← Back to companies
Company Logo

Patents Issued

// Created: Unknown

Patents, in general

For the most part, I do not believe in the value of patents in today's economic landscape, except to treat them as "trading cards" and a brand building marketing expense.

He who has the gold rules --wise attorney

The fundamental right that comes with a patent is to sue someone in court for infringement if they copy your "unique design". It is an expensive proposition, at best, to attempt prosecuting someone, especially a large well-funded tech company, for such violations. They know this, and often use patents of others with impunity, and the knowledge that even if they get caught, the expense of paying attorneys to keep fighting in court (say, $100,000 per month), is far less than they are making by using the idea.

As an illustration, my father changed the world of telescope mirrors, but history barely remembers his name. In the early 1960s, he invented and patented a revolutionary machine that could grind mirrors 20× more accurately than anything before (USPTO: 3,141,265), achieving 40nm precision—a level previously possible only with painstaking manual effort. He saw the problem that no one else did: the gear-driven machines of the time were inherently flawed, producing harmonic lissajous distortions like a spirograph toy. His solution was brilliant yet simple—replace gears with a belt driven analog computer computing an irrational number, breaking the flawed cycles and allowing for truly smooth, sub-wavelength precision optics. It worked. Almost immediately after his patent was published, the market shifted. High-precision mirrors, once out of reach, were now mass-produced and sold at 100× lower prices. He walked through factories and saw his ideas in action, but he never saw a dime from it. No financial backers, no legal team, no protection—his revolutionary invention became industry standard, while he remained an outsider to the wealth and recognition that followed. His legacy lives on in every high-precision mirror that came after. But in the business of invention, sometimes the greatest breakthroughs are the ones history forgets.

I'd rather not deep dive further on the topic here, so I am switching gears (irrationally {sic!}) to summarize the patents that Apple filed and received assignment of for my work, as well as one which I did file myself.

Summary of US Patent 5,875,354 (1999) and Its Relevance Today

Title: System for Synchronization by Modifying the Rate of Conversion by Difference of Rate Between First Clock and Audio Clock During a Second Time Period

Inventors: Paul Charlton, Keith Gurganus

Assignee: Apple Computer, Inc.

Core Innovation:

This patent describes a method and apparatus for synchronizing the presentation of time-dependent media (audio, video, etc.) by dynamically adjusting the rate of conversion of media samples. The system ensures that playback remains synchronized even when different clocks (e.g., system clock vs. audio clock) drift over time.

Key Algorithmic Approach:

  1. Rate Conversion Process: • The media stream is initially processed at a predetermined rate. • The system continuously measures drift between two clocks: • First clock: System (real-time) clock • Second clock: Audio subsystem clock (e.g., DAC clock) • Based on measured drift, the system modifies the rate of conversion dynamically to maintain synchronization.
  2. Dynamic Adjustments: • The rate of conversion is adjusted periodically based on sample count errors. • A phase-locked loop (PLL)-like feedback mechanism (software-based) fine-tunes playback speed over time. • Ensures smooth media playback by compensating for clock drift, latency issues, and jitter.

Relevance in Systems 30+ Years Later

This patent laid foundational principles for modern audio/video synchronization in digital systems, and its principles are still highly relevant in various domains:

  1. Adaptive Media Synchronization in Streaming and Broadcasting • Modern applications: • Video-on-demand (Netflix, YouTube, Disney+) • Live broadcasting (Twitch, Zoom, WebRTC) • Adaptive bitrate streaming (DASH, HLS) • Key challenge: Synchronizing media across different devices and networks. • Solution derived from this patent: • Many modern AV sync algorithms use clock drift compensation based on techniques described in this patent.

  2. High-Precision Audio Processing & DAW Software • Modern applications: • Digital Audio Workstations (Pro Tools, Ableton Live, Logic Pro) • Real-time audio synthesis and rendering • Challenge: • Digital-analog (DAC) drift and mismatched sample rates affect pitch, latency, and sound quality. • Solution: • Real-time resampling and drift correction (as outlined in this patent) are integral to modern ASIO drivers and pro audio software.

  3. IoT and Edge Computing: Real-Time Synchronization • Modern applications: • Industrial IoT (IIoT), Smart Cities, Automotive ADAS, Aerospace Systems • Challenge: • Distributed sensors and actuators operate on different internal clocks, leading to timing drift. • Solution: • Dynamic clock synchronization and rate conversion algorithms (like those in this patent) ensure precision in autonomous vehicles, real-time analytics, and control systems.

  4. Machine Learning & AI in Video Processing • Modern applications: • AI-powered video compression, motion interpolation, and deepfake detection. • Challenge: • Synchronizing AI-generated frames with real-time audio requires temporal alignment. • Solution: • Rate conversion methods from this patent are embedded in AI-based video synthesis tools (e.g., DeepMind’s Wavenet, NVIDIA Maxine, Adobe Premiere Pro AI features).

Conclusion

The real-time synchronization method in US5875354 remains a cornerstone in modern media processing, AI-driven video/audio synchronization, and real-time data analytics.

Its impact is still seen 30+ years later in: • Streaming services (Netflix, YouTube, Twitch) • Professional audio & video processing (DAWs, Adobe, Final Cut) • IoT & Edge computing (industrial automation, smart cities) • AI-enhanced media (deep learning for motion sync, VFX)

This patent was ahead of its time and helped shape robust media synchronization frameworks that underpin the global digital economy today.

Summary of US Patent 5,825,359 (1998)

Title: Method and System for Improved Arbitration of a Display Screen in a Computer System

Key Concepts & Contributions • The patent describes a method for arbitrating access to a computer display screen among multiple rendering processes in a preemptive multitasking environment. • It introduces a queue-based arbitration system that resolves conflicts when multiple processes attempt to write to overlapping display areas. • The system maintains two collections:

  1. Rendering Collection – Currently active display areas.
  2. Waiting Queue – Pending display areas waiting for dependencies to clear. • The arbitration method: • Determines dependencies between candidate display areas and currently rendering/waiting areas. • If dependencies exist, the candidate display area is placed in the waiting queue. • If no dependencies exist, it is placed in the rendering collection and processed immediately. • Ensures that no two conflicting areas are drawn simultaneously, reducing flicker, corruption, and latency. • The approach is designed to optimize graphical output scheduling, particularly in windowed environments where multiple applications need to render graphics efficiently.

Relevance 30 Years Later • Modern GPU & Window Compositing Managers: Today’s graphical systems, such as macOS’s Quartz Compositor, Windows DWM (Desktop Window Manager), and Linux Wayland/X11 compositors, handle window compositing in a way that is conceptually similar to this arbitration method. • Virtual Desktop Environments: Remote desktop and virtualization solutions, such as VMware, Citrix, and RDP, still face similar challenges in arbitrating screen updates when dealing with multiple rendering processes over a network. • Real-time Graphics & Gaming: The principle of rendering arbitration applies to modern game engines (Unreal, Unity) and video compositors (OBS, FFMPEG, AV1 encoding), where frame synchronization and efficient scheduling remain critical. • Cloud-based UI Rendering: The cloud-based application delivery model, such as Amazon AppStream or Windows 365, must intelligently schedule UI updates to optimize bandwidth and latency—making arbitration techniques crucial.

This patent essentially anticipated compositing window managers and multiprocess display arbitration techniques that are now a standard part of modern graphical user interfaces.

Summary of US Patent 5,949,434 (Paul Charlton, 1999)

Title: Method and Apparatus for Scaling Graphic Images Filed: March 20, 1997, Issued: September 7, 1999,

Key Innovations

This patent describes an improved method for scaling digital images while minimizing visual artifacts, particularly banding effects that occur during image enlargement or reduction. Traditional scaling methods involve replicating or dropping pixels, which can create visible horizontal or vertical bands. This invention introduces pseudo-random adjustments to pixel placement, reducing artifacts while maintaining image quality.

  1. Bresenham’s Algorithm Enhancement • The method builds on Bresenham’s algorithm, commonly used in computer graphics, but modifies the error term to introduce semi-random shifts in pixel placement. • This breaks up repeating patterns that would otherwise create visible bands.
  2. Precomputed Series Approach • The algorithm precomputes scaling factors and stores them in a lookup table. • Instead of dynamically computing pixel placement during processing, it retrieves precomputed values, improving efficiency.
  3. Directional and Alternating Processing • Alternates row and column processing (e.g., even rows left-to-right, odd rows right-to-left) to further disrupt banding patterns. • Uses different scaling sequences for even and odd rows, ensuring visual artifacts do not persist in a regular pattern.

Relevance in Modern Systems (30 Years Later)

This patent introduced adaptive, computationally efficient image scaling, a problem that remains critical today in various applications:

  1. Modern Image Scaling (AI-based Super Resolution & Streaming) • The concept of adaptive pixel placement has been expanded in AI-based super-resolution algorithms like NVIDIA’s DLSS, AMD’s FSR, and Apple’s Neural Engine. • Video streaming services (Netflix, YouTube) rely on real-time upscaling for adaptive resolution adjustments.
  2. GPU Texture Scaling & Game Rendering • Video games and 3D applications use dynamic resolution scaling where images are rendered at lower resolutions and upscaled in real time to maintain frame rates.
  3. Mobile & Web Graphics Smartphones and web-based applications (e.g., HTML5 Canvas, WebGL) scale images dynamically across different screen resolutions.
  4. Machine Vision & Medical Imaging • High-precision scaling techniques are used in CT/MRI image processing and machine vision applications where retaining fidelity while resizing is crucial.
  5. AI-based Image Enhancement • AI-driven denoising and super-resolution algorithms now incorporate stochastic elements to remove visual artifacts—a technique conceptually similar to this patent’s approach.

Conclusion

This patent was ahead of its time in addressing visual artifacts in image scaling. The pseudo-random error modification technique aligns conceptually with modern AI-based super-resolution and adaptive scaling, making it highly relevant even today in fields like streaming, gaming, machine vision, and medical imaging. The principles laid out in the patent are now fundamental to image-processing advancements across cloud computing, GPUs, and AI-driven graphics rendering.