Nvidia Broadcast has been with us for some time now, updating and adding more and more features to make streaming or recording content creators easier and with better quality. We will no longer need the huge chroma behind to eliminate the background, equalizers or channel mixers or soundproofing panels, all this is capable of being done by this small application.
Index of contents
- What is Nvidia Broadcast
- Nvidia Broadcast Settings
- Microphone settings and options
- speaker output
- camera settings
- Integrate everything with OBS
- conclusions
The world of streaming is undoubtedly growing faster and faster, and with it the demands of viewers to receive higher quality and more elaborate entertainment. Thanks to Nvidia RTX cards and in particular their Tensor Cores , these filtering and image processing functions are within the reach of a single click on both desktop and laptop computers. In fact, we have used an AORUS 15 2023 with Nvidia RTX 4070 for these tests, and you will see that it works great even using the device’s battery mode.
What is Nvidia Broadcast
NVIDIA Broadcast is a free application from Nvidia that offers an enhanced live streaming experience for content creators , gamers, and general users. It is designed to improve the audio quality, both input and output, and video of live broadcasts.
To do this, it uses Artificial Intelligence- based technologies that run with the help of the Tensor Cores of the Nvidia RTX 20, 30 and 40 series cards to eliminate background noise, improve voice quality and create visual effects in real time . At the time of this article, it is running at version 1.4.0, offering support for all types of Nvidia cards, including RTX Laptop versions . Since the graphics card is mostly used, it does not place an additional load on the CPU.
These are the improvements that we can implement in our transmission:
- Background noise suppression: Background noise cancellation technology uses AI algorithms to suppress the sound of a fan, PC or street noise, so that the user’s voice is clear and crisp. We don’t need to implement any type of sample to cancel the noise, it is done automatically by the application.
- Bokeh Background– NVIDIA Broadcast also offers the option to blur the background of the image to keep the focus on the user and hide. This is especially useful for streamers without the need to use a DSLR camera.
- Improved audio quality: Echo cancellation and voice enhancement technology improves audio quality to make speech sound more natural and pleasing, with two noise stages to perfectly match the output.
- Real-time Visual Effects– NVIDIA Broadcast offers a variety of real-time visual effects, such as virtual backgrounds, camera filters, and face-tracking effects, so users can personalize and animate their live broadcasts.
All this ensuring compatibility with different live broadcast applications such as OBS, Streamlabs or Discord , which makes it easy to use for any user. We only have to select Nvidia Broadcast as the source of resources in the different scenes that we create and the effects will be applied directly to the transmission.
Nvidia Broadcast Settings
We are now getting into the matter, and the first factor to take into account is of course what equipment we should or can use. Well this is easy; everyone who has an Nvidia RTX card from the 20 series , either in desktop or laptop version.
To check how well it works, we will use an AORUS 15 laptop with Intel Core i7-13700H CPU and Nvidia RTX 4070 Laptop GPU, using battery resources , and therefore limiting its power a bit to see if everything runs smoothly on the application . The transmission equipment consists of a Razer Kiyo 1920x1080p webcam at 30 FPS and a Razer Seiren V2 X microphone that the guys from Nvidia have lent us for the test.
Before using the application we should connect the devices and install their drivers, in our case it is necessary as they are USB peripherals that have some management capacity from Razer Synapse
We will download the application from the official Nvidia page and install it. The application consists of a small window with 3 sections focused on Microphone (audio input), Speakers (audio output) and Camera (video input), something like a simplified mixing for streaming and videos.
Microphone settings and options
We start with the microphone section, where we will select the Seiren X as the audio input source from the drop-down list. At this moment the effects available for it will be activated and on the right a test area where we can record our voice and check how well it is heard. Two effects can be applied simultaneously, which today are the two that are available.
Audioplayer
00:00
00:00
Use the up/down arrow keys to increase or decrease the volume.
On the one hand we have noise removal , which allows us to filter the ambient noise with respect to our voice through AI and internal patterns that it has learned to give an input that is as clean as possible. In our case we have a fan on top and the fans of the laptop itself, and this mode really manages to remove all that noise without distorting the voice . In the previous clip, you will notice the difference if you turn your headphones/speakers up enough volume.
The second filter is to eliminate the echo in the room , well, it is clear, the idea is to detect echo effects and eliminate them in real time. The echo can be generated when we have a large empty habituation or when we speak very loudly. It is not our case, so the difference between enabling or not that effect has not given us differences.
speaker output
The second section of Nvidia Broadcast is focused on the output, or rather, the return of the audio, and again the program will give us a drop-down list with the options it has found, whether they are speakers or headphones . On the right side we can listen to test clips to see what the effect of the filters is.
Again, two types of effects can be applied to the output; Elimination of noise in case we have fans or air conditioning above us, and elimination of the echo effect in the room. Basically the same as before, but to listen to a clean return, and the truth is that it does an excellent job at least with the tests we have done through headphones.
camera settings
We are entering the section where we will have more room for customization , that of the camera, and the work it does with webcams in Full HD resolution is impressive, imagine with a 4K like the Razer Kiyo Pro Ultra.
We start by selecting the device in the first section and we will immediately see the image -almost without delay- in the right area. If we click on the configuration icon next to the drop-down menu, we will access a wide list of properties where we can modify brightness, sharpness, gain, zoom, focus, exposure, etc. all those that are admitted, of course. In our case we will leave everything automatic.
In this section we can apply only two filters at a time , it’s not much, really, but we can make many interesting combinations. It allows to capture an image and flip the scene if we want the mirror mode. Let’s see all of them:
Background blur and vignette effect
Through the Tensor Core, Nvidia Broadcast detects the outline of the image in the foreground , it can be said that with great precision, and generates a bokeh effect in the background whose blur we can modify. Selecting the performance mode will give more FPS or fluidity when possible in the scene, and with the quality mode it will improve the precision in the crop. The vignette effect simply adds a black effect around us, it is of little use.
If we look closely, in the second image there is a small region between the headband of the headphones and our head that remains unblurred. Except for this small detail that certainly needs to be improved, the effect is perfect and the application in real time is impeccable . We will avoid the commented detail by using headphones that do not generate gaps in the image. As for the focus and quality of the foreground, this will depend on the features of the camera itself, which in this case has room for improvement.
Background replacement and background removal
In the same way that it can generate a blur effect, the application can also apply a background behind it , cutting the image as if we had a chroma. In this case, it has also managed to fill in the small gaps generated by the headphones, giving a solid image even in motion, much better than a chroma and without the need for a high quality of lighting in front of us.
The same as before, but leaving the background completely empty so that another layer of the scene can be seen behind us. This can be perfect for broadcasting commented gameplay in the old fashioned way, without frames, giving the game behind us as much prominence as possible.
noise removal
In case of having a bad camera or poor lighting in the setup, the application will detect an area where noise is generated to try to eliminate it. This is usually a common problem with webcams, so we will have strong and weak modes depending on how comfortable our scene looks. We have turned off the lights and with the help of the filter the image capture has been of excellent quality.
Eye contact
This effect makes the AI detect our pupils and redirect their movement in such a way that they are always facing the camera . Obviously it will act on slight movements that are more or less in the field of vision, because if we look completely to the sides, up or down, then it will respect the orientation so as not to make it “unreal”. Even though it looks like I’m looking at the camera, I’m really looking at the screen and even up, but the filter redirects the pupils incredibly accurately.
What use can this have? Well, the most obvious will be that it is not noticeable when we are using a telepompter in video recording. In streaming, just chatting or gameplay we should not use this filter because it is too artificial, since we are not always going to be staring at the camera because it is unnatural.
auto framing
Finally we have the filter that perhaps gives the most striking results when making gameplays and video reactions, it is none other than the automatic focus. If we zoom in to center our face in the scene, the app will follow us wherever we move within the camera’s maximum focal range. This intelligent tracking can work with a second filter, and although it has a slight response time , the behavior at the level of fluidity and definition is excellent.
Integrate everything with OBS
Above Nvidia de Broadcast we must place the application that we use to stream or record videos, which can perfectly be OBS. The integration procedure will be most intuitive, since this application will appear as another resource in the program .
We create our scene as we have always done, and we add the different video and sound sources that we deem appropriate. When we choose “ Audio input capture ”, “ Video capture device ” or “ Audio output capture ” we will select Nvidia Broadcast from the list, in this way the final output resource will have all the filters and effects active in the application .
conclusions
Nvidia Broadcast makes certain common sections of streaming or recording much easier, and we find it very useful for all those who have Nvidia RTX cards, which are actually the majority of content creators. It’s light, it hardly consumes CPU resources and it integrates perfectly with streaming programs , you couldn’t ask for more.
The behavior that it has offered us with a webcam and microphone like these from Razer is wonderful despite being rather normal peripherals. We find the noise suppression in the microphone especially useful and effective, as well as the filters to remove the background, blur it and follow the movement in the camera. With them we save chromas, effects in the scene that loads the CPU more and even video post-editing .
