Dynamically generated audio in advertisements
A content server provides a client device with audio content including an audio advertisement, which is provided in response to receiving a request for digital audio content from a client device associated with a user. The content server obtains user information about the user and retrieves advertisement text received from an advertiser, which are used to generate a personalized text advertisement. The personalized text advertisement is generated according to an advertisement template specifying an ordered combination of text components. The personalized text advertisement includes the received advertisement text, user information text selected from the obtained user information, and template text. The client device is provided with an advertisement based on the personalized text advertisement and is configured to play an audio version of the personalized text advertisement. The audio advertisement is generated using a text-to-speech algorithm at the client device or at the content server.
1. A computer-implemented method for providing an audio message using personalized text, the method comprising:
receiving, from a song streaming application installed on a client device associated with a user, by a song streaming server, a request to stream song content to the client device for playback at the client device;
generating personalized text for the user based at least in part on retrieved message content;
identifying song features of an item of song content provided to the client device as part of the streamed song content for playback immediately adjacent to the audio message based on a position of the item of song content in a playlist associated with the streaming of the song content;
selecting a voice for a text-to-speech algorithm, the selected voice having vocal parameters determined based on the identified song features of the item of song content; and
causing the audio message to play on the client device by instructing the song streaming application to play back the audio message, the audio message based on an audio version of the personalized text generated by the text-to-speech algorithm using the selected voice and played adjacent to the item of song content.
2. The method of claim 1 , wherein generating the personalized text comprises:
retrieving content data describing the item of song content; and
generating the personalized text based at least in part on the content data, the personalized text comprising content text determined using the content data.
3. The method of claim 2 , wherein generating the personalized text further comprises:
retrieving targeting criteria received from a provider of the message content, the targeting criteria specifying bibliographic information about the item of song content, the targeting criteria associated with the message content;
determining whether the item of song content matches the bibliographic information specified by the targeting criteria; and
generating the personalized text responsive to determining that the item of song content matches the bibliographic information.
4. The method of claim 1 , wherein generating the personalized text further comprises:
retrieving content data describing the item of song content;
identifying supplemental content information relevant to the content data; and
generating the personalized text based at least in part on the supplemental content information.
5. The method of claim 1 , wherein selecting the voice for the text-to-speech algorithm comprises:
determining audio features describing a vocalist in the item of song content; and
determining the vocal parameters for the selected voice responsive to the song features describing the vocalist.
6. The method of claim 1 , further comprising:
determining preferences of the user for song features; and
determining the vocal parameters for the selected voice responsive to the preferences of the user for song features.
7. The method of claim 1 , further comprising:
inferring profile vocal parameters of the user responsive at least in part to user profile data describing demographic information about the user;
inferring content vocal parameters of the user responsive at least in part to content data describing song content provided to the client device; and
determining the vocal parameters for the selected voice responsive to a weighted combination of the profile vocal parameters and the content vocal parameters, wherein the weighted combination weighs the content vocal parameters more heavily than the profile vocal parameters.
8. A non-transitory computer-readable storage medium comprising computer program instructions executable by one or more processors to perform operations comprising:
receiving, from a song streaming application installed on a client device associated with a user, by a song streaming server, a request to stream song content to the client device for playback at the client device;
generating personalized text for the user based at least in part on retrieved message content;
identifying song features of an item of song content provided to the client device as part of the streamed song content for playback immediately adjacent to the audio message based on a position of the item of song content in a playlist associated with the streaming of the song content;
selecting a voice for a text-to-speech algorithm, the selected voice having vocal parameters determined based on the identified song features of the item of song content; and
causing the audio message to play on the client device by instructing the song streaming application to play back the audio message, the audio message based on an audio version of the personalized text generated by the text-to-speech algorithm using the selected voice and played adjacent to the item of song content.
9. The computer-readable medium of claim 8 , wherein the instructions for generating the personalized text comprise instructions for:
retrieving content data describing the item of song content; and
generating the personalized text based at least in part on the content data, the personalized text comprising content text determined using the content data.
10. The computer-readable medium of claim 9 , wherein the instructions for generating the personalized text further comprise instructions for:
retrieving targeting criteria received from a provider of the message content, the targeting criteria specifying bibliographic information about the item of song content, the targeting criteria associated with the message content;
determining whether the item of song content matches the bibliographic information specified by the targeting criteria; and
generating the personalized text responsive to determining that the item of song content matches the bibliographic information.
11. The computer-readable medium of claim 8 , wherein instructions for generating the personalized text comprise instructions for:
retrieving content data describing the item of song content;
identifying supplemental content information relevant to the content data; and
generating the personalized text based at least in part on the supplemental content information.
12. The computer-readable medium of claim 8 , wherein selecting the voice for the text-to-speech algorithm comprises:
determining song features describing a vocalist in the item of song content; and
determining the vocal parameters for the selected voice responsive to the song features describing the vocalist.
13. The computer-readable medium of claim 8 , the operations further comprising:
determining preferences of the user for song features; and
determining the vocal parameters for the selected voice responsive to the preferences of the user for song features.
14. The computer-readable medium of claim 8 , the operations further comprising:
inferring profile vocal parameters of the user responsive at least in part to user profile data describing demographic information about the user;
inferring content vocal parameters of the user responsive at least in part to content data describing song content provided to the client device; and
determining the vocal parameters for the selected voice responsive to a weighted combination of the profile vocal parameters and the content vocal parameters, wherein the weighted combination weighs the content vocal parameters more heavily than the profile vocal parameters.
15. A system for providing an audio message using personalized text, the system comprising:
one or more processors; and
a non-transitory computer-readable storage medium comprising computer program instructions executable by the one or more processors to perform operations comprising:
receiving, from a song streaming application installed on a client device associated with a user, by a song streaming server, a request to stream song content to the client device for playback at the client device;
generating personalized text for the user based at least in part on retrieved message content;
identifying song features of an item of song content provided to the client device as part of the streamed song content for playback immediately adjacent to the audio message based on a position of the item of song content in a playlist associated with the streaming of the song content;
selecting a voice for a text-to-speech algorithm, the selected voice having vocal parameters determined based on the identified song features of the item of audio song content; and
causing the audio message to play on the client device by instructing the song streaming application to play back the audio message, the audio message based on an audio version of the personalized text generated by the text-to-speech algorithm using the selected voice and played adjacent to the item of song content.
16. The system of claim 15 , wherein the instructions for generating the personalized text comprise instructions for:
retrieving content data describing the item of song content; and
generating the personalized text based at least in part on the content data, the personalized text comprising content text determined using the content data.
17. The system of claim 16 , wherein the instructions for generating the personalized text comprise instructions for:
retrieving targeting criteria received from a provider of the message content, the targeting criteria specifying bibliographic information about the item of song content, the targeting criteria associated with the message content;
determining whether the item of song content matches the bibliographic information specified by the targeting criteria; and
generating the personalized text comprising the content text and the message content responsive to determining that the item of song content matches the bibliographic information.
18. The system of claim 15 , wherein selecting the voice for the text-to-speech algorithm comprises:
determining song features describing a vocalist in the item of song content; and
determining the vocal parameters for the selected voice responsive to the song features describing the vocalist.
19. The system of claim 15 , the operations further comprising:
determining preferences of the user for song features; and
determining the vocal parameters for the selected voice responsive to the preferences of the user for song features.
20. The system of claim 15 , the operations further comprising:
inferring profile vocal parameters of the user responsive at least in part to user profile data describing demographic information about the user;
inferring content vocal parameters of the user responsive at least in part to content data describing song content provided to the client device; and
determining the vocal parameters for the selected voice responsive to a weighted combination of the profile vocal parameters and the content vocal parameters, wherein the weighted combination weighs the content vocal parameters more heavily than the profile vocal parameters.