Resolves YES if already happened.
I tell the robot to play me <insert popular violin music>, and it does, fairly decently (it searches for the notes or whatever on its own, must be physically manipulating the violin, not digital synthesis of the music)
by 2025 = end of 2024
Related questions
@CarsonGale You know what? I will take it, after seeing the figure 01 performance, I'll take a bit of risk
@Seeker figure 01 was impressive tho! But yeah I think we still have time before that fine of motor skills are possible
@firstuserhere I became aware of this market and bought a bunch of YES (nothing happened in the news that I'm aware of)
@extent_of_foxes ah i see. Thanks for clarification. Ofc it's spiked because you bought YES, but was wondering why you bought yes.
Shifting can be avoided and glissando and vibrato are definitely solved problems - see this three-person project from a year ago. The whole thing lacks finesse due to awkward bowing technique but we are clearly on the right track and 20% feels low. https://nime.pubpub.org/pub/i9nvg1yi/release/1.
Note that programmatically manipulating a violin to make a pleasant sound is a problem solved before transistors were invented, albeit only one string per violin and no glissando etc.. If a robot played a requested tune using multiple violins, how would this resolve? https://www.classicfm.com/discover-music/instruments/violin/features/self-playing-violin-orchestrion/.
The ability to shift, use vibrato, and perform effective bow techniques seems extremely far away in the field of robotics. Similarly I don't believe there is any effective technological method of imitating musical phrasing well.
A Yes resolution would require both incredible advances in fine motor skill robotics and moderate advances in AI musical phrasing capability. I would be more convinced as to the latter if an AI, using only the electronic sounds that a violin can make, could realistically imitate a great Concerto.
There's a big difference between a robot that can play e.g., Schindler's List vs the Sibelius Violin Concerto. What level of proficiency are you looking for? Does Mary Had a Little Lamb count?
My proposal is the Mendelssohn Violin Concerto as a difficulty benchmark.
The level of finesse to play some of the more renowned literature is extreme. I've been playing for 23 years and I still can't handle the more advanced material.
@CarsonGale There has already been a robot that can play very simple pieces https://youtu.be/auxjSYCKL4s
+1 to Mendelssohn Concerto as benchmark
@IC Wow, thats great! I am happy to take suggestions for benchmark, but i listen to violin in the background while studying and am not really familiar with individual pieces much; and so, would you think playing at least 3 (or 2?) out of these 20 would be a decent difficulty? @CarsonGale https://www.classicfm.com/discover-music/instruments/violin/best-pieces-ever-written/
@firstuserhere I think 3 or more of these pieces work well for a benchmark. Schindlers List and Lark Ascending are probably the easiest pieces technically (though I don't know the Beach / Boulanger / Georges). The rest are very challenging and would be a good benchmark.
@Feanor can I trade with the understanding that 3 or more of the pieces in that article is the rough benchmark? And is there any clarity you could provide regarding the quality of the play? (e.g., lay person can't tell the difference in a blind playing test or something)
@CarsonGale Yes, you can trade with that understanding, I should've said so instead of just liking the comment. The quality should be that a lay person isn't able to tell any significant difference most of the time. This is mostly a qualitative test and I'll use my subjective taste to evaluate this if there is no survey. Anyway, i will resolve only according to the spirit of the market and not anything which "technically meets the criteria"