All Activity

1. When it comes to platforms like Suno, that should be regulated: it should not be used in competitions, detection should be used, and if there is abused it should be banned. When it comes mock ups in the Daw or in notation (Note perfomer, vst, etc): For the composer plays the notes into the program, and such. Sometimes, we do get lucky and have live performance. I think it would be cool, if there was sub forum to share our live performances. 2. A pianist and composer, I'd prefer if someone who did not use AI. 🙂 3. I agree with @UncleRed99 Idea. In traditional music, you can generally see the key, meter, and etc, but with lyrical songs we can't see that.

Absolutely! It's a real pain in the arse to do in Musescore though 😄 It's added to my to-do list (along with some dynamics shadings). I'm acutely aware of the problems of creating actually good-sounding music and I'm very grateful for you pointing this out. One of the nice things about the Baroque ritornello theme in a vocal context, is that your theme does not even need to be singable. The tenor aria from BWV 81 provides a particularly striking example. A common and easy solution for this problem is to just have your soloist sing an independent melody in counterpoint with the ritornello theme whenever it is reused. I've attached an example of this technique for my aria.

My view on the matter is that while any pieces of "music" that are generated using tools like Suno should be completely banned, the use of AI tools like Cantai to generate mock-ups from midi or notation should be allowed as they are not really any different from generating a mock-up using sound samples (as others have already said). With regards to the issue of how we should go about detecting AI "music", I like @UncleRed99's idea, though I recognise @TristanTheTristan's concern that asking for technical music theory things like starting keys/time signatures may unfairly exclude composers who do not work in theory/notation heavy genres. I propose that we require users to write a short description of their piece naming some musical elements/techniques that were used (e.g. chord progressions, instrumentation etc.) alongside any additional details they might want to mention, like maybe what inspired them?

Probably, but it's not actually doing anything that isn't already a standard feature really in drum kits for years. Premade beats. I would point out that on marketplaces like gamedev market, they forbid AI art assets, but allow composers to use AI cover art.

Hi @muchen_! I am not an expert on the Baroque style, recitatives, arias nor cantatas, but I feel like the recitative could have been more humanized through liberal use of hidden tempo changes in order to make it sound more like an accompanied cadenza and in free time which I'm assuming that that's how recitative are supposed to be like. This rendition sounds very metronomic which hurts my impression of it. I like the aria a lot though! Although between bars 8 - 14 the singer doesn't seem to have any opportune place to breathe and it's hard to imagine those 6 measures sung in one breath although I could be wrong. Other than that I really enjoyed it though! Great job and thanks for sharing!

Hi @PaavolaPyry and welcome back to the forum! I think the solo Clarinet line dancing microtonally around a single central tone in the beginning sets the stage well for the piece that follows and is yielded from that beginning line. I feel the piece is very logical and cohesive. I can listen to it many times in a row and not get bored of it - which I usually cannot do with an unaccompanied solo piece. But this isn't unaccompanied and only starts that way which is a very entrancing way to start a piece that eventually becomes a Wind Quintet (although for some reason the instruments aren't labeled in your score? Assuming that it's for Flute, Oboe, Clarinet, French Horn, and Bassoon). The harmony is dissonant but in a deliberate way that is very digestible (at least to me). It somewhat reminds me of the slow Adagio movement of Mahler's 10th Symphony (the one movement he actually finished himself). The microtones on the Clarinet don't seem to disturb the harmony but enhance it. It's around the 4 minute mark where your Clarinet melody reminds me of Mahler 10 (measure 51). I am a Clarinetist myself but haven't played in a while and wouldn't dream of using these extended techniques that you've employed here - but you're informed by your own experimentation so by all means keep going! Great job and thanks for sharing!

This is a secco recitative in the late Baroque style for alto, part of my cantata WIP. The text is taken from Goethe's Die schöne Nacht. An example translation of the text can be found here. I have also attached a little preview of its paired aria 🙂

Thanks, @Henry Ng Tsz Kiu, for your kind words. I took up your suggestion and tried to compose a fitting end for this set. So without further ado: a draft for a fast-paced 2/4-Presto with the basic structure |: A :| b1 b2 b3 |: A : |: c1 : | c2 |: A :| The audio file needs some polishing. Any comments welcome.

Thank you, @J. Lee Graham and @Kvothe for your comments which are very inspiring and stimulating for me! Since you don’t find my 10/8 D lydian fugue exposition to weird, I’ll go forward with it and will share my progressions here as soon as possible.

1. How do you feel about including AI generated music on YCF? In my opinion, AI generated music is misplaced in the general submission threads and competitions/events on this forum, since the purpose of the forum is to exchange about the composition process, results, difficulties etc. (and not about which music one likes or not). However, to ban it completely would mean to neglect a tendency which is now existing and will further increase in the future. Thus, a special, separated sub-forum could be a compromise, but I have to admit that @muchen_ 's objection and demand for quarantine cannot be dismissed out of hand. 3. If we decide to regulate AI music, how should we detect it? I am (strongly) opposed to the use of online AI detection tools, as they can lead to “false positives” and, in my opinion, their use contradicts the goal of publishing and reviewing only high-quality, human created compositions in this forum. Therefore, the judgement can only be done in the same way as requested for the submissions. With this background, the distinction between genuine compositions and compositions generated (wholly or partly) by AI can only be achieved by making the composition process used by individual members more transparent. One idea could be to create a special section in the forum where each member can post their own topic explaining in detail how they normally work, including the notation software they use, the type of recording (either live recording or with which DAW software), etc. I think anyone who intends to regularly publish works in the forum would appreciate it if others could take a look behind the scenes, so that we could learn a lot from each other. A regular submission of a piece could then include a link to this explanatory post and should, of course, include a PDF score, an MP3 audio, possibly an intermediary MIDI file, a brief description of the musical form, key, time signature, instrumentation, etc., as well as a little background information about the inspiration and idea behind the piece. I think, that if we could encourage members to post their contributions in this way, the overall quality of the forum would improve and, more importantly for the submitters, it would be easier for reviewers to familiarize themselves with the pieces, which would certainly lead to more and profound reviews. (I say that also with the background that some members tend to „flood“ the forum with a dozen of uncommented posts at once – which might be fine but not necessarily leads to many or instant reviews. Sometimes, less is more.) I don't know if such a procedure could be introduced as a “rule,” perhaps more as a “recommendation.” However, if the “experienced” members were to post in this way, they could serve as example for the others.

Here is a list of Manually-Awardable Badges that aren't directly related to a competition or event: Musical Debator Melophile Holding Strong Opinions History Buff Philosopher of Music Theory Buff Orchestration Buff Keen Ear Keen Eye Musicologist Ivory Tickler Symphonist Chamber Guru Troubadour Concertophile Outstanding Orchestrator Maestro Soloist Master of Improvisation Skilled Arranger Film Buff Video Game Buff You've Got Rhythm Got Published Got Performed Tune Smith Prolific Composer Counterpoint Wizard Sick Beats Period Composer Ardent Producer Hepcat Aleatoric Composer Bando Drum Major Pop Song Aficionado Brass Aficionado Woodwind Aficionado String Aficionado Film Composer Recording Artist Musical Dramatist Challenger Avid Listener Musical Socialite Ingenious Harmonizer Dance Music Impresario Headbanging Rocker Musical Storyteller Musical Explorer VGM Enthusiast Film Music Enthusiast Sculptor of Sound Audiophile Idea Peddler Star Performer Choir Master Musical Architect Immaculate Engraver Impeccable Taste Local Lurker Abstract Composer Guitar Guru Collaborator Accomplished Virtuoso Musical Advisor Elegant Classicist Amorous Romanticist Programmatic Composer Master of Subtlety Dedicatee Lo-Fi Audiophile Better Than Thatguy Lavender Hands Effect Ardent Reviewer Feel free to suggest badges as well! - PeterthePapercomPoser

Does GarageBand Drummer count as AI? I've used this once or twice in the past. Though I made lots of manual tweaks to create alternating rhythms and percussion techniques in different sections of the piece. So it wasn't purely AI driven. I've also used AI to create cover art for my scores.

1. I'm not sure if the anti-AI detection algorithm would work 2. if we accept AI in compositions, submissions must thoroughly declare any AI involvement in composition (explain which exact ideas/parts/transformations/revisions/etc. were driven by AI and how exactly was AI used in those parts like, total LLM prompt? vague idea? proofreading? etc. generative vs analytic?) 3. I think a complete ban on AI might look close-minded (even though I am actually not against this entirely at all at this point in time), but the world is still at its infancy of dealing with AI/LLM (which itself is also at infancy), and right now at most I would be open to a severely restricted use of AI. As time goes by I might eventually be more open. 4. the problem with enforcing these rules is we might "overcorrect" and label genuine works as "AI"esque. this goes back to my first point, we don't want to mirror YouTube's AI-takeover in doing copyright claims and reports which mostly just benefits big companies.

Last year, I came across Sara Teasdale's poetry and a lot of what she wrote about on love, beauty, fear, even death resonated with me. What really peaked my interest was her poetry book of love songs where she expresses her ideas and beliefs about what it means to be loved and to love. "Child, Child" is a heartfelt ballad set in the key of D-major with many legato phrases. The opening theme is the driving point that follows the piece through the beginning, middle and end and changes meaning through each repetition. I use it to tell this story of romance sweeping us off our feet in ways that feel enchanting, and breaking us down when we least expect it. Yet, through it, we still find the courage to love and start anew time and time again. Love--be it heaven or hell--is to hold dear and to cherish. After a pretty rough time last year emotionally, Teasdale's poetry gave words to my feelings and inspired me to write. I hope this piece resonates with you as well. Please enjoy <3. Also Happy New Year!!! Text: Child, child, love while you can The voice and the eyes and the soul of a man; Never fear though it break your heart -- Out of the wound new joy will start; Only love proudly and gladly and well, Though love be heaven or love be hell. Child, child, love while you may, For life is short as a happy day; Never fear the thing you feel -- Only by love is life made real; Love, for the deadly sins are seven, Only through love will you enter heaven.

I developed a method of getting the panning and the loudness of every partial (sine wave sound component) in the song exactly where I want it in a matter of seconds. It's been yielding results that sound pretty good to my ears. My theories seem to be working. Incidentally, the principles behind this method of mixing were inspired by my faith in God.

Hi @PeterthePapercomPoser, I forgot to answer to you! I've done a couple guitar impro sessions or compositions where while playing I have intentionally detuned - by ear - to a result that made me feel right. Then afterwards I've finetuned it. It is more about a general colour (and the way guitar reacts to small, constant microtones like these) than anything else. It's completely based on ear :).

Also, Windows 95 and 98. Yes. They are staring at you right now.

They do, you can play Microsoft Minesweeper on it!

I remember one of my friends' computers doesn't...

Do they even exist at this point?

Ai should only be used for learning composition, not doing composition.

I will also add that I think video should be encouraged as the primary detection means. Almost every computer today has built-in screen capture software that will allow you to record daw/notation playback effortlessly.

Just saw this and voted. Those results so far