System and methods for training a trainee to classify fundamental properties of media entities

• US 6,913,466 B2
• Filed: 08/21/2001
• Issued: 07/05/2005
• Est. Priority Date: 08/21/2001
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Stanfield, Geoffrey R., Bassman, Eric
• Primary Examiner(s): ADAMS, CHANDA L
• Application Number:
  US09/934,774
• Publication Number:
  US 20030041066A1
• Time in Patent Office: 1,414 Days
• Field of Search: 434/307.R, 434/219, 434/353, 844/70.R, 707/7, 707/104.1, 707/5, 707/102, 369/30.39, 700/214, 705/11
• US Class Current: 434/219
• CPC Class Codes: G06F 16/634 Query by example, e.g. query by humming
• CPC Class Codes: G06F 16/636 by using biological or physiological data
• CPC Class Codes: G06F 16/639 using playlists
• CPC Class Codes: G06F 16/683 using metadata automatically derived from the content
• CPC Class Codes: G09B 15/00 Teaching music metronomes G04F5/02
• CPC Class Codes: G09B 5/00 Electrically-operated educational appliances working with questions and answers G09B7/00; simulators G09B9/00; advertising or displaying in general G09F

Abstract

A system and methods are provided for training a trainee to analyze media, such as music, in order to recognize and assess the fundamental properties of any piece of media, such as a song or a segment of a song. The process includes an initial tutorial and a double grooving process. The tutorial phase exposes the trainee to a canonical set of classifications and then exposes the trainee to certain definitive song examples for each classification level of fundamental properties. The double grooving phase leverages the skills of the experts that defined the canonical set of classification terms to ensure that new listeners, even though exposed to the tutorial, appropriately recognize all fundamental musical properties. Thus, for specific song examples, a new listener matches results with the system experts within a degree of tolerance. When a high enough degree of cross-listening consensus is reached, the new listener becomes a groover and can classify new songs or segments of songs.

Claims

1. A method for training a trainee to assign to a media entity at least one value corresponding to at least one fundamental property of a plurality of fundamental properties of the type of media entity via a computing device, comprising:

rendering definitional classification information to the trainee to educate the trainee as to the nature of said plurality of fundamental properties;

receiving by the computing devices at least one trainee-assigned value corresponding to said at least one fundamental property of the plurality of fundamental properties of the media entity after said rendering, said at least one trainee-assigned value equal in number to at least one expert-assigned value assigned to said at least one fundamental property of said plurality of fundamental properties of the media entity by at least one expert;

comparing by the computing device the at least one trainee-assigned value to the corresponding at least one expert-assigned value; and

determining based on said comparing a first group of said at least one fundamental property of the plurality of fundamental properties for which said trainee is qualified to code values for new media entities, wherein the media entities are one of song and song segments.

2. The method for training according to claim 1, wherein said comparing includes comparing, value by value, the at least one trainee-assigned value to the corresponding at least one expert-assigned value.

3. The method for training according to claim 1, further including determining based on said comparing a second group of said at least one fundamental property of the plurality of fundamental properties for which said trainee is not qualified to code values for new media entities.

4. The method for training according to claim 3, further including redoing the rendering, receiving, comparing and determining for fundamental properties of the second group until all properties in said second group are in said first group.

5. The method for training according to claim 4, wherein when all fundamental properties are in said first group, said trainee is a groover for all fundamental properties.

6. The method for training according to claim 1, wherein said definitional information includes definitional information about rhythm, zing and mood.

7. The method according to claim 6, wherein said definitional information for rhythm comprises definitional information for tempo, a time signature, rhythm description, rhythm type and rhythmic activity.

8. The method according to claim 6, wherein said definitional information for zing comprises definitional information for consonance, density, melodic movement and weight.

9. The method according to claim 6, wherein said definitional information for mood comprises definitional information for emotional intensity, mood and mood description.

10. The method for training according to claim 1, wherein said rendering includes rendering said definitional classification information to the trainee via the Web.

11. The method for training according to claim 1, wherein said comparing includes performing statistical analysis on said at least one trainee-assigned value and said at least one expert-assigned value.

12. The method for training according to claim 11, wherein said comparing includes calculating correlations between said at least one trainee-assigned value and said at least one expert-assigned value.

13. The method for training according to claim 1, wherein said comparing includes at least one of 1) Taking a batch of songs and calculating correlation scores across a set of specified fundamental properties, 2) Taking a batch of songs and calculating the percentage of songs in which the at least one expert and the trainee are within plus/minus one classification scaling from each other across a set of specified fundamental properties and 3) examining song-by-song every property for that song, with co-listening and discussion of what is heard via the definition parameters provided for each fundamental property.

14. The method for training according to claim 1, wherein said comparing includes at least one of comparing with a statistical analysis and comparing with a non-statistical analysis.

15. The method for training according to claim 1, wherein said trainee is authorized to code new media entities for said first group of said at least one fundamental property of the plurality of fundamental properties.

16. The method for training according to claim 1, wherein said rendering of said definitional information includes rendering at least one of a song segment and song to said trainee, said at least one of a song segment and song serving as at least one example of said at least one fundamental property of said plurality of fundamental properties.

17. A The method for training according to claim 16, wherein said at least one of a song segment and song are selected from a playlist generating engine capable of matching songs to the at least one fundamental property of said plurality of fundamental properties.

29. The computer executable instructions of claim 1 further comprising instructions whereby said comparing includes comparing, value by value, the at least one trainee-assigned value to the corresponding at least one expert-assigned value.

30. The computer executable instructions of claim 1 further comprising instructions whereby said rendering includes rendering said definitional classification information to the trainee via the Web.

18. A computing device comprising:

means for rendering definitional classification information to a trainee to educate the trainee about a plurality of fundamental properties associated with a type of media entity;

means for receiving by the computing device at least one trainee-assigned value corresponding to said at least one fundamental property of the plurality of fundamental properties of the media entity, said at least one trainee-assigned value equal in number to at least one expert-assigned value assigned to said at least one fundamental property of the plurality of fundamental properties of the media entity by at least one expert;

means for comparing by the computing device the at least one trainee-assigned value to the corresponding at least one expert-assigned value; and

means for determining based on at least one output from said means for comparing a group of said at least one fundamental property of said plurality of fundamental properties for which said trainee is qualified to code values for new media entities, wherein the media entities are one of songs and song segments.

19. A computing system for training a trainee to classify music, comprising:

a display for rendering definitional classification information to the trainee to educate the trainee as to the nature of a plurality of fundamental music properties associated with musical compositions; and

audio rendering means for rendering at least one of a song segment and song to said trainee, said at least one of a song segment and song serving as at least one example of at least one fundamental property of said plurality of fundamental music properties associated with musical compositions;

means for receiving from a trainee classification data for classifying said at least one of a song segment and song rendered by said audio rendering means;

means for analyzing said classification data; and

means for determining whether said trainee is qualified to enter classification data for said at least one fundamental music property of said plurality of fundamental music properties associated with musical compositions.

20. The computing system for training according to claim 19, wherein said means for analyzing includes means for comparing said classification data to known classification data for said at least one of a song segment and song.

21. The computing system for training according to claim 20, wherein said means for comparing includes means for performing statistical analysis on said classification data from the trainee and said known classification data.

22. The computing system for training according to claim 21, wherein said means for comparing includes means for calculating correlations between the classification data from the trainee and said known classification data.

23. The computing system for training according to claim 19, wherein said definitional information includes definitional information about rhythm, zing and mood.

24. The computing system according to claim 23, wherein said definitional information for rhythm comprises definitional information for tempo, a time signature, rhythm description, rhythm type and rhythmic activity.

25. The computing system according to claim 23, wherein said definitional information for zing comprises definitional information for consonance, density, melodic movement and weight.

26. The computing system according to claim 23, wherein said definitional information for mood comprises definitional information for emotional intensity, mood and mood description.

27. A method for training a trainee via a computing device to analyze music in order to recognize and assess the fundamental musical properties thereof, the method comprising:

providing a trainee with a list of fundamental musical properties grouped into three main areas;

rhythm, zing and mood;

providing the trainee with written definitions for the three main areas;

rhythm, zing and mood;

selecting via the computing device one of the three main areas;

rhythm, zing and mood;

displaying to the trainee a list of song examples organized by classification level;

playing via the computing device to the trainee the song examples one-by-one, progressing through each classification level;

repeating the selecting, displaying and playing for the remaining two of the three main areas;

rhythm, zing and mood;

playing via the computing device songs to the trainee and a previously trained listener;

receiving codes assigned to the songs by the trainee and the previously trained listener according to the three main areas;

rhythm, zing and mood and according to classification level;

comparing the codes assigned by the trainee with the codes assigned by the trained listener; and

determining based on said comparing which, if any, of said three main areas said trainee is qualified to assign codes for new media entities.

28. A computer readable medium bearing computer executable instructions for training a trainee to assign to a media entity at least one value corresponding to at least one fundamental property of a plurality of fundamental properties of the type of media entity via a computing device, said executable instructions comprising instructions for:

rendering definitional classification information to the trainee to educate the trainee as to the nature of said plurality of fundamental properties;

receiving by the computing device at least one trainee-assigned value corresponding to said at least one fundamental property of the plurality of fundamental properties of the media entity after said rendering, said at least one trainee-assigned value equal in number to at least one expert-assigned value assigned to said at least one fundamental property of said plurality of fundamental properties of the media entity by at least one expert;

comparing by the computing device the at least one trainee-assigned value to the corresponding at least one expert-assigned value; and

determining based on said comparing a first group of said at least one fundamental property of the plurality of fundamental properties for which said trainee is qualified to code values for new media entities, wherein the media entities are one of songs and song segments.