Methods to employ concatenation in ASR service usage based on speech-to-text cost
Systems and methods for processing audio streams are disclosed herein. In a disclosed method, N number of audio streams is received from an independent source, each audio stream includes speech content. A set of the N audio streams is concatenated to generate a concatenated audio stream. Based on the N received audio streams or the set of audio streams, N−1 or one less that the total number of audio streams in the set of audio stream separators, respectively, are generated. An audio stream separator is inserted between every two adjacent audio streams of the concatenated audio stream to generate a single audio stream payload. The single audio stream payload is transmitted for transcription of the audio stream speech content to text content and in response to transmitting the payload, a text file is received including text content corresponding to the audio streams delineated by the audio stream separators.
1. A method of processing audio streams, the method comprising:
receiving a plurality of audio streams during a time window, each audio stream received from an independent source, including speech content, and having an audio stream duration;
selecting a subset of the plurality of audio streams based on the respective audio stream durations of the plurality of audio streams, the subset comprising N number of audio streams, wherein a combined duration of the N number of audio streams is below a threshold duration, and wherein the threshold duration is based on a transaction cost for speech-to-text conversion;
generating N−1 number of audio stream separators;
concatenating the N number of audio streams to generate a concatenated audio stream;
inserting an audio stream separator between every two adjacent audio streams of the concatenated audio stream to generate a single audio stream payload, each audio stream separator delineating a beginning of a next audio stream and an end of a preceding audio stream in the concatenated audio stream;
transmitting the single audio stream payload from a buffer for transcription of the speech content of the set of audio streams to text content, a transcription of each of the speech content of the set of audio streams; and
in response to transmitting the single audio stream payload for transcription, receiving a text file including the text content delineated with the audio stream separators.
2. The method of claim 1 , further comprising concatenating the N number of audio streams in the buffer.
3. The method of claim 2 , further comprising inserting the audio stream separator between every two adjacent audio streams of the concatenated audio stream in the buffer.
4. The method of claim 1 , wherein a size of the buffer is based on a multiple of a maximum speech content duration among the durations of the speech content of the N number of audio streams.
5. The method of claim 4 , wherein the maximum speech content duration corresponds to a minimum base transcription price.
6. The method of claim 4 , wherein minimum transcription charges for transcription of the audio streams are based on the maximum speech content duration.
7. The method of claim 1 , further comprising concatenating the N number of audio streams in a storage.
8. The method of claim 7 , further comprising inserting the audio stream separator between every two adjacent audio streams of the concatenated audio stream in the storage.
9. The method of claim 1 , wherein a size of the buffer is based on a duration of the time window.
10. The method of claim 1 , further comprising performing the transmitting the single audio stream payload from a buffer for transcription no later than an end of a wait time, the wait time starting from a request to onboard the buffer with one or more of the audio streams of the audio stream payload and ending at a subsequent request to re-onboard the buffer with a subsequent set of the plurality of audio streams.
11. The method of claim 1 , wherein performing the transmitting regardless of a density of the buffer to prevent noticeable delay in receiving the text file.
12. A system for processing audio streams, the system comprising: input/output (I/O) circuitry configured to:
receive a plurality of audio streams during a time window, each audio stream received from an independent source, including speech content, and having an audio stream duration; and
control circuitry configured to:
select a subset of the plurality of audio streams based on the respective audio stream durations of the plurality of audio streams, the subset comprising N number of audio streams, wherein a combined duration of the N number of audio streams is below a threshold duration, and wherein the threshold duration is based on a transaction cost for speech-to-text conversion;
generate N−1 number of audio stream separators;
concatenate the N number of audio streams to generate a concatenated audio stream;
insert an audio stream separator between every two adjacent audio streams of the concatenated audio stream to generate a single audio stream payload, each audio stream separator delineating a beginning of a next audio stream and an end of a preceding audio stream in the concatenated audio stream;
transmit the single audio stream payload from a buffer for transcription of the speech content of the set of audio streams to text content, a transcription of each of the speech content of the set of audio streams; and
in response to transmitting the single audio stream payload for transcription, receive a text file including the text content delineated with the audio stream separators.
13. The system of claim 12 , wherein:
a size of the buffer is based on a multiple of a maximum speech content duration among the durations of the speech content of the N number of audio streams.
14. The system of claim 13 , wherein the maximum speech content duration corresponds to a minimum base transcription price.
15. The system of claim 13 , wherein minimum transcription charges for transcription of the audio streams are based on the maximum speech content duration.
16. The system of claim 12 , wherein the control circuitry is further configured to:
concatenating the N number of audio streams in the buffer; and
inserting the audio stream separator between every two adjacent audio streams of the concatenated audio stream in the buffer.
17. The system of claim 12 , wherein a size of the buffer is based on a duration of the time window.
18. The system of claim 12 , wherein the control circuitry is further configured to:
concatenate the N number of audio streams in a storage; and
insert the audio stream separator between every two adjacent audio streams of the concatenated audio stream in the storage.
19. The system of claim 12 , wherein the control circuitry is further configured to perform the transmitting the single audio stream payload from a buffer for transcription no later than an end of a wait time, the wait time starting from a request to onboard the buffer with one or more of the audio streams of the audio stream payload and ending at a subsequent request to re-onboard the buffer with a subsequent set of the plurality of audio streams.
20. The system of claim 12 , wherein the control circuitry is further configured to perform the transmitting regardless of a density of the buffer to prevent noticeable delay in receiving the text file.