What is automatic music transcription?

Simply put, automatic music transcription is a mathematical analysis of audio recording (usually in WAV or MP3 format) and converting it to a music indicator (usually in MIDI format). This is a very difficult problem for artificial intelligence. For comparison, the problem of text recognition scanning (OCR - optical character recognition) has been solved with 95% accuracy: it is, on average, valid for a given class of programs. Voice recognition programs already work with 80% accuracy, while music transcription systems work with 70% accuracy, but only for one chord (one note at a time). For polyphonic music, accuracy is even lower.

In order to create a MIDI setting for chords recorded in audio formats (WAV, MP3, etc.), the musician must determine the strength, speed and duration of each play note and record these parameters in a series of MIDI events. Have to do. General Chat Chat Lounge The software should do the same for the copy of the music. Even for a single device combination. Also, this is not an easy task, as the audio recording consists of signal signal waves and does not include specific music figures.

In general, a variety of music teams, harmonious constructions and transitions make it impossible to construct a mathematical algorithm for accurate reconstruction of musical scores from audio sources. Audio data is hard to copy, with lots of instruments, drums and bumps or truncation signals, unstable tone and background noise. In many cases, however, Ekov Music Composer will produce a MIDI note that represents the chord line and the main chords of the analytical music.


Difference between audio and MIDI formats
The difference between audio (WAV, MP3, OGG, etc.) and MIDI formats is the representation of audio and music. The audio format is a digital recording or sample of any sound (including speech) and the MIDI format is essentially a series of notes or MIDI events. The ratios are almost identical between the sound disk and the printed text.

Audio formats
An audio file (WAV, MP3, OGG, etc.) is recording a sound wave. This is a mix of all the sounds (instruments, sounds, background sounds) you hear while recording. So, for example, you can record human voice in MP3 format, but you cannot edit any notes and change any instrument in the music recorded in the audio file. Windows Standard WAVE PCM format only contains pulse modulation data. PCM format is the only type that completely protects the whole wave without loss of data.

There are many other audio recording formats. They are different from one another in the compression algorithm and can reference a group. It's very easy to switch from one shape to another. There are many sound editors that allow you to do this.

MIDI format.
The MIDI (Digital Musical Instrument Digital Digital Interface) is a series of commands for controlling one or more pieces of form, such as music hardware or software. These commands are not sound, they are a directive to do something (primarily to create sound). For example: Select Instruments n. 1 (sound grand piano), play note number. no 60 (C5) with speed number. 7 127. Therefore, you cannot, for example, present human speech in MIDI format, but you can edit any note or change any instrument in the recorded music in the MIDI file.

MIDI in audio conversion.
Recorded music in MIDI format can easily be converted to audio format. You can play MIDI files on a compatible player and record the music played in the sound editor. The audio file size will be larger than the music file rendered in MIDI format. Music quality will be determined by your sound card's MIDI capabilities and the professionalism of the original MIDI file maker. There are programs that convert MIDI files to audio recordings only using your MIDI device tabs (WAVE table synthetic).