Audio File Quality and Accuracy Assessment

• US 20150331941A1
• Filed: 05/16/2014
• Published: 11/19/2015
• Est. Priority Date: 05/16/2014
• Status: Active Grant

• Current Assignee: Gracenote Digital Ventures LLC (RR Donnelley & Sons Company)
• Original Assignee: Tribune Digital Ventures, LLC (Nexstar Media Group, Inc.)
• Inventors: Defouw, Gregory P., Panguluri, Venkatarama Anilkumar
• Granted Patent: US 10,776,419 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/61 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 3/165 Management of the audio stream, e.g. setting of volume, audio stream path
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 25/03 characterised by the type of extracted parameters
• CPC Class Codes: G10L 25/51 for comparison or discrimination
• CPC Class Codes: G10L 25/69 for evaluating synthetic or decoded voice signals
• CPC Class Codes: H04N 21/233 Processing of audio elementary streams monitoring, identification or recognition of audio in broadcast systems H04H60/58
• CPC Class Codes: H04N 21/25891 being end-user preferences retrieval of video data in a video database based on user preferences G06F16/739; arrangements for recognizing users' preferences H04H60/46; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18
• CPC Class Codes: H04N 21/26258 for generating a list of items to be played back in a given order, e.g. playlist, or scheduling item distribution according to such list retrieval of multimedia data based on playlists G06F16/40
• CPC Class Codes: H04N 21/8106 involving special audio data, e.g. different tracks for different languages
• CPC Class Codes: H04N 21/84 Generation or processing of descriptive data, e.g. content descriptors systems specially adapted for using meta-information in broadcast systems H04H60/73

Abstract

Disclosed computer-based systems and methods for analyzing a plurality of audio files corresponding to text-based news stories and received from a plurality of audio file creators are configured to (i) compare quality and/or accuracy metrics of individual audio files against corresponding quality and/or accuracy thresholds, and (ii) based on the comparison: (a) accept audio files meeting the quality and/or accuracy thresholds for distribution to a plurality of subscribers for playback, (b) reject audio files failing to meet one or more certain quality and/or accuracy thresholds, (c) remediate audio files failing to meet certain quality thresholds, and (d) designate for human review, audio files failing to meet one or more certain quality and/or accuracy thresholds by a predetermined margin.

Claims

1. A method comprising:

receiving a plurality of audio files from a plurality of audio file creators, wherein an individual audio file of the plurality of audio files comprises an audible version of a corresponding text-based news story;

storing the plurality of received audio files in a tangible, non-transitory computer readable media;

determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file;

in response to determining that the individual audio file is acceptable for distribution, approving the individual audio file for distribution; and

in response to determining that the individual audio file is not acceptable for distribution, at least one of (i) rejecting the individual audio file, (ii) remediating the individual audio file, or (iii) designating the individual file for review by a human.

2. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

generating a text file from the individual audio file;

for each word in the individual audio file'"'"'s corresponding text-based news story, determining whether the word matches its corresponding word in the generated text file; and

determining whether a number of words in the individual audio file matching corresponding words in the individual audio file'"'"'s corresponding text-based news story is greater than or equal to a word match threshold.

3. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

determining a number of words in the individual audio file'"'"'s corresponding text-based news story;

determining a duration of the individual audio file;

estimating a number of words in the individual audio file based on the determined duration and an estimated words-per-minute rate of the individual audio file; and

determining whether a difference between the determined number of words in the individual audio file'"'"'s corresponding text-based news story and the estimated number of words in the individual audio file is less than or equal to a word count threshold.

4. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

determining a number of sentences in the individual audio file'"'"'s corresponding text-based news story;

estimating a number of sentences in the individual audio file based at least in part on an analysis of an audio waveform corresponding to the individual audio file; and

determining whether a difference between the determined number of sentences in the individual audio file'"'"'s corresponding text-based news story and the estimated number of sentences in the individual audio file is less than or equal to a sentence count threshold.

5. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

estimating a duration of the audio file based on a number of words in the audio file'"'"'s corresponding text-based news story and a target audio pace;

determining an actual duration of the audio file; and

determining whether a difference between the estimated duration and the actual duration is less than or equal to a duration threshold.

6. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

determining a number of words in the individual audio file'"'"'s corresponding text-based news story;

determining a duration of the individual audio file;

calculating a words-per-minute audio pace of the individual audio file by dividing the determined number of words in the individual audio file'"'"'s corresponding text-based news story by the determined duration of the individual audio file; and

determining whether the calculated audio pace is both (i) less than or equal to a maximum audio pace threshold and (ii) greater than or equal to a minimum audio pace threshold.

7. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

determining whether background noise in the individual audio file is less than or equal to a background noise threshold.

8. The method of claim 7, further comprising:

in response to determining whether background noise in the individual audio file is not less than or equal to a background noise threshold, generating a noise-reduced version of the audio file by applying one or more noise reduction algorithms to the audio file; and

determining whether background noise in the noise-reduced version of audio file is less than or equal to a background noise threshold.

9. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

determining whether an average volume of the individual audio file is (i) less than or equal to a maximum average volume threshold and (ii) greater than or equal to a minimum average volume threshold.

10. The method of claim 9, wherein the audio file has a dynamic range, and wherein the method further comprises:

compressing the dynamic range of the individual audio file in response to determining that the average volume of the individual audio file is neither (i) less than or equal to a maximum average volume threshold nor (ii) greater than or equal to a minimum average volume threshold; and

determining whether an average volume of the dynamic range compressed audio file is (i) less than or equal to a maximum average volume threshold and (ii) greater than or equal to a minimum average volume threshold.

11. The method of claim 1, wherein determining whether the individual audio file is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file comprises:

determining whether a peak volume of the individual audio file is less than or equal to a peak volume threshold.

12. The method of claim 11, further comprising:

in response to determining that the peak volume of the individual audio file less than or equal to the peak volume threshold, applying at least one of a limiter function or a dynamic range compression function to the individual audio file; and

determining whether a peak volume of the limited or dynamic range compressed audio file is than or equal to the peak volume threshold.

13. The method of claim 1, wherein remediating the individual audio file comprises one or more of (i) applying a noise reduction algorithm to the individual audio file, (ii) applying a dynamic range compression algorithm to the individual audio file, and/or (iii) applying a limiting function to the individual audio file.

14. The method of claim 1, wherein rejecting the individual audio file comprises transmitting an indication to an audio file creator associated with the individual audio file, the indication comprising one or more of:

(i) a notification that the individual audio file has been rejected, (ii) at least one reason for rejecting the audio file, and (iii) a request that the audio file creator resend a corrected version of the individual audio file.

15. A system comprising:

a machine review queue comprising tangible, non-transitory computer readable media, and configured to store a plurality of audio files received from a plurality of audio file creators, wherein an individual audio file comprises an audible version of a corresponding text-based news story;

an audio analysis engine comprising one or more processors, wherein the audio analysis engine is configured to analyze one or more accuracy metrics of an individual audio file against one or more corresponding accuracy thresholds, and based on the analysis, determine whether the individual audio file is sufficiently accurate for distribution to a plurality of subscribers for playback; and

a distribution database comprising tangible, computer readable media, and configured to store individual audio files determined by the audio analysis engine to be sufficiently accurate for distribution to the plurality of subscribers for playback.

16. The system of claim 15, wherein the one or more accuracy metrics comprise one or more of:

a word match metric based on a number of words in a speech-to-text version of the individual audio file that match corresponding words in the individual audio file'"'"'s corresponding text-based news story;

a word count metric based on a difference between (i) a number of words in the individual audio file'"'"'s corresponding text-based news story and (ii) an estimated number of words in the individual audio file;

a sentence count metric based on a difference between (i) a number of sentences in the individual audio file'"'"'s corresponding text-based news story and (ii) an estimated number of sentences in the individual audio file;

a duration metric based on a difference between (i) an actual duration of the individual audio file and (ii) an estimated duration of the audio file, wherein the estimated duration is based on a number of words in the audio file'"'"'s corresponding text-based news story and a target audio pace; and

an audio pace metric based on a difference between (i) a target audio pace and (ii) a calculated audio pace for the individual audio file, wherein the calculated audio pace is based on the number of words in the individual audio file'"'"'s corresponding text-based news story and the actual duration of the individual audio file.

17. The system of claim 15, wherein the audio analysis engine is further configured to analyze one or more quality measures of the individual audio file against one or more corresponding quality thresholds, and based on the analysis, determine whether the individual audio file is of sufficient quality for distribution to the plurality of subscribers for playback.

18. The system of claim 17, wherein the one of more quality measures comprise one or more of:

a background noise metric based whether background noise measured in the individual audio file exceeds a background noise threshold;

an average volume metric based on whether an average volume of the individual audio file is (i) less than or equal to a maximum average volume threshold and (ii) greater than or equal to a minimum average volume threshold; and

a peak volume metric based on whether a peak volume of the individual audio file is less than or equal to a peak volume threshold.

19. The system of claim 18 further comprising:

a remediation engine comprising one or more processors, wherein the remediation engine is configured to;

(i) apply a noise reduction algorithm to the individual audio file in response to the audio analysis engine determining that the individual audio file exceeds the background noise threshold;

(ii) apply a dynamic range compression algorithm to the individual audio file in response to the audio analysis engine determining that the individual audio file is neither is (ii-a) less than or equal to the maximum average volume threshold nor (ii-b) greater than or equal to the minimum average volume threshold; and

/or (iii) apply a limiting function to the individual audio file in response to the audio analysis engine determining that the individual audio file exceeds the peak volume threshold; and

a human review queue comprising tangible, non-transitory computer readable media, wherein the human review queue is configured to store individual audio files that the audio analysis engine determines should be reviewed by a human based on the audio analysis engine'"'"'s analysis of at least one of (i) the one or more accuracy metrics against the one or more corresponding accuracy thresholds, or (ii) the one or more quality metrics against the one or more corresponding quality thresholds.

20. A tangible, non-transitory computer readable media comprising instructions executable by one or more processors to cause the one or more processors to perform functions comprising:

receiving a plurality of audio files from a plurality of audio file creators, wherein each audio file comprises an audible version of a corresponding text-based news story;

storing the plurality of received audio files in a tangible, non-transitory computer readable media;

determining whether an individual audio file in the plurality of audio files is acceptable for distribution to a plurality of subscribers for playback based on a computer analysis of the individual audio file;

in response to determining that the individual audio file of the plurality of audio files is acceptable for distribution, approving the individual audio file for distribution; and

in response to determining that the individual audio file is not acceptable for distribution, at least one of (i) rejecting the individual audio file, (ii) remediating the individual audio file, or (iii) designating the individual file for review by a human.