Note: More recent transcription demonstrations, using the system developed together with Matti Ryynänen, are here.
The transcription examples below are produced by an automatic transcription system, which gets only the acoustic musical signal as input (with no side-information) and produces a musical notation for the sounds that make up the piece (i.e., a MIDI-like representation). Please notice, that the overall transcription system has not yet been optimized at all. It is simply:
Typical problems in the transcription are illustrated by the last few examples in the table. Such problems are still very common in the automatic transcription of real-world music.
|Original||Transcribed, then resynthesized||Comments|
|Chopin: Nocturne (piano)||chopin_nocturne1_t||See above.|
|Larrissey: Mist on the mountain (folk)||Transcription||No drums, easy instrument.|
|Zucchero: Senza una donna (pop)||Transcription||Easy pop|
|Taylor: Lady my whole world is you (pop)||Transcription||-|
|Saint-Saens: Sarabande op. 93 (classical)||Transcription||-|
|Abba: Waterloo (pop)||waterlooT9||Accompaniment pops up from the background|
|Abba: Dancing Queen (pop)||dancing_queenT9||-|
|Beethoven: 5th Symphony, part III (classical)||beethoven_5thT9||-|
|Edvard Grieg: Zur Johannesnacht (classical)||zur_johannisnachtT9||Harmonically rich|
|Claudio Roditi: Rua Dona Margarida (jazz/easy)||rua_dona_margaridaT9||-|
|Hilton Ruiz: Something Grand (jazz)||something_gT9||-|
|Kool and the Gang: Funky stuff (funk)||Transcription||Problem: difficult piece|
|The Police: Its allright for you (rock)||Transcription||Problem: difficult genre|
|Stevie Wonder: Master Blaster (pop)||master_blasterT9||Problem: short notes|
|Stevie Wonder: You Are the Sunshine of My Life (pop)||you_are_the_sunshine_of_my_lifeT9||Problem: difficult instrument to transcribe|
|Glen Miller: Over the Rainbow (big band)||over_the_rainbowT6||Problem: old recording, low quality|
Accurate and realistic evaluation of a transcription system is best achieved by transcribing synthesized MIDI songs. In this case, the exact reference score is readily available. Also, high-quality MIDI songs are complex enough to simulate the real case.
Transcription of synthesized MIDI gives somewhat optimistic transcription results. It is more difficult to transcribe music from CDs. However, MIDI songs allow much more accurate comparison between the reference and the transcription and this is more important. Pure listening of transcription results is not sufficiently informative.
Below, 20 randomly selected MIDI pieces and their transcription results are shown (I excluded 3-4 pieces due to bad transcription. Considering it now it is very pity because the selection is no more completely random. On the other hand, preparing a new and completely random set of transcriptions is too tedious, no time for such now.)
Again, the transcriber gets only the acoustic musical signal as input. Musical meter estimation was used for temporal segmentation, which improves the system accuracy compared to using one-by-one onset detection. No musical knowledge or the context was utilized. Each temporal segment was transcribed independently using a multiple-F0 estimator.
|Piano examples||Georgia||No major problems. Many missed notes, though.|
|Opus 1||Poor results. Difficulties: played in the high register, drums.|
|Blue Moon||Good transcription.|
|Israel||Poor results. Difficulty: loud snare drumming.|
|Moaning||Drums cause no problem.|
|No More||Difficulty: staccato (short notes).|
|Skylark||Relatively easy case.|
|Swing Cafe||Drums cause no problem. The rhythm gets blurred.|
|Wishstar||Rich polyphony. Some notes missed, but generally quite good job.|
|Pop music examples||Mamma Mia||Ok.|
|You Are the Sunshine...||Not very successful. High-pitched strings.|
|Get Back||Snare drumming, transcription ok.|
|I Shot the Sheriff||Quite successful.|
|Oye Como Va||Recognizable, no more.|
|Classical examples||Bach RC 13||Ok.|
|Bach RC 21||Ok.|
More? Check the DAFx2000 demonstrations page.