Written on the Wind, for piano, two versions, generated by NotaGen скачать в хорошем качестве

Written on the Wind, for piano, two versions, generated by NotaGen

Simple, playable, generated in the Satie style, but I named the original version, 2nd version, after the 1956 Sirk Western movie starring Rock Hudson. The first version is converted to the Dorian 4 mode, also known as "Romanian Minor." I did this to make it sound more exotic, and like Satie. I like the original (2nd track) more though, but maybe it is because I listened to it more, probably 60 times. In NotaGen, you can select from a variety of composer styles, but the Impressionist styles (Debussy, Ravel, Satie) rarely sound like Impressionism, because apparently that is hard to capture. Impressionism involves extended chords, chromaticism, use of the whole tone and/or modal scales. The Mozart style is different, and nearly all generations sound like bad Mozart. When using a style like Satie, the source file used in the LLM could be one of the Satie works that is not Impressionism, e.g., Satie's "Je te veux." With Debussy, I've had NotaGen source a style in a particular piece I recognize, such as Second Arabesque. I wish it would source the First Arabesque more. NotaGen is unique in that it generates in MusicXML format, which is a notation format that can be imported into Music Writing software such as the free MuseScore4. From there instruments can be applied and it can be exported as audio (WAV), PDF for the score, MIDI and other formats. I'm using the Arturia "American Home Grand" piano sound VST3 in MuseScore4, but that has to be installed separately, and is tricky to use but the best free piano I've found. The advantage with generating in notation, especially rather than audio, is that every note can be edited, the volume changed, the tempos changed, the pitch changed, articulation style changed etc. Also, for musicians, working in notation is more satisfying than the piano roll view of MIDI software, most of which can take months to learn well. MuseScore4 is not that easy for people new to notation software to learn but it is worth learning and it can do so much, and is free. That makes learning it worthwhile. I always like reading comments from people who like a particular piece, and also welcome questions about using NotaGen, or the other of the dozen AI music generators I've used in the last three years. I'm using NotaGen now because I like it the most. I delete negative comments and understand people hate AI music, think it is not music, or of the devil or whatever. Obviously, with over 300 AI pieces posted, I like AI music, and I'm doing this to show people what kind of music can be generated with a given generator, and give them a place to ask questions if they get stuck or don't understand something. My interest in providing examples, is that when I discovered OpenAI's MuseNet, the first AI music generator I used -- it is now dead -- I missed out on the first three years of its existence, just because people weren't posting their pieces. AI music will continue to advance. Some AI generators, like the composer versions of Allegro Music Transformer or Giant Music Transformer, can generate individual measures to build a song measure by measure, making them more collaborative. NotaGen generates a whole file in two minutes, but it generates so many bad ones, and adjusting the performance quality and rewriting the music can take hours. Each of my pieces takes from 3 hours to 20 hours to work on. I'm not just posting pieces straight out of the generator. I've written instructions to make using NotaGen in COLAB possible for non-Python users, and they are in the top posted comment under this video:    • One Night in Granada -- generated using No...