Sarah Guo’s Post

🌟 Thrilled to Re-Share Our Big News! Recently, Meeami Technologies proudly announced our groundbreaking low-footprint noise suppression solutions for Cadence Tensilica HiFi DSPs. This achievement is not just a leap in Audio AI but a testament to the power of collaboration and innovation. 🔍🎯 At Meeami Technologies, we're revolutionizing Audio AI. Specializing in compressing complex models for low compute platforms, we've achieved a significant breakthrough with our unique deep learning model design for edge-based processing. This innovation delivers clear, distortion-free audio and is tailored for each HiFi DSP model, exemplifying our commitment to performance and versatility. 🤝 Our collaboration with Cadence leverages their DSP strengths—optimized audio processing and lower power consumption—for our advanced noise suppression. This partnership enhances user experiences across a spectrum of voice-controlled devices and audio applications. 👏 I'm incredibly proud to lead a team of experts who are not just skilled but are also pushing the boundaries of what's possible in AI. Their dedication and innovative spirit have enabled us to stay at the forefront of our industry. 🌍 To our partners and the tech world: Our work demonstrates that advanced, efficient AI can thrive even on minimal resources, expanding its reach beyond high-end devices. #AI #Audio #audioengineering #audioinnovation #HiFi #HiFiMini #Noise #NoiseSuppression #DeNoising #DeNoise #noisereduction #Tensilica #dsp #HIFIDSP #spatialaudio #MeeamiTechnologies #Cadence #AudioAI #Innovation #Teamwork #IndustryLeadership #EfficientAI

We've been talking a lot about earables or hearables over the past 4 years. This is an emerging category where all the action happens in and around the ear. What's interesting about this part of the body is that it can passively and continuously measure our moods/ biometrics, it can change our moods via vagus nerve stimulation and it is a good place for brain control interfacing. (Our thoughts, via the ear, can control devices.) All these use cases and applications exist today although they have not yet merged into one integrated solution and that's when it gets...interesting. [Apple has a number of patents in these areas. Airpods are soon going to do a lot more. I'm interested to see Apple's Gen. AI on the Edge news, and if and how Airpods will be used.] What is critical to us and our clients is the future of earables and how it enables the coming Algorithm to Algorithm economy (A2A). We have seen glimpses of what this could look like from science fiction. In Ender's Game, to name one of a few classics, AIs are interfaced through the ear. This company https://iyo.audio/ is an interesting signal. They are betting that voice control of personal AI agents via the ear is going to be a dominant future interface paradigm. We know that Humane.ai and others struggled with v1 but these are vital signals of the Algorithm to Algorithm economy. Interfaces and companies like IYO are the building blocks of the future. (Finally, if you want to dive in more to the future of autonomous agents...this is a good place to start. https://lnkd.in/g8GCmrVN)

How about some multimodal video to audio? Check out Google's fresh clips (with the sound on 👂🏻). Google DeepMind has made remarkable progress in V2A generative technology, capable of adding sound to silent clips that match the scene's acoustics and desired on-screen action. Incorporating V2A into our workflow could significantly enhance future post-production pipelines as having preliminary audio post-production & SFX done already in the offline phase will streamline the entire process, especially in the less demanding use cases. 🛠️🎶 More 👉🏻 https://dpmd.ai/v2a Thanks Alku Sirén for the link!

🌟 We are happy to introduce the "Essensys Tech News" series! From now on we’re going to shine a light on the best and brightest in our industry. From insightful blogs to must-read PDFs, news updates, and invaluable resources, our #EssensysTechNews series is your go-to for staying up to date with everything tech. Get ready to explore, learn, and discover with us as we curate top-notch content. Here’s the latest: 1. After 35 hours of negotiation, the EU becomes the very first continent to set clear rules for the use of AI 🇪🇺. The AI Act is a forward-thinking legislation covering issues like facial recognition. The law comes into effect in 2025. Read more about this milestone: https://lnkd.in/eHQWRjWx 2. Spatial audio is delivered by Apple and Dolby Atoms. Apple is rewarding musicians and record labels for creating music with spatial audio technology. They say that music sounds more lifelike and engaging with Dolby Atoms, making you feel like you are at the very center of everything. Some of the artists who have released music in Dolby Atmos are Coldplay, Beyonce, Billie Eilish. We’re sure more will follow. Check out more on this here: https://lnkd.in/eBNUu_DK 3. Transitioning from treadmill walks to VR goggles marks a seismic shift in mouse experiments. With these cutting-edge goggles, mice can experience 3D virtual environments. But why the fuss? Because decoding these processes helps us understand how our brains perceive the world around us. It's not just about mice, it's about unlocking the mysteries of our own minds. Details here: https://lnkd.in/edDXNdqk #EssensysSpotlight

Catching up on this really great talk from this year's #NAMM hosted by GPU Audio. One really inspiring aspect of the talk is that all of the panelists seem to agree that the key to the success of AI in the music space is that the tools produce not ready-made outputs (ie, DallE) but produce malleable and adaptable elements that *augment* and not replace human creativity. https://lnkd.in/ePdemaZe

🎧 #Day 22 of 360: Today, I dove into the fascinating world of wavelet transformers and their application in extracting domain features from audio data. 🌊🔍 Wavelet transformers offer a powerful way to analyze signals across different scales, making them particularly effective for tasks like audio feature extraction. By decomposing signals into wavelet coefficients, we can capture both time and frequency domain information with great precision. Understanding these techniques not only enhances my grasp of signal processing but also opens up exciting possibilities in fields ranging from speech recognition to music analysis. Looking forward to applying this knowledge to enhance my work in audio processing and beyond! #WaveletTransformers #AudioProcessing #SignalProcessing #MachineLearning #DataScience #ContinuousLearning

🎧 Exciting Developments in AI Audio Generation! Dive into an in-depth analysis of AudioCraft - Meta's revolutionary AI tool reshaping the audio and music generation landscape. Perfect for those in the audio, tech, or AI sectors. Explore the future landscape of sound through this article: https://lnkd.in/gbVvEumW #AIaudio #AudioCraft #FutureOfSound #TechInnovation

Many of the recent advancements in AI-generated image, video, and audio are developed using diffusion models. Technical deep dive: Read how diffusion models function and how to improve on your own models. ➡️ https://nvda.ws/3thtzoo

Imagine this: Three tech giants come together, each a master in their field, all to put the boAt Supreme to the test. They dive into its coolest features, like AI ENx technology that makes every call crystal clear, the convenience of multipoint connectivity that lets you switch between devices seamlessly, and the smart in-ear detection that knows exactly when you're listening. With the introduction of the boAt Supreme, we have also redefined audio excellence by incorporating spatial audio support with 24-bit audio processing technology. Could we be more SUPREME? Check out the video: https://lnkd.in/gZYGW9V6

My team IC3-AI is organizing two Grand Challenges for ICASSP 2024. The first is the ICASSP 2024 Speech Signal Improvement Signal Processing Challenge, which is the second iteration of this challenge. This challenge is focused on improving speech quality if the signal is distorted due to a poor-quality microphone, audio clipping, a highly reverberant room, or poor-quality audio processing. Details are at https://lnkd.in/gnFWWvZD and in the below document. The second is the ICASSP 2024 Audio Deep Packet Loss Concealment Challenge, which is also the second iteration of this challenge. Packet loss is still a top issue that impacts audio quality, and this version of the challenge has even more challenging scenarios. Details are at https://lnkd.in/g9mGUB8t.