We have already spoken on previous occasions about NVIDIA’s interest in artificial intelligence, and although NVIDIA Broadcast may seem to have no relationship with these interests, the truth is that it is just the opposite. This application may not be as striking as what we see in other practical AI applications, but on the contrary, and making use of it, it offers us an incredibly practical application that has a place in the daily lives of many of us.
In case you don’t know it, NVIDIA Broadcast is an application that is compatible with any PC equipped with an RTX 2060 graphics card (desktop or laptop) or later and at least eight gigabytes of RAM, and that allows us to apply, in real time, a series of effects to our video (camera) and audio (microphone) output signals, as well as what we hear through the speakers or headphones. In this article we tell you what it is and how it works, and here you can see how to integrate it into OBS Studio.
As we already told you here, when NVIDIA introduced RTX Voice, the effects offered by NVIDIA Broadcast take advantage of the capabilities of its graphics cards in terms of artificial intelligence (in this regard, GPUs are much more efficient than CPUs, due to their high calculation capacity of floating point operations). In this way, and after previous training, these simple but effective solutions offer more than adequate results.
Recently, the company has published version 1.4 of the application, which you can download here and, in addition to some performance improvements in it, the great novelty is the Eye Contact function that, as you can see in the video above, it will simulate that you keep looking at the camera when you are not doing so. In this way, either for a moment and carelessness, or permanently because you are reviewing something on the screen or in a physical document or on another device, your interlocutors will not be aware of it.
We must bear in mind, of course, that this function is currently in beta phase, so its operation may experience some failure, since there are millions of possible lighting and eye color scenarios that the company has not been able to test for Check if the algorithm works correctly. However, and especially if users share their experiences with technology, it is possible that in a short time the results can be completely reliable and that, in this way, we no longer have to choose between looking at the camera or reviewing the notes we had made. prepared for that meeting.