Steamed Hams but it's a MIDI piano roll 

**points at the band room** "Good Lord, what is happening in there?!"

@@deadrozes575 a concert.

@@SandNukes A Concert!? At this time year, at this time of day, in this part of the country, LOCALIZED entirely within your band room?

@@xenalin1 Yes!

@@karateka_95 ....may I see it?

Superintendent! I was, uh, just stretching my calves on the windowsill! Isometric exercise - care to join me?

Completely unrelated to the video, but I have that Lego figure in your icon!

@@ProfessorYana why is there smoke coming out of your oven, seemore?

@@evmoore22 Oh, that isn't smoke, that's steam! Steam, from the steamed clams we're having. Mmm, steamed clams!

@@ProfessorYana **Chalmers gets out from the kitchen**

i think the original is played underneath

@@juneguts it isn't

@@juneguts No, that’s the magic of converting things to midi. You should ask someone who doesn’t know steamed hams if he can understand it.

Yeah. I'd like to hear a similarly vocoded Simpsons clip that I haven't seen before just so I would know do I actually recognize the speech or not.

@@juneguts I agree with you.

I prefer 0:46

A midi conversion is just a severely limited version of the spectrogram of the original audio, so a good enough conversion program should be able to be recognisable. Though in this case the original audio is still being played.

Do we have confirmation that the original audio is being played? People can experience auditory hallucinations from these midi transcriptions. I forget the name of the phenomenon but Adam Neely talks about it in one of his videos.

assuming the midi bars are remotely accurate, there are numerous times where there are new words being said but no new notes. The first sentence is a good example, as at the start, the midi roll is quite complex, but then it gets almost empty, yet the dialogue is just as audible. Agnes's lines are another, as despite having deeper tones to her voice, the midi only has the higher tones.

@@KingBlonde the "Midi Illusion" thing is a myth, if audio is "converted to midi" it will be recognizable as it's literally turning every sound into a midi key, there's no illusion, it's what you're actually hearing

@@negljbreakergaming nope you've misunderstood

the most underrated comment

Omg 😂

*It's all piney*

@@alexb0nd *Take that, nature!*

The Super Nintendo's gonna need his medication

Hi, Supernintendo Chalmers!

Ralph Wiggum or Wreck it Ralph?

I get a feeling you never used a Super Nintendo.

@@Jtmissile This video's a Simpsons clip. You finish the puzzle

i love this instance of random optimism. makes you appreciate life.

I was steamed in the right generation.

@@ryanhernandez8324 that's more of an Albany expression.

*timbre

@@captprostate Thank you, I had no idea I'd been spelling that wrong this whole time 😅

Nah I'm not seeing it

@@becauseimapotato7599 the rhythm

True

And at 0:55 you can see a trollface

​@@This_handleHOLY CRAP ITS TRUE

Truly

Thunder Enforcer they used a flute sample, which is better than piano since it's closer to sin wave

*phonemes

Are you sure the actual audio isn't layered in with this though? The tone would be completely different if it was just the midi piano, it isn't

@@maJastoL It’s actually just the midi piano without any layering. You can tell for things like the yawn or Chalmers yelling “SEYMOUR!” Where they have a sort of auto tuned feel to them. You can hear the same effect with other midi conversions of videos too.

@@tomlawson4713 That's crazy, sounds like it still has the timbre of their actual voices underneath

It's basically additive synthesis.

​@@maJastoL That's because timbre and harmony are basically the same thing. All natural sounds can be represented as several pure tones playing at the same time, that's the basis of additive synthesis, a synth technique where timbre is created by addition of many-many beeps with different frequencies (and it's also pretty much the core concept of all audio theory), and this means that you can also approximate timbres by playing chords on instruments. Here's how an mp3-to-midi converter works: it does harmonic analysis on the given audio, finds the most prominent overtones, and creates MIDI notes out of those frequencies. The end result will obviously sound extremely distorted, because a ton of harmonic information is lost during this process, as the resulting file will only contain notes from the chromatic scale as opposed to the vast frequency range of original sound, and there are also tons of additional harmonies added by the own timbre of the given MIDI instrument. Still, the mangled skeleton of the original harmonic content remains intact within all this mess, so that's why you can hear it: because it is there, despite everything.

"Yes, and you say that everybody's got a water buffalo despite the fact they obviously do not"

@@stardestroyer81 “Yes, and you are looking for your hairbrush despite the fact you’re obviously bald.”

not really...

*EEEEEEEMORRRRRRR!!!!* xd

Yeah, of all the voices I tried it was the one that was the simplest waveform (and thus preserved the original sound the best)

@Jordan may I see it?

@@captainsimon1 no

Timestamp?

Why isn't it pinned?

@1:13

I don't remember watching this and I don't remember leaving these comments but I'm not surprised by any of this.

Yes

from the movie that is

It's pure midi, your brain fills in the gaps. It's a famous illusion. If you play it near your grandma or some other poor soul who's never seen steamed hams before she'll probably hear only a ton of gibberish, without being able to get as much you do. Only the initiated ones can.

@@yggdarsilyae6807 its not /exactly/ an illusion! all sounds including clarinet and human speech are made up of sine waves. this algorithm represents the original sine waves as closely as possible with the sine waves of stacks of clarinets, which isnt perfect by can get pretty close just by balancing missing and extra information equally. you are hearing like 80% of the original signal being accurately recreated, which is even enough for people who've never heard steamed hams before to make out most of the words. this also has 20% of random new signals which yes do add a little confusion to new listeners. point is, your brain isnt making up the human speech that you can hear, everyone can hear that base 80% the same, but it /is/ fluffing the gaps a bit from memory to make it easier to make out, by swapping the 20% clarinet artifacts out for the 20% of the original signal, as best as it can, which is like maybe 5%. like you still hear it sounds wrong. but its enough to make out all the words from memory when others might miss them.

It has a simpler waveform, so there's less incorrect information being encoded. Just lets you hear the pitch of the audio without as much extra junk to get in the way of it.

The short answer is that MIDI pitch bends are a little finicky -- there's a global multiplier applied to how much pitch bends scale, which restricts the ability to sounds to be perfectly replicated by the MIDI data. When more than one sound isn't totally 'in tune' with the rest of the audio on an equal-tempered scale, stuff starts getting approximated.