US 5802068

I'll verify this patent against live sources before summarizing. Let me run targeted searches.

Let me check the CAFC dockets and the specific district court case tied to this patent.

US Patent 5,802,068 — Verification Summary

I cross-checked the USPTO-based text (Google Patents, Justia, uspto.report) against the record you supplied. Note on identifiers: I interpreted "5802068" literally as US 5,802,068 and did not auto-correct anything. One caution: I found no evidence of any CAFC 2026 docket involving this patent — see "Uncertainties" below.

Bibliographic data

Abstract (paraphrased from the record)

A plurality of "packs" are built from a first packet carrying high-bit-rate data (e.g., video) and a second packet carrying low-bit-rate data (e.g., audio). The apparatus detects whether the data length of the second data is smaller than a maximum data length permitted for the second packet. If so, a pack is assembled from a redundancy (padding) data packet whose length equals the difference between the maximum packet data length and the actual second-data length, together with the second-data packet — so the pack is filled to a constant size.

Plain-language overview of the independent claims

• Claim 1 — A multiplexing apparatus that forms "set assemblies" (packs) from a first set part with at least high-bit-rate data and a second set part with at least low-bit-rate data. It has (a) means for detecting whether the second data's length is smaller than the second set part's maximum data length, and (b) means for building one of the set assemblies out of a redundancy data set part equal in length to that difference plus the second set part, when the detector finds the second data is short.

• Claim 4 — A multiplexing apparatus that multiplexes high-bit-rate first data and low-bit-rate second data into a "set of programs." It has (a) calculating means that computes the second set part's data length from a predetermined parameter, (b) varying means that adaptively varies that data length based on the calculation, and (c) forming means that builds a set assembly from the varied-length second set part plus a redundancy data set part filling the varied portion. (The core idea: packet length is computed dynamically rather than fixed.)

• Claim 6 — A video/audio multiplexing apparatus forming program sets out of packets. It has (a) audio packet data length calculating means, (b) audio packet data length varying means that adaptively varies that length, and (c) pack forming means that forms a pack containing the audio data from the audio packet plus a redundancy data packet.

• Claim 15 — An MPEG-system multiplexing apparatus where packs are built from a first packet with at least high-bit-rate video data and a second packet with at least low-bit-rate audio data. It has (a) means for detecting whether the audio data length is smaller than the second packet's maximum data length, and (b) means for constructing a pack from a redundancy data packet of length equal to the difference between the maximum packet length for the audio and the actual audio data length, together with the audio data packet.

Dependent claims (context): 2–3 specify video/audio/padding data; 5 specifies video and audio; 7 defines the calculation inputs (audio bit rate, sampling frequency, encoding layer, pack length); 8 uses one audio frame as the unit; 9 adds redundancy-packet length calculating means; 10–13 add frame-number deciding means (to improve multiplexing efficiency when bit rate shifts from a middle to a high rate); 14 adds source selecting means among video, audio, and redundancy data.

Technical gist: The invention adaptively sizes the audio packet (as an integer number of audio frames) and pads the pack up to the fixed maximum packet length, which shrinks the audio-induced initial delay and the required input video buffer (the spec's worked example cites ~43.2 kB → 16.2 kB and "40% or less" of conventional initial delay).

Prior art cited on the face

US 3,838,221; US 4,224,473; US 4,864,562; US 5,650,825 (Matsushita — "sending private data instead of stuffing bits in an MPEG bit stream").

Litigation

The Google Patents record flags this family as having litigation and links two items: a Darts-IP global litigation entry (family ID 16220430) and a Texas Eastern District Court case, 2:05-cv-00484 (per the Unified Patents litigation link). I was not able to independently retrieve the docket details or party names for that case within this session.

Uncertainties (explicit)

1. No CAFC 2026 docket found. My searches for "5802068" / "5,802,068" in a 2026 Federal Circuit context returned nothing. Given the patent expired 2016-06-25, a 2026 CAFC appeal would be unusual (though not impossible for an old case). I cannot confirm any such docket exists.
2. My search results did not surface a 2026 docket for this specific patent — the MediaTek 2025–2026 litigation results that appeared involve other patents (e.g., Freedom Patents LLC asserting 8,284,686 / 8,374,096 / 8,514,815; Daedalus Prime; ServStor; Mimir IP asserting SK Hynix patents). None of those reference 5,802,068.
3. Assignee chain and dates are taken from the assignment records reproduced in the authoritative patent text; I did not retrieve the underlying USPTO assignment documents directly.
4. I did not locate the actual 2:05-cv-00484 docket, so I cannot state the parties, outcome, or whether 5,802,068 was an asserted claim.

If you want, I can attempt a more targeted retrieval of the 2:05-cv-00484 docket and a Westlaw/PACER-style CAFC docket check for 2026.

I'll research litigation involving US patent 5802068. Let me run several searches.

Let me search the specific litigation portal data and the Texas case referenced on the patent page.

Excellent — I found a key lead. Let me dig into the specific case details.

Summary of findings — U.S. Patent No. 5,802,068

Important caveat on my process: I was able to run several searches before hitting a tool-step limit, so what follows is what I could substantiate from the sources retrieved. I did not get to run a full PACER/Unified Patents/CourtListener docket pull, so treat this as a strong partial picture rather than an exhaustive docket history. No litigation result below has been verified against a primary docket record.

I interpreted the number strictly as 5,802,068 / US5802068 and excluded look‑alikes (e.g., RFC 2068, the Pfizer '068 patent, the Westvaco '068 design patent, etc.), which are unrelated.

Case 1 — MediaTek, Inc. v. Matsushita Electric Industrial Co., Ltd. (MEI) and Panasonic Corporation of North America

Source: MediaTek Inc., English Translation of Financial Statements, 2005 Q4 Consolidated Results (Notes to Financial Statements — legal contingencies), https://www.mediatek.com/hubfs/MediaTek%20Assets/Pdfs/Financial%20Reports/2005/Q4-Consolidated-Results.pdf — which states: "Further, on October 17, 2005, the Company filed a complaint for patent infringement against MEI and Panasonic in the Eastern District of Texas, asserting that the DVD recorders and digital TVs of MEI and Panasonic infringe upon the Company's U.S. Pat. No. 5802068 and 6118486 respectively. The Company is seeking monetary damages…"

Case 2 — Texas Eastern District Court case 2:05-cv-00484 (plaintiff/defendant not stated)

The Google Patents record for US5802068 carries an embedded "litigation" entry stating that a US case was filed in the Texas Eastern District Court, with the docket link pointing to:

Case 2:05-cv-00484, Texas Eastern District Court — via Unified Patents Litigation Data
(https://portal.unifiedpatents.com/litigation/Texas%20Eastern%20District%20Court/case/2%3A05-cv-00484)

The patent page also carries a Darts-ip "First worldwide family litigation filed" link for family ID 16220430, confirming that litigation has in fact occurred on this family.

Caution (per the strict no-auto-correct rule): The Google Patents entry gives only the court and case number — no party names or filing date. Case 2:05-cv-00484 is docketed in the same court and same year as the October 17, 2005 MediaTek v. MEI/Panasonic suit above, so they may well be the same action, but I cannot confirm that from the retrieved materials and am not treating them as identical. I list them separately.

Ownership chain relevant to standing (from the patent record)

Original assignee Nippon Steel Corporation (JP priority JP7-188251, 1995-06-30; filed 1996-06-25; granted 1998-09-01) → assigned to United Module Corporation (recorded 2005-01-19) → assigned to MediaTek, Inc. (recorded 2005-01-21). This explains why MediaTek — not Nippon Steel — appears as plaintiff in the 2005 Texas action, and why "MediaTek Inc." is the current assignee of record. The patent expired 2016-06-25 (anticipated expiration).

What I could not confirm

• No PTAB/IPR proceedings tied to 5,802,068 surfaced in my results. (The "UMC"/'686-patent IPR hits in my search were a different patent and a different UMC entity — United Microelectronics-related — and I did not attribute them to the '068 patent.)
• No CourtListener/PACER primary docket for 2:05-cv-00484 was retrieved, so I have no outcome, party list, or filing date for that docket entry.
• No CAFC appeal involving 5,802,068 was identified.
• The final disposition of the MediaTek v. MEI/Panasonic action is unknown to me from these sources.

If you need this verified to litigation-grade, the next step would be a PACER docket pull on E.D. Tex. 2:05-cv-00484 and a CourtListener/Unified Patents name search for MediaTek's 2005–2007 E.D. Tex. filings against Matsushita/Panasonic. I flag explicitly that I could not complete those pulls here.

I'll verify the canonical (empty) PTAB list against public sources before writing this up.

No AIA proceedings surfaced in those results — the "580" hits are U.S. 8,023,580 (a Rembrandt patent). Let me run targeted checks on the '068 specifically, plus its enforcement history.

PTAB / AIA Trial Landscape — U.S. Patent No. 5,802,068

Method note: The canonical proceeding list comes from the structured "PTAB proceedings on file" block (USPTO Open Data Portal). I ran targeted web verification to catch anything ODP hasn't ingested (PTAB E2E, Darts-IP litigation family data, docket aggregators, news). Search results for "5,802,068" and "5802068" consistently returned either (a) the unrelated Rembrandt patent U.S. 8,023,580 (whose IPRs reexams dominate the "580" query space), or (b) district-court/ITC activity — never an AIA trial number for this patent. I did not invent any proceeding numbers; none exist to report.

Proceedings overview

Total AIA trial proceedings on U.S. 5,802,068: 0 (active: 0 / claims invalidated: 0 / claims sustained: 0 / settled: 0 / institution denied: 0 — no proceeding, so no disposition category applies).

Bottom line for a defendant: there is no PTAB record to lean on. This is not "the patent survived IPRs and is hardened," and it is not "the claims are dead." It is a 1990s-vintage, pre-AIA MPEG multiplexing patent (Nippon Steel → United Module → MediaTek chain of title) that expired 2016-06-25 and appears never to have been challenged in an AIA trial at all. Any invalidity attack today would be a de novo effort with no Board findings, no estoppel, and no free "claims canceled" argument — but the patent's expired status may matter far more than its PTAB history.

Proceedings

None to enumerate. There is no IPR, PGR, or CBM number associated with U.S. 5,802,068 in the structured data, and no AIA proceeding surfaced in web verification.

For completeness, the reasons the three AIA vehicle types are foreclosed or inapplicable:

• PGR — unavailable. The patent issued 1998-09-01 from an application filed 1996-06-25 (priority 1995-06-30). PGR is limited to patents with an effective filing date on or after 2013-03-16. Structurally impossible here.
• CBM — unavailable. The transitional CBM program sunset on 2020-09-16, and in any event the claims are directed to packet/pack multiplexing of video and audio (see claim 15: "a multiplexing apparatus according to an MPEG … system"), not "a financial product or service." Not a CBM-eligible patent.
• IPR — the only theoretically available vehicle (pre-AIA patents are IPR-eligible at any time, 35 U.S.C. § 311(c)), but none was ever filed on this record.

Non-PTAB context (flagged as outside the AIA list)

I am deliberately keeping these separate — they are not PTAB proceedings, and I could not fully verify them beyond third-party datasets:

• Google Patents litigation flag. The family record carries a Texas Eastern District Court litigation entry, case 2:05-cv-00484 (link embedded in the patent page: https://portal.unifiedpatents.com/litigation/Texas%20Eastern%20District%20Court/case/2%3A05-cv-00484), plus a "family has litigation" Darts-IP pointer. Source page: https://patents.google.com/patent/US5802068/en. These are district-court artifacts, not AIA trials; Google expressly disclaims accuracy of the assignee/litigation fields.
• Alleged counterclaim use against Matsushita. A National Taiwan University thesis (secondary source: https://tdr.lib.ntu.edu.tw/bitstream/123456789/7103/1/ntu-109-1.pdf) states MediaTek counterclaimed on three patents, including U.S. 5,802,068 ("MULTIPLEXING APPARATUS OF A PLURALITY OF DATA HAVING DIFFERENT BIT RATES"), which MediaTek acquired from Nippon Steel. This is consistent with the recorded 2005 assignments (Nippon Steel → United Module 2004-11-24 → MediaTek 2005-01-20). Treat as an unverified tertiary characterization of litigation posture, not a PTAB citation.
• No reexamination found. My searches surfaced no ex parte or inter partes reexamination certificate for 5,802,068. I cannot rule one out with high confidence; ODP's reexam records were not part of the provided structured block.

Strategic summary

Claim status: all 15 claims are UNTESTED at the PTAB. Claims 1–15 (including independent claims 1, 4, 6, and 15, and dependents 2, 3, 5, 7–14) have never been adjudicated by the Board. There is no FWD to cite, no canceled claim to point to, and no patentability holding — express or implied — to build estoppel on. If a demand letter cites claims 1–3 or 15 (the MPEG-framed claims), you cannot answer with "the Board already canceled that claim," because it never did.

Estoppel landscape: empty. Section 315(e)(2) estoppel is petitioner-specific and attaches only upon a final written decision on a claim. With zero proceedings, no party — MediaTek included — is estopped from anything as to this patent, and no defendant inherits a privy-based estoppel from an earlier challenger. Conversely, you also inherit no benefit: there is no prior petitioner work product or instituted ground that you can piggyback on, and no IPR in which to seek joinder under § 315(c).

Pattern signals: none. No serial-petitioner behavior (no petitioner at all), no patent-owner appeal history to the Federal Circuit, and no defensive aggregator (e.g., Unified Patents, RPX) in the chain. The 2005 assignee sequence — a Japanese steel company's patent sold to a small California entity and then to MediaTek — is the classic profile of a patent acquired for counterclaim/portfolio use rather than for an NPE assertion campaign, which fits with the absence of any post-AIA challenge: well-asserted patents eventually attract IPRs; this one appears not to have been asserted aggressively enough in the AIA era to draw one.

The bigger picture is expiration, not validity. The record shows anticipated expiration 2016-06-25 and current status "Expired - Lifetime." That has two consequences that dominate any invalidity analysis: (i) under 35 U.S.C. § 286, damages reach back only six years from filing, so a suit filed today can recover nothing for post-2016 conduct because there is none; and (ii) for any claim-construction fight, expired claims are construed under Phillips rather than the broadest reasonable interpretation — the same standard a district court would apply. An expired patent with no PTAB history is generally a staleness problem for the plaintiff, not a validity problem for the defendant.

Recommended next steps

1. Do not represent that this patent has PTAB history. Your § 282 invalidity case, if any, is a clean slate. Nothing is canceled, nothing is sustained, and no estoppel helps or hurts you.
2. Lead with the expiration and § 286 posture. Verify the expiration date against the USPTO PatentCenter maintenance-fee record (fee year 12 paid 2010-03-01 per the legal events; the next window lapsed, producing the 2016-06-25 anticipated expiration). If no maintenance fee was paid, confirm whether the patent actually lapsed earlier than the 20-year term — that changes the damages tail. This is a threshold, dispositive-style defense and requires no invalidity proof.
3. If you nonetheless need invalidity contentions, note the built-in prior-art hooks this specification creates: the patent's own Background admits the MPEG pack/packet multiplexing framework (2,324-byte packs matched to CD-ROM sectors; ~36 msec audio frames; the 1,200,000/64,000 bps example) and cites no specific references. The one U.S. reference cited on the face is US 5,650,825 (Matsushita, "Method and apparatus for sending private data instead of stuffing bits in an MPEG bit stream") — directly on the padding/stuffing point that is the invention here. That is a § 102/§ 103 starting point, but it has never been run through the Board.
4. If you have any interest in an AIA filing, understand it is likely both time-barred and pointless: § 315(b) bars a petition more than one year after service of a complaint, and an expired patent offers no prospective injunctive or ongoing-royalty relief to enjoin. Do not spend IPR budget on a 2016-expired patent.
5. Re-verify before relying on the zero count. The ODP list is only as fresh as its most recent ingest, and PTAB E2E's public search (https://ptacts.uspto.gov) is the authoritative cross-check. I found no contrary evidence, but ODP ingest lag is the one failure mode that would change this report — and even a late-discovered IPR would not change the expiration-based conclusion above.

Confidence: High that no AIA trial is on file as of 2026-09-28 (structured block + four independent search angles all consistent). Moderate that the 2005 EDTX case and the MediaTek/Matsushita counterclaim narrative are accurately characterized — those rest on third-party datasets I could not open directly, and I flag them rather than rely on them.

I'll research the assignment chain and litigation history for US 5802068 before compiling the report.

US 5802068 — Assignment & Ownership Reconstruction

Patent: US 5,802,068 A — "Multiplexing apparatus of a plurality of data having different bit rates"
App. No.: 08/668,328 · Filed: 1996-06-25 · Priority: 1995-06-30 (JP 7-188251) · Issued: 1998-09-01 · Status: Expired – Lifetime (anticipated expiration 2016-06-25)
Current assignee of record (per Google Patents legal events): MediaTek, Inc. (Taiwan)

Sourcing note: All three reel/frame citations below are reproduced from the Google Patents legal-events record for this patent (which mirrors USPTO assignment data). The USPTO Assignment Center record itself exposes an additional field — the correspondent of record — which is not carried in the Google Patents mirror. I could not retrieve correspondent names for any of the three recordings from the sources available to me, and I will not guess at them. Retrieve the Assignment Center abstract for reel/frame entries 008075/0028, 015603/0829 and 015612/0472 to fill that field. Verify at: https://assignmentcenter.uspto.gov/ (search patent number 5802068).

Inventors

No unusual pattern. A single inventor who assigned to his employer of record before filing; there is no evidence of the inventor leaving the assignee within 12 months of filing, and no second inventor whose movements would matter.

Original assignee

Nippon Steel Corporation (Japan) — named as assignee on the face of the patent and as assignee in reel 008075/0028.

• Line of business: Integrated steel manufacturing, with a diversified electronics / advanced-materials operation in the 1990s (semiconductor materials and electronic devices). Nippon Steel was not a consumer audio/video or MPEG-products company.
• Did it ship a product embodying the claims? Unclear. Nothing in the patent record or the sources reviewed shows a Nippon Steel multiplexing product. The claim set (MPEG pack/padding multiplexing) reads on a media-authoring or encoder product; the absence of any Nippon Steel assertion or licensing program involving this patent, followed by an outright sale in 2004, is more consistent with an internal R&D portfolio asset than a commercialized one.
• Current status: Operating. Nippon Steel remains one of the world's largest steelmakers. It merged with Sumitomo Metal Industries in 2012 (Nippon Steel & Sumitomo Metal Corporation) and reverted to the Nippon Steel Corporation name in 2019. Not bankrupt — the 2004 divestiture of this patent is not a fire-sale artifact.

Assignment timeline

• 1996-06-03 (executed) / recorded 1996-06-25 — Reel 008075/0028

  • Conveyance: Assignment
  • Assignor: Kudo, Yoshinobu
  • Assignee: Nippon Steel Corporation (Japan)
  • Correspondent: Not retrievable from available sources — the Google Patents mirror omits the correspondent field. Cannot be evaluated for recurrence. (Confirm via Assignment Center.)
  • Context: Ordinary pre-filing employee-to-employer assignment of rights.

• 2004-11-24 (executed) / recorded 2005-01-19 — Reel 015603/0829

  • Conveyance: Assignment
  • Assignor: Nippon Steel Corporation
  • Assignee: United Module Corporation (recorded as a California corporation)
  • Correspondent: Not retrievable from available sources. Cannot be evaluated for recurrence.
  • Context: Divestiture by an operating steelmaker of a non-core electronics patent to a California patent-holding company. This is the only "shell-entity-shaped" link in the chain — see Signals 1 and 4.

• 2005-01-20 (executed) / recorded 2005-01-21 — Reel 015612/0472

  • Conveyance: Assignment
  • Assignor: United Module Corporation
  • Assignee: MediaTek, Inc. (Taiwan)
  • Correspondent: Not retrievable from available sources. Cannot be evaluated for recurrence.
  • Context: Sale to a large operating semiconductor company — i.e., United Module Corporation's hold lasted ~57 days, which is characteristic of a brokering/intermediary role rather than a licensing program built on this asset.

Chain of title is complete and unbroken: Nippon Steel → United Module Corp. → MediaTek, Inc. (current). No security agreements, mergers, change-of-name filings, or releases were found.

Downstream litigation (not itself an assignment, included for the timeline): MediaTek asserted US 5,802,068 in MediaTek Inc. v. Matsushita Electric Industrial Co. Ltd., E.D. Tex. No. 2:05-cv-00484 (filed 2005-10-17; closed 2007-06-14) — this is the "first worldwide family litigation" flagged in the Google Patents Darts-IP entry. A secondary source (National Taiwan University thesis, citing the case) confirms MediaTek counter-asserted three non-original patents — US 5,970,031, US 5,802,068, and US 6,118,486 — acquired respectively from United Microelectronics Corp., Nippon Steel, and Sarnoff Corporation.

Timeline diagram

timeline
    title Ownership of US 5802068
    1995 : Priority application filed in Japan
    1996 : Inventor Kudo assigns to Nippon Steel Corp
         : US application filed 25 Jun
    1998 : Patent US 5802068 issues
    2004 : Nippon Steel sells to United Module Corp
    2005 : United Module sells to MediaTek Inc
         : MediaTek asserts patent v Matsushita in EDTX
    2016 : Patent expires

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT (limited to one transient link).
Reel 015603/0829 (executed 2004-11-24) moved the patent from operating assignee Nippon Steel to United Module Corporation, recorded as a California corporation. Concrete, non-naming-based evidence that UMC was a licensing-only vehicle comes from later litigation: in Microchip Technology, Inc. v. United Module Corp., N.D. Cal. No. 5:10-cv-04241 (filed 2010-09-20) and the consolidated E.D. Tex. actions, UMC is described as headquartered in Los Altos, California, with Peter Courture as "the sole director, officer, shareholder, and/or employee"; UMC held expired patents it had acquired from WaferScale Integration, took no enforcement action for roughly a decade, and later entered an exclusive license authorizing Keranos, LLC (a Texas LLC formed 2010-02-10) to sue, with UMC taking "a portion of any recovery." That is a no-products, recovery-sharing assertion profile. Caveat: the name and state of incorporation match, but the assignment abstract for 015603/0829 does not itself identify Courture or Los Altos, so the identification of this UMC with the litigation UMC rests on name + state + timing rather than a single document. Also note the link was transient — UMC held the asset only 57 days.

2. Known asserter in the chain — NOT PRESENT.
Neither Nippon Steel, United Module Corporation, nor MediaTek, Inc. appears on the enumerated lists (Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, Spangenberg entities). United Module Corp. (with Keranos, LLC) does appear as a plaintiff in 2010–2011 E.D. Tex. assertion campaigns — but on different patents (the WaferScale '719, '629 and '009 patents), not on US 5,802,068. The only identified assertion of this patent was by MediaTek, an operating chipmaker.

3. Repeat correspondent across the chain — UNCLEAR / NOT ESTABLISHED.
This is the single most informative field for this chain and it is the one field I could not obtain. No correspondent name appears in the Google Patents legal-events mirror for reels 008075/0028, 015603/0829 or 015612/0472. Because there are only three recordings and one of them is a pre-filing employee assignment, a single recurring filing attorney across the 2004 and 2005 links would materially raise the "same lawyer, rotating shells" concern — this must be checked in Assignment Center before the NPE call is finalized. I decline to name a correspondent I have not seen.

4. Cascading transfers — PRESENT (weak-to-moderate).
Two consecutive assignments covering the same asset executed 2004-11-24 (reel 015603/0829) and 2005-01-20 (reel 015612/0472) — a 57-day interval, comfortably inside the <24-month test. The pattern (operating company → no-product holding company → operating company, with the intermediary holding for under two months) is the signature of a brokered asset sale. However, this fails part of the classic test: the second assignee is a large publicly traded operating company, not a second chained LLC, and I found no evidence that the two assignees shared a correspondent address or common principals.

5. Pre-litigation transfer — NOT PRESENT.
Assignment to MediaTek executed 2005-01-20; the first identified suit naming this patent, E.D. Tex. 2:05-cv-00484, was filed 2005-10-17 — approximately 9 months, outside the 6-month window. Note the limitation: I could not determine the precise date MediaTek's '068 counterclaim was served within that case, so if the assertion pleading post-dates the case filing, the gap is even wider, not narrower.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 proceeding involving Nippon Steel or United Module Corporation appears in the record. Nippon Steel was and remains a solvent operating company; the 2004 transfer reads as a portfolio divestiture, not a distressed sale.

7. Privateering — NOT PRESENT.
There is no evidence Nippon Steel used an NPE to assert against its own competitors. Nippon Steel sold the asset outright in 2004 and had no role in the 2005 MediaTek–Matsushita dispute.

8. Defensive aggregator (anti-NPE) — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at MediaTek, a product company, which then asserted it offensively.

Verdict

Operating-company assertion.

The chain terminates at MediaTek, Inc., an operating fabless semiconductor company, per reel 015612/0472 (executed 2005-01-20, recorded 2005-01-21), and MediaTek affirmatively asserted US 5,802,068 against Matsushita Electric Industrial Co. in E.D. Tex. 2:05-cv-00484 (filed 2005-10-17) — a suit against an actual competitor, not a licensing shakedown. The one genuine NPE-flavored feature is the transient intermediary: reel 015603/0829 (executed 2004-11-24) put the patent in United Module Corporation, a California entity that later litigation describes as having a single officer/shareholder, no products, and a recovery-sharing assertion arrangement — but UMC held for only 57 days and never asserted this patent itself, so the asset's operative use was a product-company assertion, not an NPE campaign.

Confidence caveat: the verdict rests on the terminal owner, and it is the terminal owner rather than the intermediate link that drives it. If Assignment Center shows the same correspondent attorney or firm filed both the 2004 and 2005 recordings (Signal 3, currently unclear), or shows a United Module Corporation address that is a registered-agent service, the UMC link would strengthen to a second independent signal and the call should be revisited toward NPE — moderate confidence for the 2004–2005 window. Confirm at https://assignmentcenter.uspto.gov/ (patent 5802068; reels 008075/0028, 015603/0829, 015612/0472) and cross-check the E.D. Tex. 2:05-cv-00484 docket for the counterclaim service date.

Primary verification links: https://patents.google.com/patent/US5802068/en · https://assignmentcenter.uspto.gov/ · https://assignment.uspto.gov/patent/index.html

I'll research this systematically. Let me start by verifying the patent and its cited prior art.

I have the core data. Let me verify a couple of bibliographic details on the least-documented reference and confirm nothing contradicts the record.

Prior Art Analysis — US 5802068

1. Patent under review (identified literally as requested)

Note on scope: The four independent claims are cast at four different levels of abstraction — claim 1 ("set assemblies"/"set parts" plus "redundancy data set part"), claim 4 (calculate → adaptively vary the second set part length → form a set assembly with a redundancy data set part), claim 6 (audio packet length calculating means + varying means + pack forming means using a redundancy data packet), and claim 15 (MPEG-specific, "packs"/"packets"/"padding"). Claims 7–14 add the specific parameters (bit rate, sampling frequency, encoding layer, pack length), frame‑based sizing, padding‑length calculation, frame‑number deciding means, and source selecting means.

2. The cited prior art (References Cited on the face of US 5,802,068)

The patent lists exactly four U.S. patent citations. There are no foreign patent documents or non‑patent literature listed in the "References Cited" section. The four references are set out below.

⚠️ Framing caveat before the analysis: These are the references the examiner cited and considered. Because US 5,802,068 issued with all 15 claims intact, the examiner did not treat any of them as anticipating. My §102 assessment below therefore identifies where each reference comes closest and where an anticipation theory would succeed or fail — not a legal conclusion that any reference actually invalidates a claim. I flag explicitly where I am inferring rather than reading from the record.

Reference A — US 3,838,221 A

• Full citation: US 3,838,221 A, "TDMA satellite communications system having special reference bursts," inventors William G. Schmidt, Ova G. Gabbard, John M. Husted, Wilfrid G. Maillet, and Kunishi Nosaka; assignees Communications Satellite Corporation and Kokusai Denshin Denwa Co., Ltd.; App. No. 170,796; filed 1971‑08‑11; issued 1974‑09‑24; 4 claims.
• Description: A satellite time‑division multiple access system in which each earth station transmits data in bursts synchronized to a special "reference burst" that carries no data communications (only a reference unique word / preamble). Terrestrial interface modules receive data in different forms and rates, convert it to TDMA‑compatible bit form, store compressed blocks, and hold them until a multiplexer requests them for inclusion in a burst. The arrangement and timing of data blocks within a burst is controlled by words stored in memory, allowing reordering.
• Closest claims / §102 analysis:
  • Potentially superficially relevant to claim 1 because it shows (i) a "set assembly" (a burst) made up of parts, and (ii) a non‑data "reference burst" that is arguably a "redundancy data set part." A §102 argument would target the "plurality of set assemblies … redundancy data set part" language.
  • Why it fails: The reference burst is a synchronization marker occupying its own time slot — it is not concatenated into the same assembly as payload data to pad that assembly up to a maximum length. Claim 1 requires the redundancy data length to equal "a difference between the maximum data length of said second set part and the data length of said second data" — a relative‑length computation the reference does not perform. No teaching of a high‑bit‑rate first data and low‑bit‑rate second data being co‑packaged with padding.
  • Best characterization: §102 anticipation is not supportable; it is background art on multiplexing heterogeneous‑rate data into a burst and on non‑data filler, relevant to §103 context at most.

Reference B — US 4,224,473 A

• Full citation: US 4,224,473 A, "TDMA Multiplexer‑demultiplexer with multiple ports," filed 1978‑05‑31; issued 1980‑09‑23. (Assignee is listed in the Google Patents record as Digital Communications Corporation; I did not independently confirm the inventor/assignee line, so treat that field as unverified.)
• Description: A multiplexer/demultiplexer for a TDMA earth terminal that accepts plural asynchronous terrestrial inputs at various bit rates and converts them into high‑rate bursts (and back). Each input port has an elastic buffer written at the incoming rate and read at a common first rate; a compression buffer is then read at the high burst rate, and a gating device selectively gates reads at selected times. It expressly addresses bit‑rate conversion, elastic buffering, and pulse stuffing so that data of differing rates can share a common transmission format.
• Closest claims / §102 analysis:
  • Superficially touches claim 1 / claim 4 themes: multiplexing data of different bit rates, buffer management, and using added/removed bits ("pulse stuffing") to reconcile rates.
  • Why it fails: It is a rate‑conversion and buffering system for continuous TDMA input streams. It does not disclose dividing high‑rate first data and low‑rate second data into packets/parts, does not compute a maximum data length of a second set part, and does not build an assembly whose remaining space is filled by a redundancy part of length = (max − actual). The "stuffing" it references is asynchronous rate justification, not packet‑payload padding to a fixed pack length.
  • Best characterization: Background/§103 art on multi‑rate multiplexing and elastic buffering; not anticipatory.

Reference C — US 4,864,562 A

• Full citation: US 4,864,562 A, "Sub‑rate multi‑media data transmission control system," inventors Tokumichi Murakami and Koh Kamizawa; assignee Mitsubishi Denki Kabushiki Kaisha; priority JP 1986‑12‑18; filed 1987‑12‑17; issued 1989‑09‑05; Expired.
• Description: A system that multiplexes motion video, voice, and external digital data onto a transmission line. It sets a basic transmission frame of N1×8 bits at an 8 kHz cycle, uses a sub‑frame of N2×8 bits at the variable transmission speed, and — critically — provides a "dummy subframe" of (N1 − N2)×8 bits to fill the unused portion of the frame. It uses a multiframe to integrate voice‑frame, error‑correction‑frame and video‑packet synchronization, with the stated goal of minimizing the buffer memories for motion video and smoothing the transmission speed through simple control.
• Closest claims / §102 analysis:
  • This is the most conceptually on‑point of the three older references, and it is likely why the examiner cited it. Two features map onto the subject patent:
    1. Dummy/filler data sized as a difference — "(N1 − N2)×8 bits" is functionally analogous to the claimed "redundancy data of a data length equal to a difference between the maximum data length … and the data length of second data" (claim 1 / claim 15) and the "redundancy data set part corresponding to the data length of a varied portion" (claim 4).
    2. The same underlying problem — buffer‑memory reduction and synchronization of a high‑rate signal (video) with a lower‑rate signal (voice), which is precisely the "initial delay"/buffer‑capacity problem the subject patent recites.
  • Why a §102 rejection still fails:
    • The dummy subframe is defined by the fixed system parameters N1 and N2, not by a per‑transmission computation of an audio‑packet data length from bit rate, sampling frequency, encoding layer and pack length (claims 6–8) nor by a "maximum data length of a second set part" in an MPEG pack sense (claim 15).
    • The filler is a modulation‑frame sub‑slot, not a separate packet/pack added alongside a data packet inside a packetized "pack." Claim 6/15 require "pack forming means … a redundancy data packet," a packet architecture the reference lacks.
    • No "frame number deciding means … to improve multiplexing efficiency when a bit rate is changed from a middle bit rate to a high bit rate" (claims 10–13).
  • Best characterization: Strong §103 reference (and the closest art for the "filler‑data sized as a difference" idea and for the buffer/delay motivation); not a clean §102 anticipation of any claim.

Reference D — US 5,650,825 A

• Full citation: US 5,650,825 A, "Method and apparatus for sending private data instead of stuffing bits in an MPEG bit stream," inventors Saiprasad V. Naimpally and Ren Egawa; assignee Matsushita Electric Corporation of America; App. No. 08/414,526; filed 1995‑03‑31 (some databases list priority as 1995‑03‑30); issued 1997‑07‑22. Family: EP 0 805 598 B1.
• Description: In an MPEG transport stream, "stuffing" bytes fill a transport packet when there is insufficient PES data to reach the required bit rate. The invention analyzes a data packet to determine whether stuffing bytes are present and whether the packet is eligible, then re‑multiplexes by removing the stuffing bytes (from the adaptation field in the header and/or from the payload) and inserting useful "privatestuff" data in their place — recouping otherwise wasted bandwidth. Figure 3 shows the transport‑packet header/adaptation‑field/stuffing‑byte layout.
• Closest claims / §102 analysis:
  • This is the closest reference in time, subject matter, and vocabulary: both patents are MPEG‑packet stream processing patents dealing with filling fixed‑size packets/packs and with "stuffing"/"padding" data.
  • Potential §102 hooks: the "analyzing means … whether stuffing bytes are being used" maps loosely onto claim 1's "means for detecting whether a data length of said second data is smaller than a maximum data length of said second set part"; and the stuffing/privatestuff exchange maps onto the "redundancy data" (claim 1/15) / "redundancy data packet" (claim 6) concepts.
  • Why anticipation fails — and it fails on the inventive core:
    1. Opposite operation. US 5,650,825 removes stuffing and replaces it with useful private data; the subject patent adds a padding/redundancy part whose sole specified length equals (max − actual) to complete a pack. The two run in opposite directions on the same field.
    2. No adaptive packet‑length calculation. US 5,650,825 contains nothing corresponding to claims 4, 6, 7, 8 — no calculating a second/audio packet data length from bit rate, sampling frequency, encoding layer, or pack length, and no "varying means" that adaptively varies that length. Its "analyzing" step is a binary eligibility check for existing stuffing, not a length computation.
    3. No "difference‑of‑lengths" redundancy sizing. Claim 1/15 require redundancy length = maximum data length − data length; US 5,650,825 sizes privatestuff to whatever stuffing already exists, not as a computed difference to a maximum.
    4. No frame‑number deciding means (claims 10–13) and no source selecting means among video/audio/padding (claim 14).
  • Best characterization: The most relevant §102 candidate by subject matter (MPEG, packets, padding/stuffing), but it does not anticipate any claim because it neither computes/varies an audio packet length nor sizes a redundancy part as a max−actual difference. Its real value to a challenger is as a §103 reference combined with the buffer/delay motivation.

3. Consolidated matrix

Which claims each reference most directly bears on (for a §102/§103 chart):

• Claim 1 — A (reference burst as non‑data part) and C (dummy subframe) are the closest on the "redundancy data set part of a difference length"; D on the "detecting / stuffing" language.
• Claim 4 — C (adaptively matching frame fill to transmission speed) is closest; but the "calculating … data length … in accordance with a predetermined parameter" limitation is untaught by all four.
• Claim 6 — D (MPEG pack/packet with filler) is closest on architecture; A, B, C none. The "audio packet data length calculating means" and "varying means" are untaught by all four.
• Claims 7, 8 (bit‑rate/sampling/layer/pack‑length parameters; frame‑as‑unit) — no reference discloses these; they are the strongest differentiators.
• Claims 9, 14 (padding‑length calculating means; source selecting among video/audio/padding) — no reference discloses these.
• Claims 10–13 (frame‑number deciding to improve efficiency across a middle‑to‑high bit‑rate change) — no reference discloses this.
• Claim 15 — D for MPEG/padding vocabulary; C for difference‑sized filler; both fail on the audio‑packet data‑length detection/computation.

4. Bottom line and caveats

1. The examiner cited only four references — all U.S. patents; there is no foreign or NPL citation on the face of US 5,802,068. Two (US 3,838,221; US 4,224,473) are TDMA‑satellite/telephony background art, one (US 4,864,562) is the closest art for the "filler sized as a difference" concept and the buffer‑delay motivation, and one (US 5,650,825) is the closest in subject matter (MPEG packet stuffing/padding) but teaches the opposite operation.

2. None of the four supports a defensible §102 anticipation of any of claims 1–15 on the record as published. Anticipation would require a single reference disclosing (i) computing/varying a low‑rate (audio) packet data length from parameters such as bit rate, sampling frequency, encoding layer and pack length, and (ii) forming an assembly with a redundancy part whose length equals maximum − actual. No cited reference contains both elements. The patent's strongest distinguishing limitations are claims 7, 8, 9, 10–13, and 14, which appear wholly unaddressed by the cited art.

3. Confidence and verifications: The bibliographic data for US 5,802,068 and for US 3,838,221, US 4,864,562, and US 5,650,825 are corroborated by the sources retrieved. I could not independently confirm the inventor/assignee of US 4,224,473 beyond the "Digital Communications Corporation" attribution shown in the Google Patents record; treat that specific field as unverified. Note also a minor priority‑date discrepancy for US 5,650,825 (1995‑03‑31 on the U.S. front page vs. 1995‑03‑30 in one commercial database) — this does not affect any conclusion here, since either date precedes the subject patent's 1995‑06‑30 priority date.

4. Litigation/ownership context (relevant to why prior art matters): The record shows first worldwide family litigation, a Texas Eastern District Court case (2:05‑cv‑00484), and assignment of the patent through United Module Corporation to MediaTek, Inc. (2005). Any invalidity challenge in that or later litigation would most plausibly run through US 5,650,825 and US 4,864,562 in a §103 combination, not through a single‑reference §102 attack.

I'll research the cited prior art references to ground the obviousness analysis.

Obviousness Analysis — U.S. Patent No. 5,802,068 under 35 U.S.C. § 103

Technical analyst work product. This is an engineering-oriented invalidity discussion, not a legal opinion. A formal validity determination requires claim construction, the full text/file histories of every reference, and any additional art of record.

1. Analytical framework

Because US 5,802,068 ("the '068 patent") has a priority date of 1995‑06‑30 (JP 7‑188251) and was filed 1996‑06‑25, it is governed by pre‑AIA § 103(a). The controlling test is Graham v. John Deere (scope/content of the prior art; differences; PHOSITA level; secondary considerations), as refined by KSR Int'l v. Teleflex (2007), and operationalized in MPEP §§ 2141–2144.1. Key KSR principles applied below:

• MPEP 2144.1(a) — combining known elements according to known methods to yield predictable results.
• MPEP 2144.1(f) / KSR — a technique used to improve one device, where a POSITA would recognize it would improve a similar device in the same way.
• MPEP 2143.03 / In re Boesch, In re Applied Materials — optimizing a "result‑effective variable" through routine experimentation.
• MPEP 2145 — the cited art need only be "reasonably pertinent" to the inventor's problem.

PHOSITA profile

A person of ordinary skill in the art circa mid‑1995 would have: (i) a bachelor's degree in EE/CS (or equivalent) plus 2–3 years' experience in digital audio/video compression and transport; (ii) working knowledge of the MPEG‑1/MPEG‑2 Systems layer (packs, packets, PES, padding/stuffing packets, adaptation fields, PCR/SCR time stamps); and (iii) knowledge of conventional time‑division multiplexing and buffer/rate‑smoothing techniques for constant‑bit‑rate transport of variable‑bit‑rate media.

2. The claimed subject matter (independent claims)

Dependent claims add: MPEG/video/audio/padding characterization (2, 3, 5); parameter set of bit rate, sampling frequency, encoding layer, pack length (7); one frame as the unit (8); redundancy packet length calculating means (9); frame‑number deciding means to improve efficiency when bit rate changes from middle → high (10–13); and source selecting means (14).

The admitted starting point: The '068 specification itself admits the prior art of MPEG packet/pack multiplexing and frames the problem as initial delay and input buffer size when a high‑rate video stream and a low‑rate audio stream are synchronously multiplexed (col. discussing "about eight frames ≈ 288 msec" per audio packet and the resulting "initial delay").

3. The cited prior art (patent‑citations section of US 5,802,068)

All four references are of record on the face of the patent (source: https://patents.google.com/patent/US5802068/en, "Patent Citations (4)"):

Key observation: References '562, '443, and '821 are not MPEG‑specific, but each is reasonably pertinent under MPEP 2145 to the '068 problem — multiplexing data sources of different bit rates into a common frame/burst while minimizing buffer and delay. US 5,650,825 supplies the MPEG‑specific packet/padding context.

4. Claim‑element mapping and proposed combinations

Ground 1 — Claims 1, 2, 3, and 15: US 5,650,825 in view of US 4,864,562 (optionally with US 4,224,473)

Why the combination is proper (motivation):

1. Same field / same problem. Both references concern time‑division multiplexing of multiple digital media of unequal rates over a constant‑rate channel. '562's stated object — "minimiz[ing] the size of the buffer memories … and … the transmission speed smoothing operation" — is precisely the '068 object (reduce initial delay and input buffer capacity).
2. Complementary teachings, no conflict. '825 supplies the MPEG pack/packet + padding framework the '068 claims specifically invoke (claims 2 and 15 recite "MPEG"); '562 supplies the detect‑and‑fill‑with‑dummy‑data mechanism claim 1 recites generically. A POSITA implementing an MPEG packer who needed to rate‑match a low‑rate audio packet into a fixed‑size pack would naturally adopt '562's dummy‑subframe fill — an "improvement of one device using a known technique that would improve a similar device in the same way" (KSR).
3. Reasonable expectation of success. Both are deterministic digital framing schemes; sizing a filler to (max − actual) is arithmetic, not experimentation.

Ground 2 — Claims 4 and 5: US 4,864,562 in view of US 5,650,825

This is the cleanest mapping: claim 4 is essentially '562's variable‑bit‑allocation‑plus‑dummy‑fill transposed from a H.130‑style transmission frame onto an MPEG pack, a substitution of one known multiplexing frame format for another that a POSITA would recognize as equivalent.

Ground 3 — Claims 6–14: US 4,864,562 in view of US 5,650,825 (optionally with US 4,224,473)

For claims 9 and 14, US 4,224,473 adds an independent, well‑known rationale: pulse‑stuffing and elastic/compression buffering to reconcile asynchronous inputs of various bit rates, then accumulating and reading out at burst rate — i.e., the buffering and rate‑reconciliation architecture the '068 specification describes (video/audio buffer units 3, 4).

5. Consolidated motivation‑to‑combine statement

Under MPEP 2144.1 and KSR, the four references may properly be combined because:

1. Same field of endeavor — all four are digital multiplexing of multiple media/data streams over a framed/bursted channel.
2. Same problem — handling rate mismatch between sources without excess buffering/delay; '562 and '443 state this objective expressly; the '068 specification admits this was the art's known problem.
3. Predictable combination of known elements — (a) a packer/formatter that pads a short, low‑rate payload with filler to reach a fixed length, and (b) a means to compute the filler length and select whether to emit it, are individually known ('562 dummy subframes, '825 stuffing bytes) and combine with predictable results.
4. Design choice / optimization of a result‑effective variable — selecting the number N of audio frames per packet (claims 8, 10–13) is optimizing a variable (latency vs. multiplexing efficiency) that the art already identified; In re Boesch / In re Applied Materials.

6. Secondary considerations (objective indicia)

The '068 specification asserts (i) initial delay reduced to ≤ 40 % of the conventional value, and (ii) input buffer capacity reduced (e.g., ~16.2 kbytes video buffer with N = 3 frames, 108 ms delay). Under MPEP 716.02, these must be commensurate with the claimed invention and unexpected to overcome a prima facie case of obviousness. The likely rebuttal analysis:

• Commensurate? The delay/buffer benefit flows directly from reducing the audio frame count N and filling the remainder with padding — the very acts disclosed by '562 (dummy subframes, buffer minimization). The gains are therefore expected consequences of the combination, not new functional effects.
• Unexpected? No teaching‑away appears in '562 or '443; both pursue buffer/delay reduction. A patentee's only meaningful teaching‑away argument is that '825 disparages stuffing as "a waste of system resources." That is a weak teaching‑away: '825 criticizes wasting bandwidth when useful data is available, and expressly confirms stuffing was standard MPEG practice for rate matching — i.e., it bolsters the knowledge of the padding mechanism rather than directing away from proportional use of padding for decoupling the streams.

7. Likely battlegrounds / where the prima facie case is weakest

1. "Means for detecting … smaller than a maximum data length." '825's analyzer tests for stuffing presence, not literally for "audio length < max second‑packet length." A patentee may argue the analogical reading is too loose; expect reliance on '562's explicit N₂×8 vs. N₁×8 comparison to close the gap.
2. The "middle‑bit‑rate" frame‑number decision (claim 10). The rationale "so as to improve a multiplexing efficiency when the bit rate is changed from a middle bit rate to a high bit rate" is '068‑specific. '562's "adaptive" bit allocation is broader; the combination may require declaring it an obvious optimization (Boesch).
3. Claim construction of the "means" terms. Claims 1, 4, 6, and 15 recite means‑plus‑function language; under § 112(f)/Williamson, the scope and corresponding structures (units 5, 51, 6, 61, 7, 8, 81) must be mapped, which could narrow or broaden the art mapping.

8. Bottom line

On the four references of record, all independent claims (1, 4, 6, 15) and their dependents appear susceptible to a § 103 rejection under a KSR/MPEP 2144.1 combination:

• Primary combination: US 5,650,825 + US 4,864,562 (with US 4,224,473 as a secondary reference for the multi‑bit‑rate buffering/stuffing architecture, and US 3,838,221 for framed burst assembly). '825 supplies the MPEG pack/packet + padding/stuffing framework and the detect‑then‑fill logic; '562 supplies the variable‑length second set part plus dummy/redundancy fill sized to the difference and the express buffer/delay‑reduction motivation.
• The strongest, most defensible single combination is '825 as primary + '562 for the varying/dummy‑fill disclosure, because together they disclose every element of claims 1 and 4 (and, by extension, 15) with a clear, same‑problem motivation.
• The principal vulnerability of the rejection is the specificity of claims 10–13 (middle‑bit‑rate frame‑number optimization) and the § 112(f) construction of the "means" limitations. A patentee's best counter is non‑commensurate objective evidence of reduced initial delay/buffer size; that argument is weakened because those benefits are the predictable result of adopting the '562 dummy‑fill/short‑packet approach.

Caveat: I have worked from the abstracts, claims, and definitions surfaced for the four citations of record (US5650825A, US4864562A, US4224473A, US3838221A) plus the '068 disclosure. The examiner‑cited art is not necessarily the best art, and a definitive conclusion would require the full text of each reference, the '068 file history, and claim construction — none of which is available to me here. Where I inferred an element from an abstract rather than an express passage, I have flagged it; anything beyond that should be verified against the full reference documents.