Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
(Rolling in flour, halfbaking my ass off)

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.



Multi Instrument to Midi

Better than Pitch to Midi
  (+2, -1)
(+2, -1)
  [vote for,

When searching google for software that could convert sound data (.wav files) to midi, the best that I could find was a "pitch to midi" program.

That program didn't try to figure out what kind of instrument was being played, but merely noted what the pitches of the notes were (and how long each was played).

I would like to see a program which can be given a set of instrument samples and a sound file, which would then analyse that sound file, identifying which sounds were made by which instruments (plus of course pitch, duration, and volume), and generate a midi file.

Ideally, one could take a recording of a symphony orchestra, and convert it to midi, in such a way that the midi file when played back sounds like the original recording.

This idea is of course very half-baked, since I haven't a clue as to how it would actually work.

goldbb, Mar 30 2009

Melodyne http://www.celemony.com/cms/
[tatterdemalion, Mar 30 2009]

TallStick AudioToMIDI http://audioto.com/
Supports polyphony and waveform analysis (in other words it can tell which instruments play which notes and make multiple tracks). The results aren't that great in my experience though. [Spacecoyote, Mar 31 2009]

lev and thumpbot http://vimeo.com/313131
[Spacecoyote, Mar 31 2009]

Polyphonic pitch-to-MIDI http://www.google.c...onic+pitch-to-MIDI+
Survey [csea, Mar 31 2009]


       be easier to use a stereo recording, figure out which sections/solos corresponded to which instrument, compare to a GM MIDI set, then rebuild.   

       it would sound like crap.   

       For 1/10th the effort you could put the written score into a MIDI file.   

       and it would still sound like crap.   

       anyways your climbing up the arsehole of the MIDI standard with this: MIDI (including Pitch-to-MIDI) was devised to layer sounds while playing one instrument.
FlyingToaster, Mar 30 2009

       I am of course utterly ignorant of such things, but doesn't sounding like crap or not depend to some extent on the instruments connected to the interface?
nineteenthly, Mar 30 2009

       Melodyne will do polyphonic audio-to-MIDI conversion (link). It works best when each instrument is analyzed separately, so you would probably want to run it on separate tracks. I don't know if it tries to identify the instrument type in the MIDI file but this is a simple thing to adjust yourself.
tatterdemalion, Mar 30 2009

       [19thly].. You couldn't do the individual instruments, they would have to be entire sectionals. The GM (General Midi) soundset contains a number of "orchestral" sounds but not near enough to allow for expression.
FlyingToaster, Mar 30 2009

       OK, thanks. I was thinking of MIDI in terms of the thing you use to control musical instruments, not sound card stuff.
nineteenthly, Mar 30 2009

       Maybe goldbb was [thinking of midi in terms of the the thing you use to control musical instruments], too.
Spacecoyote, Mar 30 2009

       oh... planning on hooking it up to a trumpet ? or maybe buying a $3K synth to avoid purchasing a $4 CD, and if you're doing something with > 16 parts you need one with 2 MIDI Ins or more than one synth.
FlyingToaster, Mar 31 2009

       I might be mistaken, myself not an expert and all, but isn't there ~$200 software out there that's pretty much a match for a ridiculous $3k synth? This sounds more like a tool for experimenting with making music rather than //to avoid purchasing a $4 CD//.
Spacecoyote, Mar 31 2009

       <acerbic comment deleted>   

       But anyways now you want as a precursor to playing this, a $200 softsynth (good luck with that though I'm admittedly clueless as to their capabilities and qualities of their soundbanks) as well as a computer capable of playing x amount of parts without glitching.   

       I'll grant a usefulness factor for making a MIDI recording of a garage band for practice or something.
FlyingToaster, Mar 31 2009

       Given the amount out effort people put in to making MIDI sound like an orchestra, I'm amazed you want to do the reverse. The question is not how, but why?
wagster, Mar 31 2009

       Yeah but there's little point in bickering about it I guess. This sort of thing is already baked albeit far from perfect (I use the word "perfect" lightly as we're talking about MIDI here).
Spacecoyote, Mar 31 2009

       There are actually quite cheap keyboards which do MIDI. We've got one which originally cost a hundred and fifty quid.
nineteenthly, Mar 31 2009

       Yeah I've got a cheap Casio. Don't use it much though as I'm not any good.
Spacecoyote, Mar 31 2009

       You know, [bigsleep], in a sense "professional" can be what you want it to be. To some extent, all it needs is enough people devoting enough time to it in a serious and dedicated manner combined with publicity and widespread respect. I'm thinking here of the likes of "glitch" music, which presumably started off with amateurs but built up into a genre, like a lot of other things.
nineteenthly, Mar 31 2009

       I see what you mean. It's just that in the current economic situation with the likes of the internet and such, success is a rather shaky thing, and possibly professionalism likewise. If you offer a professional service, your major competitor may be your customer in that they may decide they can do it for themselves as well as you if they have more time and less money than they used to have. The internet is relevant here because it's a source of possibly inaccurate information, but information which may make someone at least feel that they know how to do something. Obviously, i'm talking about herbalism here but it could apply to a whole lot of other things. Right now, the market which for such things is shrinking, which may influence professional services. Someone might want to sell their house without using an estate agent, for example. I wonder if the same applies to this.
nineteenthly, Mar 31 2009

       //Can we please just get on with something useful. ?//
you do know where you are, right ?
FlyingToaster, Mar 31 2009

       There is a reason for my username!
I dunno, but i doubt it was Alonzo Church or John McCarthy.
It'd be a lot easier if there was agreement on what was worthwhile and in what sense, since it's amazing both what people are prepared to pay money for and what they aren't, and to my financially learning-disabled brain that distribution seems random. If it was clear what people would and wouldn't buy, i could at least attempt to put myself in a position where i could sell what they would pay for.
nineteenthly, Mar 31 2009

       Pitch-to-MIDI would be good for me personally if I wanted to do jazz because I haven't developed a sense for anything greater than a 9th... of course on the other hand if I was doing jazz I (probably) would.   

       The main thing is... if a musician is competent enough to perform in an orchestra then he/she can read music, therefore the idea is a non-sequitur. The exceptions are groups like high-school marching bands and amateur choirs, but they always have a conductor capable of transcribing a written score to MIDI quite easily... the alternative is to handpick some soloists to record a piece, which can be pretty time-consuming unless you're already paying for section leads (pro singers who try to keep their section in line).
FlyingToaster, Mar 31 2009

       Further to Flying Toaster's point, and to the discussion about use by professionals, I cannot think of very many circumstances where professional musicians or producers would have need of the product described in this idea. As I pointed out, the Melodyne plug-in can do this on a per-instrument/per-track basis, but even then it is a secondary function of the program and not its main purpose. Transcribing an entire piece into MIDI on one play-through doesn't seem like a need that is going to arise often.
tatterdemalion, Mar 31 2009

       So noone here thinks it would be cool to be able to take a music from a CD, convert it to midi data format, then use that midi data to direct real instruments to play the same piece of music?   

       I'm assuming that the midi data that controls sound cards is the same type of midi data which is used to control musical instruments. Am I mistaken here?
goldbb, Mar 31 2009

       Yup, its the same. In fact, the program I linked would work for the job. I'd like to see lev and thumpbot [link] play the music.
Spacecoyote, Mar 31 2009

       //Am I mistaken here?// nope, I just don't see the point except as a (granted nifty) exercise in computer programming using sound elements.   

       The only viable utility I see is a garage-band doing originals can "record" a practice then rehearse at home "minus 1" style... but that just puts it on par with making an audio recording (perhaps using some of the same routines to separate the instruments without bothering to make discrete source tracks... okay that'd be neat but still sorta pointless).
FlyingToaster, Mar 31 2009

       Polyphonic pitch-to-MIDI would be useful as a front end to a MIDI-to-sheetmusic system (Finale, Sibelius, etc.)   

       Integration of the two capabilities would allow auto-transcription as a basis for producing custom arrangements.   

       Tallstick looks interesting, but the general problem is very difficult, and I doubt it works well for all types of music. At best, the current products might save some time, but most material will require post-editing. See [link] for some additional references.
csea, Mar 31 2009


back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle