https://jdan.github.io/98.css/

In the meantime you can open the piano roll and set the "Show note labels on key" in the settings so that when it shows the key signature at the top "G Major", you can match it with the corresponding note label.

I'm going to be implementing number 1 - a "playground" that lets you sort of plink around until you're happy. I've gotten a number of requests for this.

The second request has me a bit confused, so I may be misunderstanding.

Under “Note hints” in the options, there are three different modes for practice. In all of them in Practice Mode, the very first note of the sequence is always shown in standard sheet notation. Then depending on the "Note Hints", you can basically change the exercise from an ear training exercise to a sheet reading exercise.

If you change Mode from Practice to Simon, even though it doesn't show the first note - it's always the root of the signature - e.g. if it says G Major, then the first note of the sequence is always G. I should probably make that more clear.

Does that description line up with what you’re seeing?

Maybe the answer really is just as simple as a little visual toggle that puts you in a "sandbox" mode where all the sounds still go through, but the game doesn’t respond to them until you untoggle it.

Side note but I'd love to see a nicely printed stack of physical cards with popular melodic hooks/jingles, the demonstrated intervals, notation, etc.

If you’d like to make things a bit easier, you can go into the options and restrict the key signature. That way you can keep it simple and just practice in a more common keys like one major scale like C major and its relative minor, A minor.

Where I really recommend "singing" out each note is when I'm teaching my students to improvise on the piano since it creates a sort of intentionality about what you’re about to hear and sing.

For example, if they had a chord progression or melodic idea in mind but accidentally played a wrong note, they’ll notice right away because what they’re singing won’t match what they’re playing.

Whereas if you don’t sing or whistle the notes as you play your instrument, you might not notice that you’re drifting off from what you actually intended to play because within the confines of the key signature it might still sound melodically acceptable (if that makes sense).

Another trick I like to do is take a popular song, rip out the melody, and keep the chord progression. Then I’ll usually scaffold a nice accompaniment using Band-in-a-Box so the student has something looping in the background while they try to piece out the original melody themselves on their respective instrument.

That can sometimes give them more guidance, since it locks them into a specific key signature and helps them feel the flow and explore the space.

Having basically zero knowledge about about Mojo though, it might be nice to place a FAQ section in the repo to expand a bit more on the decision to use it, how the port differs from Turbovision, etc.

I actually built a fairly unique edutainment app for the opposite skill set - learning to play by ear. It works with piano, guitar, and other instruments. I always tell my students there’s no reason to choose between sight reading and ear training as they complement each other really well. Link is in my bio if you’re interested.

I’d be curious to know what the existing apps were missing that made you decide to roll your own. Things like Sight Reading Factory, Read Ahead, Synthesia, Piano Maestro, etc.

One small bit of feedback: never ever ever use a metronome sound that has a defined pitch. In your app, it sounded like C and G respectively. That kind of pitched click can interfere with a player’s ability to internalize the music and can disrupt auditory processing. Instead, use woodblocks or other purely percussive sounds that don’t have a defined pitch.

I don't have a GH repo up for the TAS system yet - it's a bespoke mess right now since it was built with the old game "Castle of the Winds" in mind but I'll definitely consider it in the future!

https://en.wikipedia.org/wiki/Castle_of_the_Winds

Regarding the verifier that plays against the live engine, I’ve approached the problem from a similar angle by having LLM agents effectively borrow a page from the speedrunning community in the form of tool-assisted speedruns, allowing the LLM access only to a virtualized game controller.

[1] - https://store.steampowered.com/app/346850/Chips_Challenge_1

I will say, the act of eating creates a "bulge distortion" that flows down the length of the snake is a nice touch though.

With a book there's much less friction with taking a brief repose to ruminate, make connections and let your mind wander where it will.

Anecdotal but the way I’ve always made sure I can recall something is by pausing after anything I consider particularly salient and then trying to connect it to other things I already know. I deliberately use a kind of guided free association + mindfulness to weave a stronger web.

For example, I remember when I first read about the cake zuppa inglese, an Italian dessert that’s basically a soppy liqueur cake whose name translates to “English soup,” I immediately thought of the Coffee Talk SNL sketch: “The Holy Roman Empire was neither holy nor Roman nor an empire. Discuss.” which itself borrowed from Voltaire’s famous remark.

https://ebay.io/m/zzT1Xf

Honestly a lot of us who worked in the translation sector remember NMT as being a huge step up and in some language-pairs even surpassing DeepL at the time.

https://en.wikipedia.org/wiki/Google_Neural_Machine_Translat...