Text-to-Audio Phrase Generation With Rhythmic Pitch Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio generation methods fail to produce musical phrases with rhythmic changes, resulting in flat audio output when converting text data to music.
Innovation Solution
An audio generation device and system that utilize a storage configuration with (n+2) sound information storage slots, setting pitch attribute codes in specific slots to generate audio data with rhythmic variations, allowing for the creation of audio phrases with rhythmic changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If two note codes are extracted for one input Japanese character, then audio data can be generated from text data, but the music phrase becomes flat without rhythmic changes
Solution Approach 1:
The patent divides the audio generation process into multiple independent pitch attribute code extraction steps (first, second, and third pitch attribute codes) corresponding to different rhythmic positions. Each pitch attribute code is extracted independently and assigned to specific sound information storage slots, allowing rhythmic variation while maintaining ease of generation from text data.
Solution Approach 2:
The patent introduces multiple pitch attribute code dimensions (first, second, third pitch attribute codes) to represent different rhythmic positions within a beat. This dimensional expansion transforms the flat two-note-code system into a multi-dimensional system that can represent complex rhythmic patterns while still processing one input character at a time.
2Adaptability or versatility
If multiple pitch attribute codes are set in sound information storage slots, then rhythmic changes can be achieved, but the data structure becomes more complex
Solution Approach 1:
The patent assigns different pitch attribute codes to specific local positions (sound information storage slots) based on rhythmic requirements. Each slot receives appropriate pitch attribute codes (first, second, or third) according to its position and rhythmic function, creating local specialization that enables rhythmic variation without requiring complex global restructuring.
Solution Approach 2:
The patent pre-structures the packet data with designated sound information storage slots and pitch attribute code positions before actual audio generation. This preliminary organization of data structure with predefined slots and code positions simplifies the generation process by eliminating the need for complex real-time structuring during audio synthesis.
3Adaptability or versatility
If pitch attribute codes are set in specific sound information storage slots according to rhythmic patterns, then audio phrases with rhythmic changes can be generated, but the processing procedure becomes more complex
Solution Approach 1:
The patent implements periodic assignment of pitch attribute codes to sound information storage slots based on rhythmic patterns. The first, second, and third pitch attribute codes are assigned in a periodic manner corresponding to different positions within the beat structure, creating a systematic and repeatable processing pattern that reduces procedural complexity.
Solution Approach 2:
The patent changes the parameter of pitch attribute code assignment based on the position of sound information storage slots. Different parameters (first, second, or third pitch attribute code) are applied to different slots according to their position and rhythmic function, creating a systematic parameter-based assignment procedure that simplifies processing compared to position-specific custom rules.
Data Source
Figure 1(a)~1(c)
Figure 2(a)~2(d)
Figure 3(a)~3(d)
AI summary
This audio generation device: generates an audio phrase with rhythmic changes with respect to input text data, symbol data, or a combination thereof; provides various content data items by associating the same with the audio phrase; and generates, upon input of content data, audio data corresponding to that content data. The audio generation device is characterized by having the following features. Packet data is generated by associating contents data with at least one among a first pitch attribute code corresponding to a pitch name code constituting one sound, a second pitch attribute code corresponding to a pitch name code constituting a plurality of sounds, and a third pitch attribute code. A process is performed for generating the audio data composed of the pitch name code corresponding to the packet data is performed.