性能不佳和响应排队 [英] Poor Performance and Queued Responses

问题描述

我对Azure和媒体服务的使用还比较陌生.

我正在使用媒体服务对音频文件执行OCR和转录.我正在将文件上传到Blob存储，然后将它们传输到媒体服务并被触发进行转录，从而产生具有4个转录文件的资产.我发现 流程的执行速度如此之慢，在流程完成方面似乎存在滞后.

另外，当我完成一天的媒体服务工作并于第二天早上回来时，似乎有9页的转录文件已被处理，而无需我为它们进行处理.有什么理由吗?有没有 该过程实际发生的队列很长?

谢谢

Edele 

解决方案

埃德勒(Edele)，

您是否正在使用我们的v3 API执行此类处理?如果是这样，那么您可能会使用 VideoAnalyzer或AudioAnalyzer预设.有了这样的预设，您将 通过应用不同的AI技术，可以获得包含多个所谓的见解或元数据文件的输出.

如果您未使用我们的v3 API，请打开支持票证，其中包含您的媒体服务帐户的详细信息(名称，数据中心，提交工作的时间/日期等).

最后但并非最不重要的一点，如果使用v3 API，那么您将需要为您的帐户配置多个S3媒体保留单元-请参阅注释" 这里.

谢谢

Hi,

I'm relatively new to Azure and the use of Media Services.

I'm using media services to perform OCR and transcription on audio files. I am uploading the files to a blob store, they are then transferred to media services and triggered to be transcribed, resulting in an asset with 4 transcription files. I find the process has such a slow performance and there seems to be a lag in the process completing. 

Also, when I finish working with media services for the day and come back the next morning, there seems to be 9 pages worth of transcription files that have been processed without me calling the process for them. Is there any reason for this? Is there a huge queue for the process to actually happen?

Thanks, 

Edele 

解决方案

Hi Edele,

Are you using our v3 APIs to perform such processing? If so, then you are likely using the VideoAnalyzer or AudioAnalyzer presets. With such presets, you will get an output that contains multiple so-called insights or metadata files, from applying different AI tech.

If you are not using our v3 APIs, then please open a support ticket with the details of your Media Service account (name, data center, time/date when Jobs where submitted, etc.).

Last but not least, if using v3 APIs, then you will need to provision your account with multiple S3 Media Reserved Units - see the Note here. 

Thanks

这篇关于性能不佳和响应排队的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！