LocalVocal: Local Live Captions & Translation On-the-Go

LocalVocal: Local Live Captions & Translation On-the-Go v0.5.3

Go to download
Overview Updates (29) Reviews (11) History Discussion
MoChuang said:
I found the problem for me. This domain no longer exists https://ggml.ggerganov.com/ which is where the plugin is trying to download the models from. But if you use the other link for models listed in the github page (https://huggingface.co/ggerganov/whisper.cpp/tree/main) then you can find the models and download them and either name the folders correctly from the json file or just select your own model in the plugin.
Click to expand...

Hey, was wondering how exactly you managed to do this. Would love to get this plugin working since I'm a shyer streamer who doesn't like speaking with my own voice
 
royshilkrot updated LocalVocal: Local Live Captions & Translation On-the-Go with a new update entry:

v0.4.3

What's Changed​

  • Downgrade ONNX runtime to 1.20.1 for MacOS 12 compatibility by @Tabby in #249
github.com

Release OBS Localvocal 0.4.3 · royshil/obs-localvocal

What's Changed Downgrade ONNX runtime to 1.20.1 for MacOS 12 compatibility by @Tabby in #249 Full Changelog: 0.4.2...0.4.3 Checksums obs-localvocal-0.4.3-macos-arm64.pkg: 6c979aa6b04caf9afa057e3b...
github.com github.com
Click to expand...

Read the rest of this update entry...
 
royshilkrot updated LocalVocal: Local Live Captions & Translation On-the-Go with a new update entry:

v0.5.1

Small re-release to fix missing installers from the last release

What's Changed​

Download:
github.com

Release OBS Localvocal 0.5.1 · royshil/obs-localvocal

Small re-release to fix missing installers from the last release What's Changed Fix release artifact renaming by @Tabby in #258 Full Changelog: 0.5.0...0.5.1 Checksums obs-localvocal-0.5.1-macos-...
github.com github.com
Click to expand...

Read the rest of this update entry...
 
royshilkrot updated LocalVocal: Local Live Captions & Translation On-the-Go with a new update entry:

v0.5.3

What's Changed​

  • Update whisper dependencies to 0.0.12 by @Tabby in #264
    • Updates Vulkan SDK on Windows to 1.4.328.1
      • Adds more Vulkan operations that Whisper.cpp can make use of which should improve Vulkan performance
    • Build ROCm/HIPBLAS Whisper.cpp backend for all available GPU architectures available in ROCm instead of whatever the...
Click to expand...

Read the rest of this update entry...
 
Very nice work. But why have you omitted the Greek language in Transcription models?
It is inconvenient to have to download and point to an external model.

When we see that you have even included languages spoken by far fewer people/smaller countries, but Greek is left out, we can only begin wondering...

Hopefully, you will address this in one of your next updates.
 
Hello. Can you help me with the installation?
I downloaded the installer and installed the plugins. At first, OBS showed that the plugin didn't exist. Now it's listed as active in the plugin list, but I can't set the necessary filters—they just don't appear in the filter list. Perhaps I installed it incorrectly?
 
Lorandicus said:
Hello. Can you help me with the installation?
I downloaded the installer and installed the plugins. At first, OBS showed that the plugin didn't exist. Now it's listed as active in the plugin list, but I can't set the necessary filters—they just don't appear in the filter list. Perhaps I installed it incorrectly?
Click to expand...
Hello. You can't find it among the filters but you have to create it via sources by clicking on +
 
I was there was a new version on Github v0.60. When I go thru the installation video the screens are much different including the title of the filter, did someone put a fake version up? If not, are there any new instructions as I'm not able to get the transcriptions to display. Thanks. Any special instructions for using this plugin with the altium vertical plugin?
 
royshilkrot said:
Actually you can stream the subs to the 608/708 CC channel on yt or twitch or any other platform by enabling "send captions to stream" in the settings. You can also have more control by sending the captions directly using the POST endpoint: POST YouTube Captions: OBS w URL/API Source [Tutorial]

- YouTube

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.
youtu.be youtu.be
Click to expand...
Love love love this plugin and appreciate all the work @royshilkrot has put into it! I do have a question based on this past post - I've just installed 0.6.1 and I do not see a "send captions to stream" option. There is a "WebVTT" option, but I specifically need 608 captions delivery. Is that still a part of this plugin or is there a way I can add that? Thank you!
 
On Linux and Mac the text sources are different than on Windows (Freetype2 instead of GDI+) and lacking important formatting options. I can not confine the captions to a defined area on the screen. The size and shape in changing all the time, sometimes even outside the canvas. Does someone have a solution for that issue? I don't like to haul my huge win tower plus one more monitor to the live venue.
 
Loving what LocalVocal has accomplished so far, but I've hit one snag I'm trying to diagnose. I'm using the write to file option and writing to a local file. I'm then using a python script to grab that text file, format it for a caption encoder, and then telnet that data to my caption encoder for insertion onto my SDI signal. The problem is that LocalVocal isn't consistently writing to the text file. Sometimes I get a couple of words, sometimes I get a whole paragraph at once. I tried enabling the buffered output setting, setting it to character with two lines and 30 characters per line. That didn't seem to do anything to my file. I then tried changing the Sentence Duration under advanced settings from 7000 to 4000, but that didn't seem to affect my file either. What settings do I need to tweak to get LocalVocal to output at a more consistent rate to the file? When it's sending quickly, it's displaying on my caption encoder plenty fast, but when it dumps a paragraph at a time to the file, it results in both the captions being delayed, but also the captions going by so quickly they're impossible to read as the caption encoder tries to display the whole paragraph. Again, fantastic plug-in and hopefully it's a quick tweak that I'm missing to make the file output more consistent.
 
You must log in or register to reply here.
Back
Top