I am hoping by posting this someone can help me out, so here is the problem to solve: I don’t really do notation, as far as music goes I am 90% intuitive, I’m a bass player, I play what I feel, end of story. But my ears are strong, I can find melodies where others hear noise (and vice versa) so how to involve technology in this issue? The best conduit I know is my subconscious, always listening and processing and thinking thoughts I am not consciously party to, but every so often I find myself whistling a melody. Step 1 = “is this a melody I’ve heard recently?” If no, goto step 2 = “how to capture this fleeting melody before it disappears?”. This is where I get protective i.e. do not listen to anything else resemebling a melody until this IS captured in some form. So how to capture it? Well it may turn out to be my Zoom H2s most valuable use, but I grab it, get horizontal on the couch & whistle… I don’t have perfect pitch, nor would I want it, but my relative pitch is pretty good…. And its intuitive – I don’t have to find the right note, I just whistle it. But I’d like to be able to do something with my whistled melodies…
Nothing as complex as that example, but given these constraints (a simple monophonic whistled melody, recorded as a .WAV file) how would you extract MIDI from it?
When I was playing around with that senegal doves recording last week I opened it up in Audiosculpt and had a look at the spectrum of a single melodic phrase:
And then did a chord analysis, which isolated the notes:
And using the tuning tool I could find the notes manually, but I want to do this via some form of automatic process – is there a plugin or app that does this? Monophonic audio in, MIDI out?
found a similar question on gearslutz
http://www.gearslutz.com/board/music-computers/133827-wav-midi-pro-tools.html
will try melodyne…
Pingback: Tweets that mention Music of Sound » Recycling whistled melodies -- Topsy.com
I feel you on this one.
My natural sense and feel for music far outweighs my working knowledge of music theory and ability to play a instrument.
It often seems like I can’t properly express all the ideas in my head, even using all the tool available today.
I guess there’s no substitute for hard work, but perhaps these can help.
http://www.music-notation.info/en/compmus/audio2midi.html
Will try this RTAS plugin -a spectrum analyser with MIDI output!
http://www.bluecataudio.com/Products/Product_FreqAnalystPro/
tim, melodyne is AMAZING.. if you want me to analyse it for you just send me the file.
though i could work it out by ear.
L
oh i just read the whole thing.. melodyne is your friend
which version of melodyne would i need? can the plugin generate MIDI?
hi tim, here is an example done on melodyne studio 3 on osx if you want an idea of what it does to a whistle recording without any tweaking
http://cl.ly/ad28f69471a2335bace3
cheers,
Emile
that is brilliant Emile, exactly what I am aiming to achieve – thanks for doing that!
I’d agree, Melodyne is probably the best for this.
You could also use SPEAR to get rid of the harmonics, reduce your whistle to a pure sine wave, and then you have a very clean signal to process in any audio-MIDI converter.
http://www.klingbeil.com/spear/
Although, learning notes isn’t THAT hard, there’s only 12 of ‘em.
cheers
Dan
As an aside, I remember my very first “DAW” ( Opcode Studio Vision, running on a Mac II nothing, with a Sound Tools card… 4 channels!!! ).
It had a feature where you could turn any monophonic audio track into MIDI, you could even map loudness AND brightness to any desired MIDI CC. And this was in 1990 or so. You have to wonder why these things didn’t become standard later on.
hah me too! well i used Sound Tools first, then ProTools but Studio vision was my first music software and I remember that feature well!
true – i remember it (and Q Sheet AV) well!
Turbosynth was so far ahead of its time – if you released it now it would still be very relevant/useful!
I know that problem Tim.=) I tend to record it and insert Logic Pitch Correction, where I can see the notes input. It ain`t amazingly precise, but it does the job. I`m sure Pro Tools does have smthg simular.
Well, I can see two possibilities here:
The first is OpenMusic, in wich (in a similar way to max/msp) you can load the file from you analysis in Audiosculpt and convert do MIDI, by programming or using one patch from the library (if you have the license to use it from IRCAM).
The other one is by using Max/MSP, by means of programing with the extension fiddle~ (http://crca.ucsd.edu/~tapel/software.html).
With fiddle~ is simpler to achieve, but the first one offers more possibilities (including exporting to finale).
Hope I offered a different approach on the matter…
Whistle it by all means, but basically the keyboard is the easiest instrument to use that can transcribe your immediate instinctual reaction (whistling) to a Midi or pure note form to another medium.
But you are totally right in capturing the immediate which is where the purest ideas come from……
If you can play the same melodies on a keyboard as you have whistled then you are 99% there.
Melodyne Editor Stand Alone mode is the tool for this! Oh – just saw it was mentioned above. I’m pretty sure only the standalone mode supports midi export (not the plug in version)…. It really enjoy putting any audio (full mixes of songs, soundscapes, etc..) into it to see what it’s polyphonic algorithm comes up with – very cool sounds and sound design potential in the ‘blob’s that it splits into… I tend to record the best results out using audio hijack… its retuning potential for polyphonic material is great as well! I’ve been very happy with the purchase!
How is the Blue Cat Audio plugin mentioned above?
The Blue Cat doesn’t exactly give very musical results in my experience, but the lot of them do generate very useful “control voltages”. Especially the peakmeter is great, as it outputs peak/rms which can be used for lots of things.
Another thing that can turn audio into MIDI is Kyma.
Finale has pitch to notation, and can export MIDI as well. I’ve never used the feature myself, but they claim to be able to take any mono input (voice, instrument, and I would guess, whistling) and notate pitch and rhythm. I’ll have to fire it up and try it out to tell you how well it works. Of course, I can’t whistle, and my singing leaves a lot to be desired, as well!