Today at Microsoft Inspire, our Azure OpenAI Service and Azure AI Speech teams announced that OpenAI Whisper will be in preview soon. The OpenAI Whisper model has multi-lingual capabilities that offer precise and efficient transcription of human speech in 57 languages, and translation into English. It also creates transcripts with enhanced readability.
The benefits of running the OpenAI Whisper model in Azure include enterprise-grade security, privacy controls, and data processing capabilities that allow for customized solutions to fit specific business needs.
Azure OpenAI Service
Azure OpenAI Service enables developers to run the OpenAI Whisper model in Azure, mirroring the OpenAI Whisper API in features and functionality, including transcription and translation capabilities.
The Whisper model REST APIs for transcription and translation will be available from the Azure OpenAI Service portal.
Try it out in Azure AI Studio
Once OpenAI Whisper is in preview in Azure OpenAI Service, users will be able to use Whisper in Azure AI Studio.
- Users must apply for access to Azure OpenAI Service
- Once approved, users can visit the Azure portal and create an Azure OpenAI Service resource
- Once resource has been created, users can begin using Whisper model
Azure AI Speech
Within Azure AI Speech, users will be able to leverage the OpenAI Whisper model for batch transcription, to easily transcribe large volumes of audio content at scale. This capability is particularly useful for processing extensive collections of audio data stored within the Azure platform.
Users of Whisper in Azure AI Speech will benefit from our existing features including async processing, speaker diarization, customization, and larger file sizes. Azure AI Speech enhances Whisper transcription by enabling files up to 1GB in size and the ability to process large amounts of files by allowing you to batch up to 1000 files in a single request. Additionally, when using Azure AI Speech the recognition result will include word level timestamps, providing the ability to identify where in the audio each word was spoken. Speaker diarization is another beneficial feature of Azure AI Speech that identifies individual speakers in an audio file and labels their speech segments. This feature allows customers to distinguish between speakers, accurately transcribing their words, and to create more organized and structured transcription of audio files.
Try it out in Azure AI Speech Studio
Once OpenAI Whisper is in preview in Azure AI Speech, users will be able to experiment with the model in Azure AI Speech Studio. Before selecting a transcription model, it is recommended that customers assess both the Whisper model and Azure AI Speech's existing transcription model, as both offer unique capabilities and advantages.
To determine which model works best for your unique use case, experiment with both Azure AI Speech and Whisper models in Azure AI Speech Studio. Our Transcription try it out experience allows customers to compare models side by side to explore the differences between Whisper and Azure Speech models and is also available for free without signing in. The free trial limited to 5 files per batch and 1 minute of audio per file).
Getting started
The integration of the OpenAI Whisper model into Azure OpenAI Service and Azure AI Speech offers a host of benefits to businesses ranging from enhanced readability, efficient transcription of audio content at scale, multi-lingual capabilities, and customized solutions that fit specific business needs. The possibilities are endless with the integration of the OpenAI Whisper model into Azure AI services.
- Once in preview, you can use OpenAI Whisper in Azure OpenAI Service. If you don’t yet have access to Azure OpenAI Service, you can apply for access by completing the form at https://aka.ms/oai/access.
- Once in preview, you can use OpenAI Whisper in Azure AI Speech. Get started here.
- We also encourage you to explore and try batch transcription models in Azure AI Speech Studio.