Methods and apparatus to determine audio source impact on an audience of media
Methods, apparatus, systems and articles of manufacture to determine audio source impact on an audience of media are disclosed. A disclosed example apparatus includes at least one memory, instructions in the apparatus, and processor circuitry to execute the instructions to: divide audio of monitored media into successive audio segments; perform speaker identification on the audio segments; generate confidence values for speakers identified in the audio segments; generate speaker identification data for ones of the speakers having respective confidence values that satisfy a threshold; and analyze the speaker identification data to determine a speaker impact on audience ratings data.
1. An apparatus comprising:
at least one memory;
instructions in the apparatus; and
processor circuitry to execute the instructions to:
divide audio of monitored media into successive audio segments;
perform speaker identification on the audio segments;
generate confidence values for speakers identified in the audio segments;
generate speaker identification data for ones of the speakers having respective confidence values that satisfy a threshold; and
analyze the speaker identification data to determine a speaker impact on audience ratings data.
2. The apparatus of claim 1 , wherein the processor circuitry is to at least one of change a frequency spectrum of the audio, adjust properties of the audio, or change a file format to preprocess the audio of the monitored media before the speaker identification is performed.
3. The apparatus of claim 1 , wherein the processor circuitry is to execute a logistic regression algorithm to generate the confidence values.
4. The apparatus of claim 1 , wherein to generate the confidence values, the processor circuitry is to:
generate patterns from the audio segments; and
compare the patterns to training samples stored in a training sample database.
5. The apparatus of claim 1 , wherein the processor circuitry is to determine the speaker identification data by identifying a dominate speaker and time variant confidence values associated with the identified dominate speaker.
6. The apparatus of claim 1 , wherein the speaker identification data is minute-by-minute speaker identification data, and the processor circuitry is to compare minute-by-minute audience ratings of the monitored media and the minute-by-minute speaker identification data to determine how a first one of the identified speakers affected audience viewership of the monitored media.
7. The apparatus of claim 6 , wherein the processor circuitry is to correlate changes in the minute-by-minute audience ratings to changes in the minute-by-minute speaker identification data to determine if the first one of the identified speakers is associated with a decrease or increase in audience ratings.
8. At least one non-transitory computer readable medium comprising instructions that, when executed, cause processor circuitry to at least:
divide audio of monitored media into successive audio segments;
perform speaker identification on the audio segments;
generate confidence values for speakers identified in the audio segments;
generate speaker identification data for ones of the speakers having respective confidence values that satisfy a threshold; and
analyze the speaker identification data to determine a speaker impact on audience ratings data.
9. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions cause the processor circuitry to at least one of change a frequency spectrum of the audio, adjust properties of the audio, or change a file format to preprocess the audio of the monitored media before the speaker identification is performed.
10. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions cause the processor circuitry to execute logistic regression algorithm to generate the confidence values.
11. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions cause the processor circuitry to generate patterns from the audio segments and compare the patterns to training samples stored in a training sample database to generate the confidence values.
12. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions cause the processor circuitry to determine the speaker identification data by identifying a dominate speaker and time variant confidence values associated with the identified dominate speaker.
13. The at least one non-transitory computer readable medium of claim 8 , wherein the speaker identification data is minute-by-minute speaker identification data, and wherein the instructions cause the processor circuitry to compare minute-by-minute audience ratings of the monitored media and the minute-by-minute speaker identification data to determine how a first one of the identified speakers affected audience viewership of the monitored media.
14. The at least one non-transitory computer readable medium of claim 13 , wherein the instructions cause the processor circuitry to correlate changes in the minute-by-minute audience ratings to changes in the minute-by-minute speaker identification data to determine if the first one of the identified speakers is associated with a decrease or increase in audience ratings.
15. A method comprising:
dividing, by executing an instruction with at least one processor, audio of monitored media into successive audio segments;
performing, by executing an instruction with at least one processor, speaker identification on the audio segments;
generating, by executing an instruction with at least one processor, confidence values for speakers identified in the audio segments;
generating, by executing an instruction with at least one processor, speaker identification data for ones of the speakers having respective confidence values that satisfy a threshold; and
analyzing, by executing an instruction with at least one processor, the speaker identification data to determine a speaker impact on audience ratings data.
16. The method of claim 15 , wherein the generating of the confidence values includes executing logistic regression algorithm.
17. The method of claim 15 , wherein the generating of the confidence values includes generating patterns from the audio segments and comparing the patterns to training samples stored in a training sample database.
18. The method of claim 15 , wherein the determining of the speaker identification data includes identifying a dominate speaker and time variant confidence values associated with the identified dominate speaker.
19. The method of claim 15 , wherein the speaker identification data is minute-by-minute speaker identification data, and further including comparing minute-by-minute audience ratings of the monitored media and the minute-by-minute speaker identification data to determine how a first one of the identified speakers affected audience viewership of the monitored media.
20. The method of claim 19 , further including correlating changes in the minute-by-minute audience ratings to changes in the minute-by-minute speaker identification data to determine if the first one of the identified speakers is associated with a decrease or increase in audience ratings.