| 摘要 |
PURPOSE:To select music according to the sensitivity that the music has by inputting and converting the extent of the sensitivity that the music has into a music number, reproducing the performance music having the music number, and mixing a speech inputted from a ' microphone with the performance music and outputting them. CONSTITUTION:Respective neurons 34 of the input layer 31 of a neural network 30 are connected to all input-side neurons 36 of the intermediate layer 32 by synapses 35, the input-side neurons 36 of the intermediate layer 32 are connected to all output-side neurons 38 by synapses 35 and connected to output-side neurons 38, and the output-side neurons 38 are connected to all neurons 39 of the output layer 33 by synapses 35. Here, the strength of coupling with each neuron is set, synapsis by synapsis, so as to select the music number of KARAOKE music. Namely, the neuron weight of respective synapes connected to the neurons 36 of the intermediate layer 32 is set in order from the top and the weight value is optionally set to vary the amount of information sent to the next layer, thereby varying an output pattern. |
