OK, so if you are a Zoom disciple, you can see virtual backgrounds and they even have Zoom apps that let you use anything from Pexels. In addition, they have super good keyclick suppression. So if you are using Slack Video or Google Meets, then you are really in trouble if you like typing while listening to a video.
So what are the choices to get you out of this mess:
- Open Broadcast Studio. This works really well for Zoom, Google Meets and things that are not that picky about a virtual camera device. It also burns CPU like crazy (unless you are using an MacBook Pro M1 Pro/Max) then it is not an issue. You can create a green screen video simply by adding a Video Capture Device in Sources and then right clicking to Filter and add Chroma Key filter. This is not as good as Zoom which does this with ordinary backgrounds but nothing is perfect.
- Green Screen and Webcam Lighting. At that point you can use a green screen (they are $29 now for a chair version. The other things you need are light that works there typically a good for folks.
- Background Videos, you then can add Media Source and find a video then you can arrange it well so it looks nice on the screen.
- Background Removal Plugin. This is a v0.3 version works pretty well, what it does is to try to guess what is in the background and add a green color to it. So if you don't have a green screen, you load it as an AV filter and then select green and Selfie Segmentation and it uses a neural network model to guess where you are.
How to get audio click suppression with Krisp (paid) or RNNoise (free)
If you also want to get click suppression and other audio treats, then you can:
- Use OBS NDI which is the Network Device Interface, this bottles up OBS with video *and* audio input and puts it out as a virtual camera and audio.
- OBS Noise Gate and Noise Suppression. The Noise suppression uses RNNoise (see below and they have the VST 2.x plugin running)
- Krisp.ai charges $5/month (I know right), but it is implemented as a filter that creates a virtual audio device.
- nVidia Voice RTX if you are on Windows and an old Mac with nVidia graphics cards.
- The no cost option is a VST 2.x plugin called byu @werman which uses an RNN to suppress noice (and needs a bunch of magic VST SDK files). You do need an underlying production system that gets these plug-ins. Reaper is one that is mentioned. But you can get shims from Sailpdt which appear to allow this plugin to work with Apple Logic Pro and there is a OBS Plugin that appears to allow this as well so when you open a source, you can select VST 2.x Plug-in and it has a sleect plug in dialog. You need to put the plugins into magic locations which on the Mac is ~/Library/Audio/Plug-ins/VST/ which is sure not obvious and I could not get this to compile either, although someone did build a bundle for it at
How to get video and audio from OBS via NDI
If you want to do audio processing, you can use OBS using the NDI or Network Device Interface, this creates yet another virtual device called an NDI, but this time it bundles the audio processing of OBS with the image itself. With NDI, you are actually transmitting the entire OBS output over NDI to the network. And it creates a virtual audio *and* video device.
The base virtual camera in OBS 27.0 and higher just do lack video. So do this:
- Go to the NDI Video site and click on download and you will get an email with the magic link.
- There you will get a package which you can load and then you restart your computer
- Now you have to start "NDI Video Input" and you probably want to make it a startup item.
- Go to the OBS and install the OBS NDI tools with brew install obs-ndi and then in OBS go to Tools/NDI Output Settings and you see see the Main Output and click on this.
- Now OBS will always output to the NDI Video and Audio.
- In your applications like Zoom, you will see a new virtual camera called NDI Video and a new virtual sound called NDI Audio and this should now be the output.
All of this chews up way more CPU. In fact on an M1 Max, I'm now using 35% of the CPU because the video goes from the camera to the OBS to NDI and then to Zoom, so you are using lots of CPU, but if you want audio processing, this is the way to go.
OBS does not work with Slack, but you can use Krisp.ai (paid)
Note that while the video doesn't work on Slack, the audio does, so at least you can get that processing. This works for krisp as well so if you are willing to pay $5/month, you can get that keyboard suppression software.