Generating dynamically controllable composite data structures from a plurality of data segments

• US 10,452,996 B2
• Filed: 09/28/2015
• Issued: 10/22/2019
• Est. Priority Date: 08/10/2011
• Status: Active Grant

• Current Assignee: Konlanbi
• Original Assignee: Konlanbi
• Inventors: Drame, Cyril
• Primary Examiner(s): Hill, Stanley K.
• Application Number:
  US14/868,149
• Publication Number:
  US 20160155065A1
• Time in Patent Office: 1,485 Days
• Field of Search: None
• CPC Class Codes: G06F 3/165 Management of the audio stream, e.g. setting of volume, audio stream path
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

A method of producing dynamic controllable data composites from two or more data segments includes: building or training one or more function mappers to map between one or more extracted feature envelopes sets from the original data and one or more general parametric representations of the data; combining the extracted feature envelopes or the function mappers using two or more audio segments; and feeding the extracted feature envelopes or combined feature envelopes to the function mappers to obtain synthesis parameters to drive a synthesis process.

Claims

1. A method of producing dynamically controllable data composites from two or more data segments of captured data, the method comprising:

training one or more function mappers, wherein each function mapper has interpolation and extrapolation capabilities and includes parameters used to fit to training data in a training process to map between one or more feature envelope sets extracted from the data and one or more parametric representations of the data and by using at least one training parameter to discriminate between the data segments in a training process for the one or more function mappers, the one or more feature envelope sets corresponding to one or more characteristics that are estimated from the captured data, and the one or more parametric representations of the captured data corresponding to a parametric analysis process that fits the one or more parametric representations to the captured data, each parametric representation being a set of one or more parameters estimated from the captured data in the parametric analysis process;

forming a composite data structure by using at least one control parameter to combine values for a plurality of the data segments, each of the plurality of data segments being associated with a corresponding feature envelope set and a corresponding function mapper; and

generating synthesis parameters for the composite data structure by feeding the one or more feature envelopes sets corresponding to the data segments of the composite data structure to the corresponding one or more function mappers and combining outputs of the one or more function mappers in accordance with the composite data structure, the synthesis parameters being used to drive a synthesis process.

2. The method of claim 1, wherein the captured data corresponds to perceptual data and the one or more parametric representations of the captured data include spectral characteristics of the captured data.

3. The method of claim 1, wherein training the one or more function mappers includes performing the parametric analysis process that fits the one or more parametric representations to the captured data.

4. The method of claim 1, wherein an output of the synthesis process includes output characteristics that are based on the one or more perceptual characteristics of the captured data.

5. The method of claim 1, wherein the captured data is received as an output of a measurement device.

6. The method of claim 1, wherein generating the synthesis parameters includes combining two or more feature envelope sets at an input level of a first function mapper.

7. The method of claim 1, wherein training the one or more function mappers includes combining two or more feature envelope sets at a body level of a first function mapper by training the first function mapper with a first training parameter that operates as a discriminant input on the two or more feature envelope sets.

8. The method of claim 1, wherein generating the synthesis parameters includes combining values at output levels of a first function mapper and a second function mapper, each of the first and second function mappers being fed with two or more feature envelope sets.

9. The method of claim 1, wherein the captured data includes audio data and the one or more feature envelope sets that are extracted from the data segments of the captured data include at least one of a pitch envelope, a loudness envelope, or a brightness envelope for the audio data.

10. The method of claim 1, wherein the captured data includes audio data and the one or more parametric representations of the captured data include values for spectral peaks in one or more of the data segments.

11. The method of claim 1, wherein the captured data includes body-motion data and the one or more feature envelope sets that are extracted from the data segments of the captured data are based on at least one of position, rotation angle, or velocity for one or more body joints or one or more end effectors.

12. The method of claim 1, wherein the captured data includes body-motion data and the one or more parametric representations of the captured data include values for at least one of position, rotation angle, or velocity for one or more body joints or one or more end effectors.

13. The method of claim 1, wherein the captured data includes texture data, and the data segments include two-dimensional arrays corresponding to portions of the texture data.

14. The method of claim 1, wherein the captured data includes texture data and the one or more feature envelope sets that are extracted from the data segments of the captured data are based on neighboring image-texture values for a given location.

15. The method of claim 1, wherein the captured data includes texture data and the one or more parametric representations of the captured data include values for image texture for a given location.

16. The method of claim 1, wherein generating the synthesis parameters includes:

using a first control parameter to interpolate values between output levels of a first function mapper and a second function mapper, or using the first control parameter to extrapolate values from the output levels of the first function mapper and the second function mapper.

17. The method of claim 1, wherein generating the synthesis parameters includes using a first control parameter to compute a linear combination of values from the output levels of a first function mapper and a second function mapper.

18. The method of claim 1, wherein the one or more function mappers include at least one neural network that is trained adaptively in the training process for the one or more function mappers.

19. The method of claim 1, wherein generating the synthesis parameters includes using a discriminant input to combine two or more of the data segments in a first function mapper.

20. The method of claim 1, further comprising:

determining boundaries of the data segments in the captured data from changes in the one or more feature envelope sets that are extracted from the data segments of the captured data;

associating one or more categories with the data segments, each category corresponding to a perceptual descriptor for a corresponding data segment; and

training the one or more function mappers by using the one or more categories in association with e at least one training parameter to discriminate between the data segments.

21. The method of claim 1, further comprising:

providing values for a first segment of the captured data to a user interface;

receiving from the user interface values for a first perceptual descriptor for the first segment;

associating the first perceptual descriptor with the first data segment; and

training a first function mapper by using the first perceptual descriptor in association with a first training parameter to discriminate between the first data segment and other data segments.

22. An apparatus for producing dynamically controllable data composites from two or more data segments of captured data, the apparatus including at least one computer that performs operations for computer-implemented modules including:

a training module for training one or more function flappers, wherein each function mapper has interpolation and extrapolation capabilities and includes parameters used to fit to training data in a training process to map between one or more feature envelope sets extracted from the data and one or more parametric representations of the data and by using at least one training parameter to discriminate between the data segments in a training process for the one or more function mappers, the one or more feature envelope sets corresponding to one or more characteristics that are estimated from the captured data, and the one or more parametric representations of the captured data corresponding to a parametric analysis process that fits the one or more parametric representations to the captured data, each parametric representation being a set of one or more parameters estimated from the captured data in the analysis for synthesis process;

a data-structure module for forming a composite data structure by using at least one control parameter to combine values for a plurality of the data segments, each of the plurality of data segments being associated with a corresponding feature envelope set and a corresponding function mapper; and

a synthesis module for generating synthesis parameters for the composite data structure by feeding the one or more feature envelopes sets corresponding the data segments of to the composite data structure to the corresponding one or more function mappers and combining outputs of the one or more function mappers in accordance with the composite data structure, the synthesis parameters being used to drive a synthesis process.

23. The apparatus of claim 22, wherein generating the synthesis parameters includes using a discriminant input to combine two or more of the data segments at an input level of a first function mapper.

24. The apparatus of claim 22, wherein the training module is further configured to perform operations including:

determining boundaries of the data segments in the captured data from changes in the one or more perceptual characteristics of the captured data;

associating one or more categories with the data segments, each category corresponding to a perceptual descriptor for a corresponding data segment; and

training the one or more function mappers by using the one or more categories in association with the at least one training parameter to discriminate between the data segments.

25. The apparatus of claim 22, wherein the training module is further configured to perform operations including:

providing values for a first segment of the captured data to a user interface;

receiving from the user interface values for a first perceptual descriptor for the first segment;

associating the first perceptual descriptor with the first data segment; and

training a first function mapper by using the first perceptual descriptor in association with a first training parameter to discriminate between the first data segment and other data segments.

26. A non-transitory computer-readable medium that stores a computer program for producing dynamically controllable data composites from two or more data segments of captured data, the computer program including instructions that, when executed by a computer, cause the computer to perform operations including:

training one or more function mappers, wherein each function mapper has interpolation and extrapolation capabilities and includes parameters used to fit to training data in a training process to map between one or more feature envelope sets extracted from the data and one or more parametric representations of the data and by using at least one training parameter to discriminate between the data segments in a training process for the one or more function mappers, the one or more feature envelope sets corresponding to one or more characteristics that are estimated from the captured data, and the one or more parametric representations of the captured data corresponding to a parametric analysis process that fits the one or more parametric representations to the captured data, each parametric representation being a set of one or more parameters estimated from the captured data in the parametric analysis process;

forming a composite data structure by using at least one control parameter to combine values for a plurality of the data segments, each of the plurality of data segments being associated with a corresponding feature envelope set and a corresponding function mapper; and

generating synthesis parameters for the composite data structure by feeding the one or more feature envelopes sets corresponding to the data segments of the composite data structure to the corresponding one or more function mappers and combining outputs of the one or more function mappers in accordance with the composite data structure, the synthesis parameters being used to drive a synthesis process.

27. The non-transitory computer-readable medium of claim 26, wherein generating the synthesis parameters includes combining two or more feature envelope sets at an input level of a first function mapper.

28. The non-transitory computer-readable medium of claim 26, wherein training the one or more function mappers includes combining two or more feature envelope sets at a body level of a first function mapper by training the first function mapper with a first training parameter that operates as a discriminant input on the two or more feature envelope sets.

29. The non-transitory computer-readable medium of claim 26, wherein generating the synthesis parameters includes combining values at output levels of a first function mapper and a second function mapper, each of the first and second function mappers being fed with two or more feature envelope sets.