When trying to compose I realize that at the keyboard (and even more so on the guitar) rather than composing from my mind (or soul) I still fall into patterns I am comfortable with on the keyboard (or fretboard). as a way to break that trend (yes years of playing will help) I have hummed my compositional ideas directly into a recorder (now digital). Then transcribe by ear my humming. trouble is, it is slow ( I have a so-so ear) and I know the delay to get things transcribed effects my fleeting musical inspiration. I assumed with the modern age of sequencer and midi software ( which gave me great speed in other ways), I could use a good voice to midi tool, which worked right with all my other tools ( say in cubase) so my voice was just another controller sort of speak for getting melody down. but every year or so, I check out the newest versions of current audio to midi offerings ( digitalear, intelliscore, recognisoft, ...) and for me they how can I put this, well they suck (getting 50% of notes wrong, dumping core, ...) I know my voice is hoarse and I can't sing very well but the pitch is there(I think). Even playing with pitch and scale limitation tools on these programs, they still barely work. Has anyone found any audio to midi tool that works well with this type of goal - that is - tone deaf composer trying to hum their melodies with reasonable accuracy and right into a composing stream (sequencer). Have they improved since last year, it is just me and my gravely voice, are the hardware system better? Are there techniques that others have found with this problem.
Click on "Audio/Video Examples", from the Download list on the left. Then watch the 4 individual demo videos under Melodyne Plugin: "The Idea", "In Tune", "On Time" and "Integration". Then watch the one under Melodyne Studio, and then the BBC one. All of them are great.
Aw heck, here's the direct link to the Videos page:
Whether you want to change your vocals to actual MIDI format (in order to trigger other sounds), or else just gain all of the editing flexibility of MIDI, by making your audio MIDI-like, Melodyne looks like it'll do either one easily. I love how it breaks that pre-recorded trumpet sample instantly into individual pitches, in tempo, just as if it were a MIDI loop. Wow. (I just hope that that was a regular wav file and not some special kind of wav, designed specifically for use in Melodyne.)
But a FREE suggestion would be to try a new audio host called Reaper. It is a free alternative to Cubase and comes bundled with a few useful VST effects. One of the effects is a freeware voice tuner, like AutoTune, and it allows sending of midi notes that are recognized. This is really cool! You can put on headset, and sing or hum a tune, and it will play your instrument in realtime!
Thanks, I'll check out reaper. Meanwhile I am playing with Melodyne and it is very very strong. More than anything, its' tools convinced me that I really am singing out of tune (funny how the mind hears things differently). And melodyne makes it easy to grid to scale or simply correct my voice tuning by hand. Remember for me the original idea was to try and use new compositional tools, since when I go to the keyboard or guitar I tend work with known ( worked for me in the past) patterns ( scales, muscle memory, ...). My experiment now, to break out of that compositional style is to walk around outdoors, compose in my head then when ready, hum it into my mp3 player ( with audio in). Do as much as I can that way before I then take that directly into a channeled sequencer (cubase currently) to edit. Thereby keeping my playing style out of the process. humming as a brain to sequencer controller. I'll let you know how it works. But for those interested in audio to midi - melodyne is much much better than intelliscore or digital ear.