Systems and methods for aligning lyrics using a neural network

A first electronic device comprising:

• one or more processors;
• and
• memory storing instructions for execution by the one or more processors, the instructions including instructions for:
• receiving audio data for a media item;
• generating, from the audio data, a plurality of samples, each sample having a predefined maximum length;
• using a neural network trained to predict character probabilities, generating a plurality of probability matrices for the plurality of samples, including, for a respective sample of the plurality of samples, generating a probability matrix of characters for the respective sample of the plurality of samples, including:
• downsampling at least a portion of the respective sample to reduce a dimension of the respective sample;
• convolving an output of the downsampling of the first respective sample;
• and
• upsampling an output of the convolution of the first respective sample to increase the dimension of the respective sample;
• wherein the probability matrix includes:
• character information,
• timing information, and
• respective probabilities of respective characters at respective times;
• concatenating a set of two or more of the plurality of probability matrices to create a single probability matrix, the single probability matrix including:
• character information,
• timing information, and
• respective probabilities of respective characters at respective times;
• and
• identifying a first sequence of characters using the single probability matrix.

A non-transitory computer-readable storage medium storing instructions that, when executed by an electronic device, cause the electronic device to:

• receive audio data for a media item;
• generate, from the audio data, a plurality of samples, each sample having a predefined maximum length;
• using a neural network trained to predict character probabilities, generate a plurality of probability matrices for the plurality of samples, including, for a respective sample of the plurality of samples, generating a probability matrix of characters for the respective sample of the plurality of samples, including:
• downsampling at least a portion of the respective sample to reduce a dimension of the respective sample;
• convolving an output of the downsampling of the respective sample;
• and
• upsampling an output of the convolution of the respective sample to increase the dimension of the first respective sample;
• wherein the probability matrix includes:
• character information,
• timing information, and
• respective probabilities of respective characters at respective times;
• concatenate a set of two or more of the plurality of probability matrices to create a single probability matrix, the single probability matrix including:
• character information,
• timing information, and
• respective probabilities of respective characters at respective times;
• and
• identify a first sequence of characters using the generated single probability matrix.