Never add video: prefix to transcription properties

[lfcnassif]

Just tested transcription on videos simply adding video into mimesToProcess in AudioTranscriptConfig. I thought it would be necessary to update the convertCommand (and that's why I exposed it to interested users) but it wasn't needed (reported by @hauck-jvsh before).

But transcription results are being prefixed by video:. That's not good because it splits results between different properties and it also avoids to embed the video transcriptions into chats properly.

[lfcnassif]

Tagging as enhancement since transcribing videos wasn't the original intent.

[hauck-jvsh]

I don't know if it is a good idea to expose this, as it will send the entire video file, which maybe large, over the network. In this case I think that the audio should be extracted in the client.

[wladimirleite]

I agree. For videos, running MPlayer locally before sending to transcription server makes more sense.

Another detail, maybe the conversion command can be slightly changed to improve performance, completely ignoring the video channel. Not sure if it already does that. I will check this later and update here.

[lfcnassif]

Thanks @hauck-jvsh, I agree, but the option was already exposed, user was already able to change the configuration, I just fixed the property name. Let's open another ticket to improve video transcription resource usage.

Thanks @tc-wleite, let us know if the command can be improved.

[lfcnassif]

I changed my mind and will tag this as bug.