Skip to main content
Lab

TsugiSilicon

Hardware engineering. Multi-stream HDR pipelines, hardware-aware compression, and the silicon that ships in front of the viewer.

TsugiSilicon is the lab where the algorithms hit physical decoders, FIFOs, PLL trees, and consumer-class SoCs. Output is reference designs, FPGA bring-up notes, measured-data exhibits, and provisional patent disclosures positioned for AOMedia AV2 and adjacent standardization tracks. The thesis underneath: resolution-layering and obfuscated dual-layer compression have a clean freedom-to-operate window opening as the foundational 1990s-era patents from Demos, Demografx, and the early Dolby Vision predictive-encoder cluster have expired or expire through 2026, and TsugiSilicon's two filed provisionals stake additive-only and obfuscated-split positions in that window.

Research lines

Mechanism overview

Two filed provisionals carry the lab's current technical thesis, operating at different scales of the same resolution-layering problem. Trinity (App. 63/987,139) addresses the FPGA-and-discrete-IC scale. A digital logic device hosts per-decoder asynchronous dual-port FIFOs (sized to several scanlines), a hysteretic watermark comparator, and a physically separate sideband signaling path that phase-locks plesiochronous discrete decoder ICs by commanding them to extend their vertical blanking interval when associated FIFO fill exceeds a high watermark. The output is scanline-accurate additive pixel-domain composition with saturating arithmetic across three fidelity-partitioned streams (base, color delta, gray delta). Additive-only reconstruction, distinct from the predictive-encoder approach in the Dolby Vision Profile 7 ecosystem.

DLC (App. 64/054,446) addresses the consumer-SoC scale. A pre-encode pixel-domain transform composes three keyed components (key-seeded tile permutation that preserves within-tile correlation, GoP-granular polarity inversion, and integer-root-of-unity chroma-axis rotation) and splits a high-precision master into two complementary streams such that no single layer is visually intelligible. Aggregate compression-efficiency cost lands in the low-double-digit-percent range against an unobfuscated split, versus near-total loss for naive visual cryptography. The decoder side orchestrates two concurrent hardware decoder contexts on a single consumer SoC under a shared PLL clock tree, with PTS-paired buffer matching and a proportional-control software throttle on the leading session. The same coordination problem Trinity solves with hardware sideband at the FPGA scale, solved with software-throttled rate control at the on-SoC scale. Hardware playback is verified on Amlogic S922X-J (Ugoos AM6B+).

DLC encoder-side schematic. Master split into Layer A and Layer B via a three-component keyed transform, recombined into a single container.
Figure 1. DLC encoder-side schematic. Source: TsugiSilicon. App. 64/054,446 figure derivative.
Filing timeline showing Trinity, DLC, and the rest of the portfolio with PCT conversion milestones.
Figure 2. Filing-and-PCT timeline. Trinity PCT 2027-02-19, DLC PCT 2027-05-01. Source: TsugiCinema portfolio analysis, 2026-05-08.

Standardization track

The lab's primary standardization vector is AOMedia, with an engagement window targeted for 2026-08-01 (ninety days post-DLC filing). Promoter-tier membership is the current recommendation. AV2 specification work remains in draft as of May 2026, and the year-end-2025 launch did not complete on schedule, so the window favors a secure-transport-profile addendum positioning aligned with the AV2 12-bit professional tier. The strategic premise is that AV2 codecs will eventually want a distribution-side obfuscated dual-layer container for premium HDR content, and DLC's keyed pixel-domain pre-encode is one of the few candidate mechanisms that preserves codec-internal compression efficiency rather than fighting it. Counsel sign-off on standard-essential-patent scope is a precondition before any working-group contribution.

Counsel posture and commercialization horizon

Same prosecution counsel for both filings. Trinity twelve-month conversion and PCT deadline 2027-02-19. DLC twelve-month conversion and PCT deadline 2027-05-01. Track One Prioritized Examination at non-provisional filing recommended for both, to secure final disposition before the 2030 commoditization horizon for resolution-layered HDR distribution. PCT geographic scope decision targeting JP, KR, CN, and EU. For DLC, the open-issue list flags software-throttled rate control as a potential KSR-style obviousness risk against the Trinity sideband mechanism (counsel review at non-provisional drafting), and a measured-data exhibit is queued to strengthen the otherwise-constructive working examples ahead of Track One conversion.