#
2024
#
October
#
31st October 2024
Waltzing With Brando will premiere at the Torino Film Festival!
Cut in Final Cut Pro X by the incredible and magical Michael Yanovich, ACE.
You can read more here.
finalcutgod has released a new YouTube video talking about captions in Final Cut Pro:
Caption Pop AI by StupidRaisins is out now for USD$34 for a limited time!
The website says:
Introducing Caption Pop AI – the fastest way to make captions right inside Final Cut Pro! Forget about exporting to CapCut or waiting on other slow, clunky software. With Caption Pop AI, you can automatically create captions in just 3 easy steps: analyze your video’s audio, customize your captions, and drop them onto your video. Done!
Whether you're editing YouTube content, social media clips, or Caption Pop AI demos, Caption Pop AI delivers professional results in record time. Customize the look, enjoy accurate AI transcriptions, and work in 17 languages – all without leaving Final Cut Pro.
Get Caption Pop AI today and keep your workflow smooth and speedy!
You can watch an intro video on YouTube here:
You can download and learn more here.
What I didn't realise until now is that StupidRaisins has basically built their own versions of FxFactory - their own app, with their own marketplace - very impressive!
I've already downloaded Caption Pop AI and taken it for a quick test drive...
My first comment is I fricken LOVE the Stupid Raisins DMG!
My second comment, is their developers clearly haven't read my FCP Cafe documentation about why you should use greyscale icons for Workflow Extensions!
They also haven't set the minimum sizes correctly for their Workflow Extension:
Unlike MotionVFX's mCaptionsAI it seems like Caption Pop AI only support importing audio files:
After dragging in a WAV or AIFF you can select local or cloud processing and the model you want to use:
If you selected local, it then downloads the model:
It then processes the audio content:
Once complete, you have a lot of controls to play with:
To send back to Final Cut Pro, it's a bit more clunky than MotionVFX's mCaptionsAI - if you press Open in FCP, it will basically send a FCPXML across - prompting you to select a Library in which you want to import the FCPXML.
It then brings in a new Project with just the titles, so you need to copy and paste back into your original project/timeline:
Overall it does a pretty good job, using the recent Mac mini Announcement video as a test:
There is some maths errors/funkiness with timings, for example:
Overall though, it seems like a very useful tool to have in your toolbox, especially if you don't want to pay MotionVFX a subscription.
It'll be super interesting to see how well Apple's new caption tool compares to OpenAI's Whisper, which is what all these tools use.
It seems like Caption Pop AI is using whisper.cpp under the hood.
It stores all the models, temporary audio files, and FCPXML exports in '~/Library/Containers/com.urlaunched.StupidRaisins.StupidRaisins/Data/Library/Application Support/com.urlaunched.StupidRaisins.StupidRaisins:
If the Apple speech-to-text model is better than Whisper (which is unlikely), then I'd imagine a lot of these developers will start offering a FCPXML import option.
I think though, really the main benefit Apple has is that they can do the timing and handle timecode much better than Whisper can currently do - this might be the biggest selling point for Apple's tool.
Overall, it's great to see another non-subscription tool on the market!
Audio Design Desk v2.1.3 is out now!
It includes the following changes:
- Track automation wasn't blocking mouse events on regions beneath, so it made it impossible to drag automation points
- More region mouse event improvements and fixes
- Improvements and fixes to Audio Unit manager UI for enabling and disabling Audio Units
- Epidemic login button would be disabled if you canceled out of the web login forcing an ADD restart to reenable it
- Audio tables wasn't updating with correct count after searching
- Dragging multiple files from the library tables would result in overlapped regions in the timeline. Now, they will stack by default on new tracks, but you can hold the shift key to have them placed sequentially in a row
- Misc UI fixes
You can download and learn more here.