The Trailing Edge -- TTE.CA’s Post

I did another Game Audio Tech Tips video, this time on extending the functionality of Spatial Audio Volumes using an Actor Component to Start and Stop Wwise Events using either Overlaps or the Distance Attenuation (you can only normally Start sounds using Spatial Audio Volumes, which means they are just holding in virtual when outside of attenuation range) This is aimed at general learning and / or people or teams with lack of programmer support and gives them a means to expand upon the Spatial Audio tools to optimise CPU etc. https://lnkd.in/eS6MH8r7

With LM Studio, you can easily run local LLMs on your computers. Downloading powerful LLMs to your local computers and deploy them as your AI agents: - LLaMa - Mistral - Falcon #llms

A major update has been made to one of my primary modules in the studio  (Virtual & hardware)  This is one of the best midi sequencers available and it's worth keeping an eye on. https://lnkd.in/eEb-bd4Z (that one will add value to your midi) #prodigestudiolab #expandyourcreativity #madrid

This is a great UX-focused review of a Copilot + PC for someone like me who is on the fence about (a) whether to buy one (especially since the Recall feature was pulled) and (b) which brand/model to buy. As an unapologetic doodler, his discussion of the so-so stylus was especially interesting to me. Also: To tech vloggers out there. Please note two aspects of Colbow's style: (1) He does NOT have any loud annoying background music! There is no need for background music in a product review/demo (unless it is an audio product). (2) He does NOT yell at his audience. Unless you have a really poor microphone, there is not need to yell at your audience. We are not under a circus tent. <stepping off soapbox> https://lnkd.in/gEVUGhQ6

Using dynamic programming and NVIDIA GPUs, we present Turing Segment—an optimized version of Cellpose. It offers the same level of accuracy while cutting down processing time to just seconds and significantly reducing memory usage. Explore the release at: https://lnkd.in/gKFGmkqs #SpatialBiology #CellSegmentation #GPU

We're excited to announce Turing Segment - optimized version of Cellpose leveraging dynamic programming & NVIDIA GPUs to segment your cells in seconds, not hours. Same accuracy, much faster & with less memory usage. Explore & download: https://lnkd.in/gKFGmkqs #SpatialBiology #CellSegmentation #GPU

Hidden behind today's Advent calendar door is the practical use of text variables in SoundPLANnoise. Project and program information, as well as various calculation run information, can be used as text variables in the graphics for fast and error-resistant creation of maps. Text variables are also used pages in the header and footer of the Result Tables or Spreadsheet. #AdventCalendar #HolidayJoy #FestiveSurprises #CountdownToChristmas #SoundPLAN #noise #sound #acoustics #simulation #software #environment #SoundPLANnoise

We've built our Aurora5 MV-HEVC encoder, leveraging AI-driven video compression technologies. Moreover, through a collaboration between #Visionular and the HDR Nova team, we also contributed a patch to #MP4BOX: >> https://lnkd.in/gEZgfbVj, which allows one can re-encode spatial videos following these steps: (1) taking an Apple iPhone Pro captured .mov file, (2) obtaining the left and right view YUV file via MP4BOX, (3) encoding using the MV-HEVC #HTM reference encoder, or our Aurora5 MV-HEVC encoder, and (4) wrapping the encoded elementary bitstreams to .mov file, again via MP4BOX, ready for playback in Apple Vision Pro. The MP4BOX patch we submitted contains two key items:  (1) Parsing fix for the SPS, and (2) Embedding the 3D reference display info SEI message. Our Aurora5 MV-HEVC enables real-time encoding of spatial videos to significantly smaller sizes while maintaining high quality. We'd love to talk to anyone interested in exploring various spatial videos captured by an iPhone 15 Pro, experiencing different configurations of MV-HEVC encoding, and enjoying their diverse visual effects on an Apple Vision Pro device. Join us for further discussions!

𝐌𝐞𝐞𝐭 𝐭𝐡𝐞 𝐆𝐨𝐨𝐠𝐥𝐞 𝐏𝐢𝐱𝐞𝐥 9! 🎉 𝐖𝐢𝐭𝐡 𝐆𝐞𝐦𝐢𝐧𝐢 𝐀𝐈, Imagine having a phone that’s your creative partner, personal assistant, and privacy guardian all rolled into one. It’s like having a genius in your pocket—managing your day, snapping pro-level photos, and even letting you edit yourself into any scene. Pixel Studio is your new playground for creativity, right on your device. And don't worry, the Tensor G4 chip keeps everything running smoothly while protecting your data. 𝐑𝐞𝐚𝐝𝐲 𝐭𝐨 𝐬𝐞𝐞 𝐰𝐡𝐚𝐭 𝐬𝐦𝐚𝐫𝐭 𝐫𝐞𝐚𝐥𝐥𝐲 𝐥𝐨𝐨𝐤𝐬 𝐥𝐢𝐤𝐞? 🚀 #GooglePixel9 #AIRevolution #CreativityUnlocked #NextLevelTech

It is amazing how the projects such as https://lmstudio.ai/ make accessible running publicly available LLMs locally on laptops. Started Code llama on Mac today in a few minutes, including downloading of the model. Still exploring it, but preliminary results are promising. I'm also amazed how open source collaboration is working together and making it possible, starting with MetaAI research and open sourcing LLaMA models, llama.cpp engine for running LLMs on commodity hardware and Hugging Face for hosting LLMs, many of which are adapted for llama.cpp by enthusiasts.