We are currently serving pre-authenticated URLs from SharePoint to the Upload Endpoint. Including the access token, a few other parameters and the file name, these often exceed the 2048 character limit of the videoUrl parameter.
It would be great if you could either increase the length limit, or (even better) allow the videoUrl to be part of the post request content, so that we can continue indexing files directly from SharePoint.

Currently, Video Indexer is managed completely outside of the Azure portal via its own site.

This poses a number of enterprise DevOps challenges in being able to automatically provision and setup the service for use for customers.

It would be great to see this rolled into the Azure Portal, just like all the other Cognitive Services currently on offer, with associated automation options (eg. ARM templates) to allow for customers to easily provision the service.

Incorrect translation in Japanese Portal.

The word "平行移動"(origin "Translate") on the play / edit screen is a mistranslation.

The correct translation in Japanese is "翻訳".

got a error! that's all folks nothing describes what the error is

Video scenes are created automatically, but can you add or change scenes later?

Can you add Japanese to Thai translation and Thai to Japanese translation function?

Currently there is a function to correct subtitle data of video, but there is no function to add subtitle data, so can you create a function to add subtitles at a specific time?

HI,

There should be a way to add transition like simple fade to avoid a sharp cut. This is help to create a seamless video.

Regards

It would be awesome if there was a basic implementation for Video Indexer tools on a web app, like an MVC 5 web app, to have the ability to upload videos and saw the videos and the indexation on the interface via that application selecting them. And avoid the necessity of consult videoindexer.ai all the time.

Specific Japanese title name of video is garbled. In my test case, Japanese characters 'クロ' is garbled.

I don't see the ability to create new labels/tags. The AI got some wrong, I see where I can delete them, but would like to then add correct tags.

Isn't the 'People Model' concept same as the Face APIs 'Person Group'?
Is there a plan to have both the APIs integrated so that the Person Identification can happen through Video or Still Image inputs?

Users should be able to manually increase or decrease the segment selected, right now this is strictly done on insights.

Use case- a basketball highlight is picked up by the broadcaster saying the player's name, an insight. If the end user wanted to start the segment a few seconds before the insight, and extend it for a few seconds after, this will be HUGE!

The video is 3.26 hours long but the text provide is not almost a third of the audio.

When using Video Indexer i noticed that, even if a language is defined, the topics come in English only. Would be really wonderful if the topics were also in the specified language.