![]() ![]() You can download the latest version of Discord Canary for macOS here. I’m curious if there are any notable improvements to the platform through the Canary version. Let us know down in the comments how the native app runs for you. Its cross-platform capabilities work across iOS, Mac, Windows, Android, Linux, etc. The chip lineup includes M1, M1 Pro, M1 Max, and the all-new M1 Ultra.ĭiscord is an Electron-based app and should run more smoothly through its Canary version. Currently, all of Apple’s Mac devices run on Apple Silicon, with the exception of the Mac Pro. Many users will be happy the app no longer has to rely on Rosetta 2 – this technology translates apps designed for Intel-based Macs to work on Apple Silicon. Popular with gamers, users can communicate through private channels across most of their devices, even Xbox. This means the Discord Canary app can properly support Macs using the M1 chip or later.ĭiscord is a communications platform where users can create groups to chat through text, audio, and video. According to Reddit users, the digital messaging app has now exited its four-month long beta. Developers of software such as Slack and. ![]() Our architect, Stepan, together with the team, had made decent progress there and this progress came amazingly timely.Back in November, Discord entered beta testing for its macOS app, a version called Discord Canary, on Apple Silicon. The Electron framework that lets web developers more easily develop cross-platform apps has now been updated to work on the new Apple Silicon M1 Macs. We had an ongoing project where the goal was to reduce the size/speed of our DNN so that it can be embedded in more constrained environments. The miracle came from our incredible research team. To be clear, this was simply impossible for us. In practice this would mean dropped packets and broken voice. If you don’t process within 3ms Chrome will drop the audio packet and give you the next one. So we had to bufferize 10 frames (30ms) and then process the 30ms frame with our DNN within 3ms. The way Chrome audio filters are designed is extremely constrained.Ĭhrome feeds the filter plugin with 3ms frames, while it reads the data from the mic, and the plugin has <3ms time to complete the processing. Good engineering, Mozilla, Facebook and Google, as always ?īut apparently this was just the beginning of our journey. To be honest we were mindblown at this point. A version of Discord with native support for Apple silicon Macs is now rolling out to users, almost a year and a half after the first Macs with Apple silicon chips were shipped to customers. The result? Almost the same speed as the C++ version. After some adjustment to the build system we were able to build our code with XNNPACK for WASM. It’s a highly optimized library of floating-point neural network inference operators for WASM (and ARM, x86). So the algorithm was running 10x slower than our main C++ model on the same laptop.Īfter some research we found about Tensorflow.js and all the awesome work that Google is doing to bring DNNs to the browser. It was a hacked-together version which implemented matrix multiplication in the most naive way, in Javascript. His first version of the port came after 3 days. When we have no idea about something – we always look at Artak ? So Artak, one of our architects, started looking into WASM. How do we port our C++ codebase into browser? Well, obviously WebAssembly (WASM) but we didn’t have anyone with such experience. We set out to build Krisp for Chrome and even came up with a cool name for it – KrispX (X for extension). The filter was working fine for Google Meet, Webex and other apps. We ❤️ Toptal.Ī month after that dinner we had a POC of an extension which was able to add an audio filter to a webpage that transforms the microphone stream in real-time by adding a static noise. I went straight to Toptal to find someone who could quickly hack a proof of concept (POC) for us. I was so skeptical that didn’t want to engage our core team yet. ![]()
0 Comments
Leave a Reply. |