Music generator
AssigneeAiMi Inc.
InventorsEdward Balassanian, Patrick Hutchings
Life of this patent
Estimating the term…
Abstract
Techniques are disclosed relating to determining composition rules, based on existing music content, to automatically generate new music content. In some embodiments, a computer system accesses a set of music content and generates a set of composition rules based on analyzing combinations of multiple loops in the set of music content. In some embodiments, the system generates new music content by selecting loops from a set of loops and combining selected ones of the loops such that multiple ones of the loops overlap in time. In some embodiments, the selecting and combining loops is performed based on the set of composition rules and attributes of loops in the set of loops.
Claims20
A non-transitory computer-readable medium having instructions stored thereon that are executable by a computing device to perform operations comprising:
- generating output music content by combining multiple musical tracks such that two or more of the tracks overlap in time, wherein the generating is performed by a composition module based on statistical rules for:
- selecting tracks to add to the output music content, wherein at least one statistical rule indicates a first type of track is to be included in a specified ratio within the output music content;
- determining a number of tracks to layer in the output music content, wherein at least one statistical rule indicates a first number of tracks to layer at a specified ratio within the output music content;
- and
- determining a next section type for the output music content from among a plurality of section types, wherein different section types have different statistical rules for selecting tracks and determining a number of tracks to layer;
- and
- adjusting one or more parameters for the statistical rules used by the composition module, wherein the adjusting is based on outputs of a machine learning module that receives feedback information as input, wherein the feedback information is associated with the output music content.
A method, comprising:
- operating, by a computing system, a machine learning module to generate outputs based on output music content generated by a composition module and feedback information associated with the output music content;
- and
- generating, by the computing system based on the outputs, adjustments to one or more parameters for statistical rules used by the composition module, wherein the statistical rules include rules for:
- selecting tracks to add to the output music content, wherein at least one statistical rule indicates a first type of track is to be included in a specified ratio within the output music content;
- determining a number of tracks to layer in the output music content, wherein at least one statistical rule indicates a first number of tracks to layer at a specified ratio within the output music content;
- and
- determining a next section type for the output music content from among a plurality of section types, wherein different section types have different statistical rules for selecting tracks and determining a number of tracks to layer.
An apparatus, comprising:
- one or more processors;
- and
- one or more memories having program instructions stored thereon that are executable by the one or more processors to:
- generate output music content by combining multiple musical tracks such that two or more of the tracks overlap in time, wherein the generation is performed by a composition module based on statistical rules for:
- select tracks to add to the output music content, wherein at least one statistical rule indicates a first type of track is to be included in a specified ratio within the output music content;
- determine a number of tracks to layer in the output music content, wherein at least one statistical rule indicates a first number of tracks to layer at a specified ratio within the output music content;
- and
- determine a next section type for the output music content from among a plurality of section types, wherein different section types have different statistical rules for selecting tracks and determining a number of tracks to layer;
- and
- adjust one or more parameters for the statistical rules used by the composition module, wherein the adjustment is based on outputs of a machine learning module that receives feedback information content as input, wherein the feedback information is associated with the output music content.
Dependencies are inferred from the claim text.