Picking up from the microphone to achieve real-time escape of midi format?
13 views
Skip to first unread message
zhang chao
Oct 10, 2023, 5:14:16 AM10/10/23
to Magenta Discuss
Can you ask a question?
Now there is a requirement to identify the piano playing sounds in real-time.
Currently, a method has been found to recognize and transcribe pop-up piano recordings into MIDI files using the Onsets and Frames model in Magenta Music developed by Google AI Labs and a recurrent neural network.
Although this solution can be implemented, there is a troublesome process that requires first indexing, saving as an MP3 file, and then uploading to the Google server to escape midi. This process takes a very long time, and in addition, it still needs to go through VPN in China. This will take too long to wait in practical applications, so it cannot be considered as a implementation. Do you have any other way to escape midi directly from the microphone?
We also welcome other great deities to reply
Now there is a requirement to identify the piano playing sounds in real-time.
Currently, a method has been found to recognize and transcribe pop-up piano recordings into MIDI files using the Onsets and Frames model in Magenta Music developed by Google AI Labs and a recurrent neural network.
Although this solution can be implemented, there is a troublesome process that requires first indexing, saving as an MP3 file, and then uploading to the Google server to escape midi. This process takes a very long time, and in addition, it still needs to go through VPN in China. This will take too long to wait in practical applications, so it cannot be considered as a implementation. Do you have any other way to escape midi directly from the microphone?
We also welcome other great deities to reply
Reply all
Reply to author
Forward
0 new messages