Audio Stream Compaction for Lower ASR Transcription Charges
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio stream processing systems incur high costs due to fixed pricing models for speech-to-text services, regardless of audio file duration, leading to inefficient utilization of resources and increased expenses for enterprises handling large volumes of short audio streams.
Innovation Solution
Implementing a two-pronged audio stream processing technique that is either transaction-based or time-based, optimizing audio stream files before transmission to ASR services by concatenating or compacting them to reduce unnecessary speech content and align with cost-effective pricing models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If audio streams are transmitted individually to ASR services, then transcription accuracy is maintained, but service costs increase due to fixed pricing models
Solution Approach 1:
The patent combines multiple short audio streams into a single concatenated audio stream for batch processing by the ASR service. This merging approach reduces the number of individual transcription requests, thereby lowering service costs under fixed pricing models while maintaining transcription accuracy through proper stream separation markers.
Solution Approach 2:
The system implements a dual-mode processing architecture that can handle both individual stream transcription and batch concatenated stream transcription. This multi-functionality allows the system to select the most cost-effective approach based on audio stream characteristics and ASR service pricing models.
2Productivity
If audio streams are concatenated for batch processing, then transmission frequency increases and costs decrease, but processing complexity increases
Solution Approach 1:
The patent introduces separator markers between individual audio streams within a concatenated batch. This segmentation approach allows the ASR service to accurately distinguish and process each individual stream within the combined batch, maintaining transcription accuracy while enabling efficient batch processing.
Solution Approach 2:
The system performs preliminary concatenation and separator insertion of audio streams before transmission to the ASR service. This pre-processing step consolidates multiple streams into a single optimized payload, increasing transmission frequency and reducing individual transmission overhead while simplifying the overall processing workflow.
3Ease of operation
If fixed-fee pricing model is used for ASR services, then service simplicity is maintained, but cost efficiency decreases for short audio streams
Solution Approach 1:
The patent changes the parameter of audio stream presentation from individual separate streams to concatenated batches with separators. This parameter change enables the system to leverage fixed-fee pricing more effectively by reducing the total number of transcription requests, thereby improving cost efficiency without requiring changes to the ASR service pricing model itself.
Data Source
AI summary
Systems and methods for processing audio streams are disclosed herein. An audio stream including speech content is received. The audio stream is compacted to generate a compacted audio stream and the compacted audio stream is transmitted to an automatic speech recognition (ASR) service for transcription of the speech content to text content. In response to transmitting the compacted audio stream for transcription, text content, a transcription of the audio stream, is received from the ASR service, and transcription charges are determined.