VLC is a very well-known open-source media player. I’m sure many people have used it—my computer has it, for instance. It supports almost all audio and video formats and codecs, and it’s really quite handy.
Although it’s open source and free to use, I didn’t expect the development team to actually rent a booth at CES 2025 to exhibit, showcasing the latest AI auto-captioning and translation features. This means that in the future, videos without subtitles won’t be a problem—you can just generate them with AI, or even translate them into other languages.
At CES 2025, VLC showcases new AI-powered auto-generated subtitles and translation features.
Earlier, the VLC team shared a short video on X.com showing their latest AI-powered auto-generated subtitle and translation features at their CES2025 booth.
According to the introduction, this new AI feature is based on a local open-source AI model, meaning it can operate offline without needing the internet or cloud, and it supports multiple languages, expected to reach over 100.
Unfortunately, this video doesn’t directly show the subtitle generation and translation process, so it’s unclear how fast it is.
VLC automatic subtitles generation and translation based on local and open source AI models running on your machine working offline, and supporting numerous languages!
Demo can be found on our #CES2025 booth in Eureka Park. pic.twitter.com/UVmgT6K4ds— VideoLAN (@videolan) January 8, 2025
With this feature, users no longer need to worry about subtitle files when watching videos; they can just generate them via AI whenever needed.
VLC CMO Natacha Holtzhausser said: “Generated subtitles and translations will open up language-accessible accessibility for downloaded movies and TV shows, lectures, and live broadcasts.”
As for which AI model is used, VLC did not disclose, but Holtzhausser later added, “We will provide more details about how this technology works in an upcoming blog post.”
We are so thrilled to unveil the first of our tree game changing innovations this week in @CES : Automatic subtitles and translation in real time powered by #AI models running locally on your machine.
We will give you more details about how all of this works in a next blogpost… https://t.co/4liAyheOfm
— Natacha Holtzhausser (@NatachaHoltz) January 8, 2025
Regarding the AI transcription subtitle feature, it had actually been in development for some time. Initially, it was done through an extension using OpenAI’s speech recognition model, Whisper. However, this new demo appears to embed the AI model directly into the VLC software itself, enabling real-time translation subtitle generation.
Whisper’s translation feature currently only supports translating other languages into English, so the AI model it uses is probably a different one.
Although VLC’s built-in AI auto-generated subtitles and translation features are quite good, generating subtitles is no problem—nowadays you don’t necessarily need a large model, and accuracy is high too—but translation is hard to say.
For AI translation to be good, the model can’t be too small. The translation quality of current small models is terrible, and I think that’s a major challenge. After all, when running locally, the larger the model, the higher the demands on the computer.
So I think we shouldn’t expect too much from the AI real-time translation feature.
As for when it will be released? That’s also unclear; VLC hasn’t announced anything.
In addition to new features, this week VLC also celebrates surpassing 6 billion downloads:
For those interested, you can keep an eye on the official latest news and Natacha Holtzhausser’s community; barring any surprises, more details should emerge after CES 2025 concludes.
As AI develops rapidly, it’s only a matter of time before more players incorporate similar AI features, so readers using other players can look forward to it too. That said, I think running AI models over the network is the best solution right now, rather than locally.
Source: KOCPC Chinese