.png){kind=small}
An opportunity to purchase a revolutionary compression algorithm and patent
Our System is a groundbreaking Variable Length Coding (VLC) Accelerator. It represents a significant advancement in general-purpose data compression. Its applications span across video codecs, AI-generated video, extended reality (XR/VR), and AI neural network training, offering a true paradigm shift in these domains.
Secure a ground-breaking patent crucial for advancements in general-purpose data compression, with applications extending to video codecs, AI-generated video, extended reality (XR/VR) and AI neural network training. This patented
Variable Length Coding (VLC) Accelerator enhances the decades-old methods of handling VLC in computing, such as Huffman, Golomb, Exp-Golomb and Rice, dramatically increasing decoding speeds by a factor of 3x or more.
This innovation represents a totally revolutionary leap, not just a minor tweak, delivering unprecedented data processing speeds in leading-edge devices such as NVIDIA’s Hopper GPUs, Qualcomm’s Snapdragon, ARM and x64 ranges.
This game-changing innovation resolves critical bottlenecks, delivering a performance earthquake not seen this century - a stunning 3x performance improvement over current solutions, and is positioned to be of high interest
to the largest technology companies
The patent’s wide array of applications presents infinite licensing opportunities, particularly as the requirement to decompress data in video and AI technologies become increasingly ubiquitous. This positions it as a compelling
pure-play licensing investment, with the potential to generate annual revenues exceeding $100 million.
Alternatively, this patent is a highly attractive addition to the portfolio of an established technology giant. It offers strategic advantages for proprietary use to gain a competitive edge or to restrict rival access.
The patent itself covers a multitude of scenarios and features 27 separate claims running to 57 pages.
"MaxL Systems has broken new ground with an entropy decoder/ data decompression algorithm that redefines the standard for computational efficiency, offering unprecedented performance and scalability, for both video and AI."
"their decoder technology provides a massive turbo-charge to the processing of the entropy stream – which is in many cases the limiting factor"
"… a x5 speed improvement for entropy decode can be expected, even if implemented poorly. An x20 improvement is possible if it is implemented well!"
"[The novel math] is a beautiful idea - elegant, simple to implement and a great match for real hardware…"
While technology has evolved rapidly, the core speed of computers has changed little, over the past two decades. Apparent advancements are largely due to multicore and parallel processing techniques, which fall short for tasks
like video processing, and other compressed bitstreams, that require sequential data handling.
Modern systems need to efficiently handle vast amounts of data, but current compression techniques like Huffman, Exp-Golomb, Golomb and Rice, produce variable-length symbols that create bit-streams difficult for standard
CPUs to process efficiently, leading to inefficiencies and suboptimal performance in high-demand applications such as video processing and AI.
The MaxL VLC Accelerator Patent shatters existing limitations.
The patent revolutionizes existing decoding techniques by delivering unprecedented acceleration for Huffman, Exp-Golomb, Golomb and Rice decoding. These techniques are foundational elements of many systems including all
modern video codecs, and increasingly used in AI, making this innovative patent a critical part of modern living.
Compression of data is crucial for AI machine learning. During training, an AI model sends large amounts of data between various computation engines. This data is compressed. Applying the MaxL VLC Accelerator in these scenarios
could triple the decompression speed, enabling the analysis of more data in a shorter time.
Exp-Golomb coding is extensively used in newer video codecs such as H.264 (Advanced Video Coding) and H.265 (High Efficiency Video Coding). The widespread adoption of these codecs means that anyone consuming video on a
TV, laptop, mobile, or VR device is directly benefiting from variable length coding.
Huffman coding is used in older video encoders, such as MPEG-1 and MPEG-2, and is extensively applied in still images as an integral part of the JPEG standard.
The MaxL VLC Accelerator can be applied in all these scenarios, extensively in AI and video, resulting in a remarkable 3x increase in processing and chip speeds.
The power of this technology unleashes the potential for the use of other more elaborate, mathematically complex, compression methods, which have hitherto been intractable and overlooked. For video these could lead to a
5% to 10% reduction in required bandwidth, depending on the application.
This results in the highly desirable outcomes of lower heat generation and extended battery life in consumer devices, such as VR headsets, providing competitive advantages, as well as more accurate models in machine learning
applications.
Acquiring this technology is not just about incremental improvement; it's a paradigm shift in data decompression performance. This technology enables a company to lead the industry by offering unmatched performance and
efficiency in data compression solutions.
Alternatively, by seizing the opportunity to own this revolutionary technology exclusively, with full ownership and licensing rights, your company could consolidate a significant and on-going market advantage. In addition,
you can generate substantial external licensing fees, particularly as video and AI technologies, become increasingly ubiquitous.
The Patent has been prepared and filed by leading London based Patent Lawyers Slingsby Partners LLP.
UK Patent Application No. 2411064.5. Title: Decoding of Variable-Length Prefix Code Data. Filing Date: 29 July 2024
The sale of the patent* and technology will be concluded through the acquisition of MaxL Systems Limited, the legal owner of the patent. The process will follow a structured tender format.
However, this does not preclude any interested parties from making a speculative offer that they believe would be compelling enough to forego the formal tender process.
The Executive Summary on this
website is provided as preliminary information to allow parties to assess their level of interest and register such interest with MaxL Systems Limited.
1.Register Interest:Interested parties should register their interest with MaxL Systems Limited.
2.Receive Further Information: Upon registration, further details and the tender process timeline will be provided.
3.Tender Submission: Submit formal tender offers as per the provided guidelines.
4.Evaluation and Selection: MaxL Systems Limited will evaluate the offers and select the most suitable bidder.
5.Acquisition Finalization: Finalize the acquisition of MaxL Systems Limited, transferring ownership of the company and patent.
For inquiries or to register your interest, please register your interest below
About Us
Dr Tim Lee is a prodigious serial inventor.
Tim was a Maths scholar at Cambridge University, and gained a PhD, in the field of Computational Bayesian Statistics.
He loves finding original and pragmatic solutions to previously unsolved problems, resulting in a career full of revolutionary products for multiple companies.
Once again, Tim has combined his love of math and his knowledge of computing, resulting in another stunning invention providing a novel solution to an age-old problem. The break-through Patent has been filed and is owned by MaxL Systems Ltd.
Despite having already deployed billons of licences on previous inventions, this patent is the most far reaching and elegant of his ideas so far.
A few of Tim's many previous original ideas and solutions are listed below.
EP3340067 (A1), GB2558271 (A), US2018181539 (A1), CN201711408367, 20171222, CN105023258 (A), EP2940654 (A1), GB2526067 (A), US2015317778 (A1), US9697594 (B2), US201715628877, 20170621, CN105045559 (A), EP2940576 (A2), EP2940576 (A3), GB2525648 (A), GB2525648 (B), GB2554167 (A), US2015317126 (A1), US9785406 (B2), US201715696385, 20170906
Maths / Computational Bayesian Statistics
Invented a novel numerical method to invert very large sparse matrices.
Later this method was pivotal in the development of [the now ubiquitous] MRI medical scanner.
Computer Science
Created a CAD package which would run on personal computers years before AutoCAD appeared on the market.
Cryptology
Invented a method to securely license software – years before Public-Key encryption was publically available.
Computer Design
Devised a method to install multiple graphics cards in a PC. Built the Software to allow Windows to drive multiple-screens, years before Microsoft offered a similar system.
System sold extensively to finance houses and Air Traffic control systems.
Image Enhancement
Realised a method to add high-dynamic range to both cameras and displays… allowing
• digital cameras to shoot into the sun, or function in near darkness.
• tablets eg an iPad to be readable/usable in direct sunlight or in rapidly variable lighting eg on a train journey.
Mobile Video
Built low power video encoder and decoder … for mobiles, drones, tablets etc. Customers all over the world. Over 2 billion licences sold for the tech.
Mobile Mapping
Invented the tech to track cars/trucks and show them on maps – just like Air Traffic Control show planes. Both Tracking, Route-Finding and Mapping deployed commercially before Google-Maps existed.
LIDAR usage
Devised Software to extract Power Pylons/cables from LIDAR data, allowing utilities to inspect their assets.
We are in talks with several of the major chip manufacturers, server companies and large language model providers. To be part of the discussion please register your interest.
.png){kind=small}