Exhilarating the Tone Transfer feature by Google

0
977

Most of the time we used to hum our favorite tunes that keep running inside our mind. Everyone cannot replicate the exact essence of the tune but we tried to match up with our abilities. What happens when you tune changes into a mesmerizing one? You heard it right! Google comes up with a new idea of changing your humming into an exquisite violin solo.

Google conducted an experiment on “Tone Transfer” with the help of its new machine learning algorithm. This helps to modify the normal humming into a pleasant musical piece. The steps are really simple. First, visit the Tone Transfer site via Android mobile or Computer of OS Windows or Mac. On the website look for the “Add your own” option and select it. Now its time to record. You can play an instrument or use your own voice as an input. Once input is fed Google will utilize its machine learning algorithm and convert it into a digital signal. Now you can convert this digital signal further into the desired tune of an instrument such as Flute, Violin, Trumpet, or Saxophone.

The output quality of the tune may vary depending upon the background noise and microphone. One should be a little bit cautious while recording the audio because it will have a direct impact on the quality of the output.

One can explore the benefits of machine learning in the field of art as well. Google’s Magenta AI team has built an open-source technology called the “Differentiable Digital Signal Processing (DDSP) library. With the help of Digital Signal Processors, one can be able to generate the natural sounds of instruments. The main operation of Differentiable Digital Signal Processing is that it takes the audio input from the user and converts it into a digital signal with the help of a neural network. The converted signal is then ready for further conversion into distinct instruments of the user’s choice.

Tone transfer has the potential to create high-quality audio from the user input. Google with the help of the DDSP library, takes care of audio synthesis models. These models in turn aid Tone transfer tools to generate high-quality audio.