Transcription and Captioning Improvements

I would like to request various improvements to the automatic transcription for source files.

1. Improved accuracy in general. We are seeing a lot of incorrect words and a huge amount of incorrect punctuation in the transcripts. I understand that punctuation is not always obvious in dialogue, but the amount of mistakes has been frustrating. The automatic transcription needs to do better with introductory clauses/phrases, in particular. It also needs to handle people with accents better.

2. A way to input commonly-used acronyms and phrases so that the AI can learn them and favor them for your transcripts. Our industry uses a very large number of acronyms and unique jargon. The automatic transcription does horribly with acronyms. If I could add these to a dictionary or list for the automatic transcription to reference and favor when it thinks these words are said, it would be helpful. We use search and replace right now, but the system very often uses many different alternative words/phrases inconsistently instead of the correct one or just one consistent alternative, which makes it difficult to find all incorrect instances.

3. The ability to export transcripts without any timestamps. We do not need the timestamps for our transcripts. It is a pain to manually remove them all.

4. The ability to import untimed, plain-text transcripts and have Premiere automatically determine the timings and generate captions. Right now, we can only generate captions using the automatic transcription. If we receive a transcript that was already made by another department, however, we cannot just import it and have Premiere automatically add the timings. We are often given transcripts in plain text without the timings. We are not given SRT files. I know that YouTube is able to add timings to plain text, so this feature should, theoretically, be possible.

5. More accurate caption length and line numbers. Even though we adjust the settings for maximum number of characters, time, and number of lines, these settings are not applied accurately. For example, we frequently will end up with captions that include just one word and last one second even though we had a large duration and character limit. This seems to be some sort of bug in the way the settings are applied.

6. The ability to export the captions as a VTT file. We do not use SRT in our industry. We use VTT. I do not want to have to use a third-party tool to convert an SRT to a VTT.

We are really enjoying the text-based editing features. Improving the automatic transcription and captioning would make this process easier and even more useful for us. Thank you!