Ok, I'm kind of going crazy tryingto figure this out since no one seems to use this feature.
I don't need it for the accuracy, but I'd like to upload transcriptions from long interviews so I can then run analyze and marry the text to the speech so a producer can easily cut from the text.
Every FAQ and tutorial is clear and it looks very easy. I open up any kind of clip with VERY clean audio then click the "analyze" button in the metadata panel and and media encoder opens and works for a bit.
when it's done I go back and the field is ALWAYS EMPTY. I tried it on a friends near identical computer (2011 macbook pro) and it works fine, everything the same. No matter what I try mine comes up blank like some setting is wrong or 'off' or something. it does some work but acts like there is no audio or something. Can anyone PLEASE help me figure out what is happening, I'm stumped. I'm fine with BAD transcription but showing nothing tells me something isn't working right, and I can't get it. appreciate any constructive help or ideas.
unfortunately I have read all of this and many other tutorials outside of Adobe and cannot find anything amiss from my process. Really frustrating. As i said some mystery setting in Premiere or my machine is wrong so it 'thinks' it's working but always puts out nothing no matter what kind of video file, what kind of audio quality, identical reference script, etc. I have a very powerful machine with all legit programs. Thats why I'm extremely frustrated with it.
ok, thanks for the suggestions, I finally figured it out, and it was my original hunch.
It seems to be the MTS files. Even ones that are now proress transcode and some various other ones that are quicktime outputs from MTS. It really hates MTS (and so do I).
The other gripes I found about MTS seem to indicate that the 'analyze' button is greyed out if it's out of it's folder structure, but I never had that problem it just tried anyway and came up empty.
Also if anyone was interested in trying this feature - don't. It's really not useful enough yet. Even with a near perfect transcription added it's still way off, and even changes some of the pre-transcribed words back into nonsense. After all that dividing and re-joining clips I still have a trash text transcription thats not clear enough for someone to select from, in a cumbersome fashion that once perfected might save me 10 minutes.
I thought even after doing my own transcription it would help to have it married to the timecode, but it didn't save anything and muddled what I had already done.
For fiction through "Story", maybe. But Pre-transcribed documentary, still a hassle. One day it'll be great in a few years.
Thanks for your feedback. I'm sorry it didn't work out for your project. Speech to text is a very difficult thing to pull off, and we are constantly working to make it better. For sure, it's not for everyone at this point in time.
I have the same problem....The field of Metada is allways empty! No text!!! Algouth Media Encoder seems to work and "finish" its work making a sound and a "green mark" but in Premiere NO TEXT. .I don't undertand what to do about MTS files of Renatus1234565??? whats that? any idea?
Kevin... one *big* request I have for this feature. Please can Premiere *not* change text if a script has been provided. Make it a different color if it can't recognize the words but please, please keep the text the same!
If I go through all the trouble of creating a transcript and import it (or use the actual script of a play), the last thing I want is fo Premiere to screw it up! I like the fact that I can go in to the transcript and fix words Premiere didn't get right but I really hate having to "fix" things that were right in the first place!
one *big* request I have for this feature.
Europe, Middle East and Africa