TS-AudioToMIDI is a program developed for automated music recognition. TS-AudioToMIDI is able to both recognize pre-recorded audio files and perform on-fly recognition. This gives you numerous possibilities:
- upload your favorite melody to your web page
- play it on your cellular phone
- use it in your MIDI instrument for training or for background
- use it in a MIDI editor (some of them can print MIDI files as musical notation)
- make a preview of your mp3 collection (saves traffic on your web site)
You can also use TS-AudioToMIDI for numerous other purposes.
Recognizing pre-recorded music provides the possibility to convert digital sound to MIDI format. MIDI file takes hundreds times less space than digital sound in any format that allows you to use it as a melody for your web page or make a preview of your mp3 Internet music collection. As MIDI file is, in fact, a layout of music, you can change arrangement for it in contrast to digital sound. Instruments, number of voices or melody itself can be simply modified using any third-party score editor, which allows creating remixes or variations. Also MIDI file can be uploaded to most cellular phones as ring tone (please consult your cellular phone manual for details).
On-fly recognition is a process performed in real-time, without noticeable delay. TS-AudioToMIDI "listens" to the sound incoming from Wave In device (it may be, say, microphone or line input with some musical instrument plugged in) and detects notes it "hears". Detected notes can be either recorded or played in real time, or both. Thus, you can change instrument in real-time, for example play the guitar and get a piano sound. As MIDI files take little space you can store hundred hours of melodies without the necessity to clean up your disk. Record your improvisations, radio or anything.
After listening to the MIDI files TS-AudioToMIDI produces, people occasionally notice some inaccuracies in melody and arrangement. The problem of automated music recognition cannot be solved entirely nowadays and mistakes in melody are possible. Apt tuning of recognition parameters may significantly reduce the amount of those mistakes. But one should remember: the more complex melody one is recognizing the more inaccuracies one gets. For simple record of musical instrument the number of mistakes is usually lower than for ensemble record. The highest accuracy rate is reached when recognizing monophonic melody. After performing recognition with optimal parameters and saving the result as MIDI file you usually need to perform final correction for it in a third-party score editor.
TS-AudioToMIDI is unable to detect the instrument itself. You need to assign the instrument manually. Also drums are not recognized due to peculiarities of MIDI format.
Related topics:
How to transcribe a WAVE file
How to transcribe music in real-time
|
