-
Speech Analysis On Premiere CC 2017 Workaround
Hey guys!
A couple weeks ago I bumped into a post issue for a documentary I’ve been assisting/editing. The whole time I’d been working with it I’d been using timecode from transcripts to find bites – this is one of those 160 hours of footage with 4-5 hour interviews that carry a lot of heavy, long-winded text bites. An editor just joined the project and he was used to using speech analysis in AVID but we’re cutting on Premiere. He mentioned Premiere used to have a speech analysis feature that I couldn’t find in 2017, so after a lot of research and talking to the support team at Adobe, I finally managed to use speech analysis again and have it even work on CC 2017, I know there’s a whole lot of people out there struggling with the same issue so here’s how I did it in case it helps anyone else.
First, I downloaded Premiere 2014, the last version that had the speech analysis feature. You can find this under the ”Previous Versions” section in your creative cloud drop down menu:
(Premiere 2014 isn’t shown on my screenshot because I already have it installed)
When you install this version it doesn’t remove which ever newer version you are using and I found out you can actually run both versions at the same time! Which you will need for this.
As some of you know, you can use the video file to attach the script but if you’re working with long interviews like me, you probably want to stick to the wave files. To find these I unlinked the interview files and selected reveal in finder for the interviewee mic:
Then I imported the wave files I needed to analyze with speech analysis and downloaded my transcripts from google docs as plain text. I ran the program and it send it to Media Encoder 2014 which downloaded automatically with Premiere.
After it was done encoding, it was done! The script was attached to my wave file’s metadata and it automatically showed up on Premiere 2017 even though it doesn’t have the feature anymore. You don’t need to re-import anything or even online/offline the file, it just pops up:
Pretty cool huh?
Karina
-
6 Replies
-
Very nice!!!
Even though the feature isn’t currently implemented…the XMP metadata is still supported…and this what allows your workflow to work!!!
Thanks for the detailed step by step!!!
Alex Udell
Editing, Motion Graphics, and Visual FX
Let’s Connect on Linkedin
Examples: Retail Automotive Motion Graphics Spots
Example: Customer Facing Explainer Video
Example: Infotainment & Package editorial -
Thank you so much for this! It is exactly the fix I’m looking for. 🙂
Peace & respect,
Joy E. Reed
http://www.ohmyproductions.net -
Hopefully in the very near future, this workaround won’t be necessary due to the impending arrival of this premiere plugin:
https://www.digitalanarchy.com/demos/transcriptive.html
Adobe Premiere 2017.1
Windows 10 Pro
Samsung SSD 850 EVO system
Samsung SSD 850 EVO Adobe cache
NVIDIA GeForce GTX 970
OWC Thunderbay 12t x 2 in Raid10 configuration (thru Storage Spaces and Disk Management) -
I can confirm Transcriptive works like a charm. I’ve only tested it with Speechmatics as the transcription engine, but it’s fast and accuracte.
-
Hello!
I’m trying to bring that workflow back as well… in order to create subtitles within AE.
I ran the analysis in CS5, using a WAV file and an exact matching text file.
It worked, kinda, but I still have 3 things I can’t figure out:1. After analysis is done, most of the words are getting small caps, and all the punctuation marks aren’t showing. And that is even though they have the right cases in my text file (Caps at beginning of sentences etc). Can I make the analysis get those right cases (caps, small letters, punctuation marks…) from my text file?
Is that text-file-encoding related? I tried switching and resaving the text from Unicode to UTF8 but haven’t seen any change. And the ANSI encoding (text file default) is actually cutting 3/4 of my text, so ANSI isn’t an option.2. After an analysis is made, and the speech is being embedded in a file (my WAV at this case) – can I reset the captions info/delete the whole captions info? Such as that I’ll be able to run analysis on the same WAV from scratch, w/o taking into consideration anything that was there from earlier analysis processes?
3. There are some words that I need to change/add manually. That’s fine, however after closing the premiere project, even if I saved it – my changes and additions aren’t saved into the WAV ☹
so – how can I make changes that will be KEPT within my WAV file and will show up next time it’s used?Any info would help. Thanks!!
Ori Kuper. -
None of this effort is necessary anymore due to the AI speech analysis options that are now available online. Transcriptive, Temi and other services are so good that there’s no need to use these old methods.
https://www.linkedin.com/in/tmprods
tallmanproductions.net
Reply to this Discussion! Login or Sign Up