Azure-docs: What's the maximum limit for Batch Transcription? I mean file size and duration

[1c7]

For example, max file size is 100mb?
max duration is 1 hours?

---

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

• ID: 73fb7e76-fb80-c420-2efe-e3ca3a27b749
• Version Independent ID: 07a42557-deb3-e4b3-3f14-f26a27be49c2
• Content: What is batch transcription - Speech service - Azure Cognitive Services
• Content Source: articles/cognitive-services/Speech-Service/batch-transcription.md
• Service: cognitive-services
• Sub-service: speech-service
• GitHub Login: @wolfma61
• Microsoft Alias: wolfma

[1c7]

I am using Batch Transcription: https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/batch-transcription

I am not using any SDK, just get auth token using HTTP and create transcription task with REST API.

Here is how I do it

1. Upload audio file to Azure Storage Account's Container
2. Get a public accessible URL for that file
3. Send that URL to batch transcription REST API ("Creates a new transcription" POST api/speechtotext/v2.0/transcriptions)

Question

What's the limit? max file size 100mb or? max duration is 1 hour or?

[souravmishra-msft]

@1c7, thank you for reaching out. We are looking into this and would get back to you soon on this thread.

[ram-msft]

@1c7 Can you please add more detail about the input audio file's that you are trying.
A standard subscription (S0) for Speech service is required to use batch transcription. Please follow below details for limits.

https://azure.microsoft.com/en-us/pricing/details/cognitive-services/speech-services/
Please follow the below for faq.
https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/faq-stt
The limits above and in the doc are the default limits. We do work closely with large customers and can change the quota when necessary.

Current API_2.0 is only single file based, A version3 of the batch api is going to become available,This API version will allow you to supply a container as the input.

[1c7]

@ram-msft
Hi, the input audio file comes from the user, so it could be any format or size or duration.
I am trying to find out what's the limit,
so my program can say: Sorry, the file is too big/duration is too long

the FAQ https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/faq-stt
didn't say file size limit and duration limit.

This is request frequency limit

I don't need to train model, I just use the default one

Conclusion: problem still unsolved

Current API_2.0 is only single file based, A version3 of the batch api is going to become available,This API version will allow you to supply a container as the input.

that's great but still haven't answer what's the limit for single file?

[wolfma61]

there are no hard-coded limits in the batch transcription service itself regarding file size or audio duration.

a batch transcription request must be finished withing 48hours (currently), once it has started processing. This includes downloading the audio blob, transcribing, uploading the result data. We are transcribing with up-to double realtime speed. All these parameters are internal and can change, the service usage might be relevant, or available space (memory / disk).

I would recommend staying in a manageable space of several hours of audio. Longer files I would actually split to parallelize the upload of the audio and also the processing of the audio. Splitting 20 hours of audio in 10 segments of 2 hours might get you the transcription results in a couple of hours, as a big file you will have to wait at least 10 hours or so.

[1c7]

@wolfma61 Thank you :)