Jump to content

Leaderboard

Popular Content

Showing content with the highest reputation on 12/08/2025 in all areas

  1. Hi my dear readers, it's me again. I don't understand many things here, but first things first. Imagine that I say this piece is evidently made by AI. It would be very easy to prove me wrong, @Vonias. Instead of submitting .mid files, a better, more detailed description of the process you followed would have clarified all posters here of what you did and what you have not done with it. Personally I don't care a single bit about that but we'll get to that point later. Please, consider that you are in a place filled by trained & experienced musicians and some of us have passed hundreds of hours listening to pieces just to give feedback out of either a brief or a more in-depth analysis. For some, not specifying that you were helped by an AI to make a better recording of your work and ambiguously describe it or talk about it with phrases like: may very well lead to confusion and even anger among others who consciously decided against using this tools in order to participate in an event in which ultimately your compositional ability is what we want people to show off, review and comment. Various AI tools exist to "help" with that recognition, but they are still in beta (mostly) and their own creators advise against using them as reliable methods. I took the liberty to pass the audio to some, despite you confirmed that the recording is AI: It basically says "Pure AI" 90% and provides a model. Tell me —if you want— if it hit it, this will make this developer's work better. But the thing is that you say everything else is not AI-generated. And this is where we get into trouble here. There are two different pieces here. A pdf submitted along the creation of the post that has little to do with the final mp3 you provided and the score that apparently resembles the mp3, according to you, 1:1? What's exactly the relation between the first and second scores? I can see some, let's say "adventurous" chords in the first score but not that many in the second. First score is half the size, begins with a different voicing. Rhythms don't really coincide, lyrics rest at different places "angel____ vs. an-gel cries___" . Midis are all over the place but you can clearly see the difference. Just pointed out some. What @AngelCityOutlaw is pointing out is mostly in the right direction but: That would not work. You need a seed too. If you do that yourself twice with the same material (and without fixing the seed) you'll get different things if you don't specify the same seed to the model, whatever it be. It happens with music, but also with LLMs and Image & video generators. And there are various tools to treat, import and export midis. Generating high-quality products using it is a whole 'nother story yes, but we ain't talking about one here, this is a <1 minute composition. Finally let me tell you that I don't mind if you just made the whole thing via AI, a part, or whatever. I would prefer you and anybody else in that situation to be honest and clarify things from the beginning to the rest of us who most likely are prone to and will have a listen to your piece, perhaps even giving some also honest feedback in exchange. Regarding that, the audio doesn't seem quite fitting. The strange autotune sound that the AI apparently gives to the singers sometimes and the strange cuts in vibrato feel off. The piece lacks development but it's a short "poem" more or less, right? For more info on this topic of using AI to make/export/import music on many formats, an AI that I have enslaved using a rectangular plastic card that spits money, may show you the way. Or maybe not, who knows: https://www.perplexity.ai/search/can-suno-or-similar-programs-e-2dN_8OipSNCC93.FtFGJoA?sm=d#0 Perhaps I'm late to the party, but these are my two cents. Sorry for my possible typos in advance... I am too sleepy. Regards!
    1 point
  2. Also something else I need to say about this: So I, and others, can put in hours of work writing the music, orchestration, making detailed mockups, etc. but a guy who writes a Suno prompt and has a "piece" in 5 minutes is treated as equally-valid in this event? It isn't actually even his music. Where is the "fun" in that, exactly? Why even bother? You guys say you want more people to do reviews. What your reviews are going to become if this kind of thing is permitted is a bunch of people being like "Nice prompt bro, but I would've said 'romantic soaring strings' instead of 'cinematic'".
    1 point
×
×
  • Create New...