches to a small driver on one end, and tapers down to fit into the trombone at the other. Servos or a linear actuator to move the slide and a small amp and tone generator to drive the speaker and you've got yourself a MIDI trombone.
Won't it make sense to add a piezo element to the whistle and monitor the current pitch and adjust the position slightly in a PID loop, therefore solving both the backlash and the constant need of re-calibration?
I was thinking that if improvement was desired, step 1 would be to eliminate most/all of the backlash by going to an o-ring sealed metal (or plastic, or, heck, why not wood) piston. Once you take the slop out the feedback technique will work better - probably just a quick pre-performance automatic tuneup. I'd think closed loop while playing might make wobbly or 'chirpy' notes where you'd have a pitch transient at the beginning. Now that I think about it, a good analogy is the calibration strategy for the Volca Keys - it plays silent notes at itself while idle to keep the VCOs tuned, but doesn't while you're playing to avoid note-stepping. (the upshot is if you start playing before it's tuned up you'll get interesting results).
I really love how you structured this video with the explanations in between performances, as soon as I felt the video was over I was blessed with another ballad
I'm so glad the algorithm brought me this video. I was sold as soon as you started playing Star Wars, but then having FOUR SLIDE WHISTLES SING BARBERSHOP?? I had to leave a comment omg. Going to send this to all my barbershop friends :P
Thank you for sharing. I have listened to the last piece probably a hundred times, and I know that my hours of enjoyment pale in comparison to the hours you must have spent on this. The mechanics are beautiful in their apparent simplicity and function. Thanks again.
That’s an amazing project! And can I say superb arrangement of Mr Sandman too. As a barbershop singer, I’m sure I could detect “just” intonation used, as I’m sure I heard the odd ringing seventh harmonic. Well done 😀
I wonder what would be the best way to calculate the frequency. FFT might be overkill and I think one might already get decent results with autocorrelation.
@@kwinvdv I don't do signal processing so I might be wrong but isn't FFT the fastest way to do autocorrelation, meaning that we should just use the FFT directly anyway?
@@PascalSommerMovies I initially assumed that autocorrelation would be cheaper to calculate then FFT. However, according to wikipedia a fast algorithm for calculating the autocorrelation uses two FFTs. So FFT indeed seems to be the way to go.
As in.. Death by midi. (Also, a neverending stream of this, especially with machine learning (aka little repetition to hold on to) would probably qualify as 'torture' according to the Geneva Convention ;) *[hence, the Saw movie plot]) As cool as this thing is, a continous barrage of this sound.. Is something that I would struggle with ;) . ("Tell me ze locationz of ze missilez, Mr. Bond.. Or ze midi stringy thingie will play its own interpretation of Swan Lake...")
That might be difficult to implement because "good" or even just "tolerable" music is subjective to the human experience. So there's no really good way to train a Neural Network or other types of machine learning without a large database of examples. I guess you could just have it train on your favorite Spotify play lists but that again might wind up just making random notes without constant user feedback
This is brilliant. As an electro mechanic instrument builder myself, your work is super impressive. Thanks for the great video- so glad to have found your channel!
Here’s what you’re rewatching the video for, feel free to help me with the names of the songs I got wrong: 0:45 Tetris theme 1:01 Star Wars main theme 2:07 God only knows - The Beach Boys (Bioshock Infinite version) 6:05 Banjo & Kazooie theme 6:29 Dragon Roost Island - The Legend of Zelda: Wind Waker 10:05 Cruel Angels Thesis: Evangelion 12:20 Mr. Sandman - The Chordettes
Dude has a face that is a mashup of several famous actors. I can see Nathan Fillion, Harrison Ford, Pedro Pascal heck even Kevin Smith all combined into one.
I mean slide whistles doesn't look too hard to make other than getting the mouth piece to work properly so if there are no bass slide whistles out there, making one out of pvc wouldn't be the end of the world
I discovered your Channel through this video. And have been binge watching everyone of your videos. Haha I find myself sad that everything was so long ago! Don’t stop uploading again! Pleasseeeee