US 8812327

I'll search for authoritative information on this patent number.

Let me try alternate search phrasings.

Now let me check for any CAFC 2026 litigation involving this patent.

US Patent 8,812,327 — Analyst Summary

Search note. My live searches located the patent on Google Patents, FreePatentsOnline (FPO) and Justia. I found no Federal Circuit 2026 docket, appeal, or CAFC opinion referencing U.S. 8,812,327 (or "8812327") — searches returned empty/irrelevant hits. I cannot confirm or deny 2026 CAFC activity; treat that as unverified rather than as a negative finding. Everything below comes from the authoritative full patent text plus those patent-database hits.

Bibliographic data (literal, as published)

Abstract (verbatim)

A method of hierarchical coding of a digital audio frequency input signal into several frequency sub-bands, including a core coding of the input signal according to a first throughput and at least one enhancement coding of higher throughput, of a residual signal. The core coding uses a binary allocation according to an energy criterion. The method includes for the enhancement coding: calculating a frequency-based masking threshold for at least part of the frequency bands processed by the enhancement coding; determining a perceptual importance per frequency sub-band as a function of the masking threshold and as a function of the number of bits allocated for the core coding; binary allocation of bits in the frequency sub-bands processed by the enhancement coding, as a function of the perceptual importance determined; and coding the residual signal according to the bit allocation. Also provided are a decoding method, a coder and a decoder.

Plain-language overview of the independent claims

The patent is about scalable ("hierarchical") audio coding — a base/core layer plus "improvement" (enhancement) layers — where the core layer allocates bits purely on an energy criterion, and the invention adds a psychoacoustic masking model into the improvement layer so bits are allocated perceptually, without changing the core layer (preserving G.729.1 interoperability).

Claim 1 — Hierarchical coding method.
Core-code the input in a low frequency band at a first bit rate using a first binary bit allocation based on an energy criterion. Then, for at least one higher-bit-rate improvement coding of a residual signal in a high frequency band, do four things:

1. compute a frequency masking threshold for at least part of the bands handled by the improvement coding, where the masking threshold is normalized by its value at the last sub-band of the low band and/or the first sub-band of the high band (this normalization clause is recited in the granted claim);
2. determine a perceptual importance per sub-band of the high band as a function of (i) the calculated masking threshold and (ii) the number of bits allocated for the core coding;
3. perform a second binary bit allocation in those high-band sub-bands based on that perceptual importance; and
4. code the residual signal according to that second allocation.

Claim 7 — Hierarchical decoding method. Mirror image of claim 1: core-decode the received signal in a low band at a first bit rate with an energy-criterion first allocation; then, in an improvement decoding of a higher bit rate in a high band, calculate a masking threshold (normalized at the last low-band sub-band and/or first high-band sub-band), determine perceptual importance per high-band sub-band from the masking threshold and the core-decoding bit count, apply a second bit allocation, and decode the residual signal accordingly.

Claim 9 — Hierarchical coder (apparatus). A memory storing code instructions plus a processor configured by them to implement: a core coder (first bit rate, low band, energy-criterion first allocation) and an improvement coder (higher bit rate, residual signal, high band) comprising modules that (a) calculate the normalized frequency masking threshold, (b) determine perceptual importance per high-band sub-band from the masking threshold and core-coder bits, (c) apply a second binary bit allocation, and (d) code the residual signal per that allocation.

Claim 10 — Hierarchical decoder (apparatus). Same processor/memory framing, but for a core decoder and an improvement decoder performing the corresponding masking-threshold, perceptual-importance, second-allocation, and residual-decoding operations.

Claim 11 — Non-transitory computer-readable medium storing a program whose instructions, when executed by a processor, implement the coding method of claim 1 (including the normalization-by-value-at-last-low-band-sub-band-and/or-first-high-band-sub-band limitation).

Claim 12 — Non-transitory computer-readable medium storing a program implementing the decoding method of the claim-7 type.

Dependent claims (for completeness): 2 and 8 (two-part perceptual-importance derivation: define a first perceptual importance from masking threshold, quantized spectral-envelope values and a normalization factor, then subtract the ratio of core bits to sub-band coefficients); 3 (perceptual importance also a function of bits allocated in previous coding stages using energy-criterion allocation); 4 (masking threshold obtained by convolution of a spectral-envelope expression with a spreading function involving the sub-band's central frequency); 5 (derive a tonal/non-tonal indication; apply the masking-threshold and perceptual-importance steps only if the signal is non-tonal); 6 (improvement coding of TDAC type in an extended coder whose core is G.729.1).

Implementation context disclosed (not claim language)

• The improvement layer is applied to the TDAC (Time Domain Aliasing Cancellation) transform-coding stage of a G.729.1 core; a specific worked example adds two 8 kbit/s layers shifting 48 → 56 → 64 kbit/s in the 50–7000 Hz band.
• An advantageous spreading function is triangular: +27 dB/Bark on one slope and −10 dB/Bark on the other (FIG. 7); the masking threshold is applied to the high band (≈4–7 kHz) and normalized for spectral continuity at the 4 kHz junction (FIGS. 7–8).
• Perceptual importance follows the G.729.1 form for the low band and adds a log_mask(j) term for the high band, with the core-layer bits subtracted as nbit(j)/nb_coeff(j).
• Tonality detection opportunistically reuses the G.729.1 envelope-coding mode bit ("differential Huffman" vs. "direct natural binary").

Discrepancies / explicit uncertainties

1. Claim text changed between publication and grant. The pre-grant publication US 2012/0185255 A1 recites the claims in a broader form (no "low frequency band"/"high frequency band" split and no masking-threshold normalization clause), whereas the granted US 8,812,327 B2 claims I was given recite those added limitations. I did not retrieve the prosecution history or any certificate of correction / reexamination record, so I cannot authoritatively explain why the claims differ. Do not treat my description of the change as a legal conclusion.
2. Expiration date. Google Patents shows "Active, expires 2030-11-15." A 20-year term from the 2010-06-25 filing would ordinarily run to 2030-06-25, so the listed 2030-11-15 is consistent with a patent-term adjustment. I am reporting the database value literally and have not independently verified the PTA calculation.
3. CAFC 2026 dockets. As noted above, no 2026 Federal Circuit docket or opinion naming 8,812,327 surfaced in my searches. I have no authoritative litigation information for 2026.
4. Assignee naming. The record reflects "France Telecom SA" as original assignee and "Orange SA" as current assignee (Google Patents also flags that assignee listings "may be inaccurate" and disclaims having performed a legal analysis).

Sources: https://patents.google.com/patent/US8812327/en (and the #1/#6 fragment views) · https://patents.google.com/patent/US20120185255 (pre-grant publication) · https://www.freepatentsonline.com/8812327.html · https://patents.justia.com/patent/8812327

I'll search for litigation involving US patent 8812327.

Let me search more specific litigation databases.

Let me try more targeted searches on litigation databases and Docket Alarm.

Litigation Search Results: US Patent 8,812,327

Bottom line: I found no litigation involving US Patent 8,812,327. Based on my searches of patent-litigation-oriented sources, I could not identify any district court case, Federal Circuit appeal, or PTAB proceeding asserting or challenging this specific patent.

What I searched

• General web searches for "US patent 8812327 litigation" and "8812327 patent lawsuit defendant"
• Unified Patents portal/litigation references
• Docket Alarm references
• Party-based searches tied to the patent owner (Orange SA / France Telecom)

What the results showed

1. No hits on the patent number in any litigation context. The searches that returned the literal string "8812327" produced only unrelated items — e.g., a Nevada state-court deposition transcript transaction number, a Brazilian corporate-registry filing, a German safety data sheet article number, and a copper-shaving product number. None relate to US Patent 8,812,327.

2. The results that addressed patent litigation generally (Unified Patents material, Docket Alarm, various PTAB documents) concerned entirely different patents (e.g., U.S. 8,605,794; 8,767,824; 9,928,275; 7,454,002) and did not mention 8,812,327.

3. Matches involving "Orange" / "France Telecom" litigation were French competition/antitrust and GPL-license disputes (e.g., Digicel, SFR, Entr'Ouvert), plus European FRAND cases brought by Intellectual Ventures against Orange on unrelated patents (Paris, RG 17/3837 and 17/3838; EP patents). None of these involve US 8,812,327.

Patent background confirming the lack of assertion activity

From the patent's own record (Google Patents, fetched 2026-09-24):

• Patent: US 8,812,327 B2, "Coding/decoding of digital audio signals"
• Inventors: David Virette, Stéphane Ragot, Balazs Kovesi, Pierre Berthet
• Original assignee: France Telecom SA; current assignee: Orange SA
• Priority date: 2009-07-07; filed: 2010-06-25; granted: 2014-08-19
• Status: Active; adjusted expiration 2030-11-15
• Family: EP2452336B1, KR101698371B1, CN102576536B, CA2766864C, FR2947944A1, WO2011004097A1
• Cited By (1): only a 2020 Tencent application (US12002484B2) — a forward citation, not a litigation reference
• Citations (18) and Family Cites (4): prior-art references only

The absence of any "Cited By" litigation-type entries and the fact that the patent is held by Orange (an operating telecommunications company rather than a known non-practicing entity) is consistent with a patent that has not been asserted in court.

Important caveats

I cannot state this with absolute certainty. My searches relied on general web results and did not give me direct, live access to the specific litigation databases that would provide a definitive answer. Specifically:

• I could not directly query the Unified Patents litigation case list, Docket Alarm, or PACER/RECAP docket for this patent number, so a case that exists but is not well-indexed by general search engines could have been missed.
• The literal-string searches for "8812327" were heavily polluted by unrelated filings, so a genuine but low-visibility docket entry could be obscured.

For a definitive answer, you should run a direct docket search on a litigation database (e.g., Docket Alarm, Unified Patents Portal → Litigation → Case List, CourtListener/RECAP, or PACER) using the patent number 8,812,327, and cross-check the patent's owner Orange SA as plaintiff. If you can grant access to those tools or share specific docket results, I can analyze them.

Summary table:

If no litigation exists, no case can be listed — and that is the result my searches support.

I'll verify the structured data against live sources before drafting.

The first pass returned unrelated matters. Let me run more targeted queries.

Verification note

I ran the canonical structured block ("PTAB proceedings on file") against live sources. The structured block — sourced from the USPTO Open Data Portal — reports no AIA trial proceedings on US 8,812,327. My independent web checks (queries combining the patent number with "IPR," "PTAB," "Inter Partes Review," "petition," and the assignee names Orange SA / France Télécom, plus a litigation-defendant query) returned zero petitions, institution decisions, final written decisions, or Federal Circuit appeals directed to this patent. The only hits were unrelated matters that happened to contain the string "IPR" or the digits "8812327" (a Brazilian corporate registry entry, a lanthanide precursor patent, an optical-fiber IPR). I could not query PTAB E2E directly in this session; the ODP structured feed is the authoritative source here and is not contradicted by anything I found.

Proceedings overview

Total AIA trial proceedings on US 8,812,327: 0. There are no IPRs, no PGRs, and no CBMs — hence no breakdown into active / claims-invalidated / claims-sustained / settled / institution-denied buckets. The bottom line for a defendant: every claim of 8,812,327 is UNTESTED at the PTAB and therefore all 12 claims remain presumptively valid and enforceable. This is not a "hardened patent that survived two IPRs" scenario, and it is emphatically not a "claims 1-5 have been canceled" scenario — no claim of this patent has ever been canceled, narrowed, or even construed by the Board. A defendant facing assertion today gets no free IPR estoppel and no ready-made invalidity record; the validity fight starts from scratch, and it starts on the defendant's dime.

Because no proceeding exists, the per-proceeding sections below are omitted rather than fabricated. The substitute analysis follows.

Patent status snapshot (the facts a defendant actually needs)

Claim structure matters defensively. Every independent claim (1, 7, 9, 10, 11, 12) recites the same narrowing hook: the masking threshold "being normalized by the value of the masking threshold at a last sub-band of the low frequency band and/or a first sub-band of the high frequency band." That "and/or" normalization limitation is the point of novelty over the G.729.1 baseline and is recited identically across all six independent claims. It is the single most productive non-infringement and § 103 battleground, and it appears in the claims as granted — meaning any prior-art reference that teaches masking-threshold computation without sub-band-boundary normalization likely does not reach the claims.

Strategic summary

What is canceled vs. sustained vs. untested. Nothing is canceled; nothing is sustained. All 12 claims are untested. For a defendant, that means the full claim set — including the broad-looking method claims 1 and 7 and the system claims 9 and 10 — is live and carries the § 282 presumption of validity. There is no PTAB record to cite, no claim-construction ruling from the Board, and no Federal Circuit mandate narrowing the claims. Any invalidity position must be built and funded from zero.

Estoppel landscape. There is no § 315(e)(2) estoppel on anyone. No petitioner has ever been before the Board on this patent, so no party — and no privy of any party — is barred from raising any ground. The complete prior-art universe is available. Note, however, that the patent's effective filing date (2009-07-07) places it pre-AIA (before 2013-03-16). Consequences:

• PGR is unavailable by statute (PGR reaches only patents with an effective filing date on or after 2013-03-16). Do not let anyone tell you otherwise.
• CBM is unavailable — the transitional CBM program sunset on 2020-09-20, and in any event this is a technological audio-coding invention, not a "financial product or service."
• IPR under § 311 is the only AIA trial vehicle. Inter partes reexamination was repealed for new filings effective 2012-09-16 and does not help here; ex parte reexamination remains available to anyone at any time and is the only non-IPR administrative route. Practically, its estoppel-free, ex parte nature makes it a poor substitute for IPR, but it can be a useful fallback if a § 315(b) bar has already run.
• The § 315(b) one-year clock is the critical deadline. It runs from service of a district-court complaint alleging infringement. If you have already been served and more than a year has passed, IPR is off the table for you — and for your privies — and you are limited to ex parte reexam plus district-court invalidity.

Pattern signals. None of the usual patterns are present. There is no repeat petitioner (there are no petitioners at all). There is no defensive aggregator (no Unified Patents, RPX, or similar entity) in the record for this patent. The patent owner, Orange SA, has not pursued any PTAB appeal on this patent because there has been no PTAB proceeding to appeal. The only forward activity the public record reflects is a 2024 citation by Tencent (US12002484B2) and a cluster of 2013–2014 France Télécom family filings (FR3003682A1, FR3003683A1), none of which indicate validity attacks.

The absence is itself a signal — but a weak one, and you should read it carefully. Well-asserted patents eventually attract IPRs. Here, the highest-probability explanation is that 8,812,327 has not been asserted in a campaign-scale litigation that would draw petitions. Orange's G.729.1 scalable-extension work is standards-oriented (ITU-T), which historically has produced patent-pool / licensing behavior rather than serial district-court assertion — and pool members' targets often conclude that filing an IPR burns a licensee relationship for little gain. That is an inference from the record's silence and from the standards context, not a documented fact, and I flag it as such. The alternative explanation — that licenses and settlements are quietly resolving disputes before petitions are filed — is untestable from public data. Before relying on the "nobody has bothered to attack it" heuristic, confirm whether the patent (or its EP/KR/CN/CA siblings) has been asserted anywhere, since a foreign counterpart suit would not produce a US PTAB docket but would change the calculus.

Recommended next steps

1. Do not assert or assume PTAB activity. There is none. Confirm this is still true at the time you rely on it by searching PTAB E2E (ptacts.uspto.gov) by patent number 8,812,327, and by checking the patent number on CourtListener's docket search for any CAFC appeal from a PTAB decision. The ODP feed is authoritative as of the most recent ingest, but a very recently filed petition may not yet appear. Re-verify within 30 days of any filing deadline you are managing.

2. Preserve your § 315(b) window. If you have been served with a complaint asserting 8,812,327, calendar the one-year IPR deadline immediately. Because there is no existing petitioner, you have a clean slate: no § 315(e) estoppel risk from a prior filer's grounds, and no risk that the Board will exercise Fintiv/General Plastic discretion against you as a serial filer. That makes IPR an unusually attractive option on this patent relative to the typical crowded-assertion scenario.

3. Build the validity case around the normalization limitation. The "last sub-band of the low frequency band and/or first sub-band of the high frequency band" limitation is recited in all six independent claims. Strong IPR grounds will center on (a) G.729.1-family and TDBWE/TDAC references combined with standard perceptual-audio-coding masking texts — the specification itself cites M. Bosi & R. Goldberg, Introduction to Digital Audio Coding and Standards (Kluwer, 2003), and (b) the Panasonic and Motorola references already cited on the face of the patent (US 2010/0017204; US 2010/0017200; US 2011/0046946; US 8,200,496; US 8,209,190; US 8,219,408), all of which are § 102(b)/§ 103 candidates given the 2009 priority date.

4. Exploit the family structure for prior art. Orange's own earlier application US 2010/0121646 A1 (France Télécom, priority 2007-02-02) is cited on the face of the patent and sits in the same subject-matter space. Family members and sibling filings are often overlooked § 102(e)/§ 102(a) art and should be charted.

5. Calendar the patent's economics. Adjusted expiration is 2030-11-15. This is not a near-expiry patent, so there is no "wait it out" defense and no cost-of-capital discount worth much. Conversely, confirm maintenance-fee payment status via USPTO Patent Center — a lapse would end the exposure outright, and the ODP legal-status field ("Active") is expressly a non-legal conclusion.

6. If you are a licensor or patent owner. The corollary is that 8,812,327 has never been tested. Every one of its 12 claims is intact but unconstrued, and the first petitioner to file will get a meaningful say in how the normalization limitation is read. The narrowness of that limitation across all independent claims is your principal exposure if the patent is ever asserted aggressively.

Nothing above should be read as a legal opinion on validity, infringement, or the accuracy of USPTO legal-status data. The proceeding count is drawn from the structured ODP block supplied in this prompt and was not contradicted by web search; where I could not verify a fact (e.g., whether the patent has been asserted in litigation, or maintenance-fee status), I have said so rather than assumed it.

I'll research the assignment record before writing. Let me pull live sources.

Let me dig for the specific application's assignment records.

Assignment-record reconstruction — US 8,812,327 B2

Method note / scope limitation. I worked from (a) the authoritative full text and legal-events block of US 8,812,327 B2 served from Google Patents, and (b) live web searches for corroborating USPTO records. The USPTO Assignment Center records themselves (reel/frame images) were not directly retrievable in this session — I hit a tool step limit on the targeted queries for application 13/382,786. Anything below marked unverified was not confirmed against the reel/frame image and should be checked at assignmentcenter.uspto.gov before being relied on. I have not invented any reel, frame, or correspondent.

Inventors

All four are listed as assignors in the recorded inventor→France Télécom assignment event ("Assignors: BERTHET, PIERRE, VIRETTE, DAVID, KOVESI, BALAZS, RAGOT, STEPHANE") shown in the patent's own legal-events record, dated 2012-03-23.

Unusual-pattern check: No departure pattern. This is a corporate R&D team ("Orange Labs"), not a startup founding team. Ragot in particular continues to appear as an Orange inventor on later Orange-owned codec patents (US 9,984,696 B2 "Transition from a transform coding/decoding to a predictive coding/decoding," applicant Orange, Paris; and French filings listing "RAGOT Stéphane – Demandeur: ORANGE Société anonyme"). That is affirmative evidence the team stayed with the assignee, i.e. no fire-sale precursor signal.

Original assignee

France Telecom SA (France Télécom), 6 place d'Alleray, Paris 75015, France — named on the face of the patent and in the inventor-assignment record. Google Patents additionally lists Orange SA as current assignee, reflecting the France Télécom → Orange corporate rebrand.

• Business: incumbent telecommunications operator and network/R&D house (Orange Labs). The patent is core codec R&D — an enhancement-layer bit-allocation/MDCT-masking method layered onto ITU-T G.729.1 (claims 1 and 6 recite G.729.1-type TDAC improvement coding).
• Product embodying the claims: the claimed subject matter is implemented as an improvement layer to the G.729.1 scalable codec, i.e. an ITU-T standardized speech/audio codec algorithm, rather than a standalone retail product. Orange's codec work feeds voice-over-IP and conversational-network equipment; the specification itself frames the invention as boosting G.729.1 widened-band quality ahead of its super-wideband extension — a carrier-driven standardization artifact. There is no evidence of a discrete consumer SKU, but this is operating-company R&D in the ordinary course, and Orange's portfolio is implemented across its own network and licensed into the codec ecosystem.
• Current status: operating, not dissolved, not in bankruptcy. France Télécom was renamed/rebranded Orange S.A. (change of name; entity continued, SIREN 380129866 per INPI records).

Assignment timeline

IMPORTANT: The fetched authoritative record exposes exactly one post-filing ownership event on the patent's own docket. There is no chain of LLCs, no security interests, and no third-party assignee anywhere in the record.

1. Executed 2010-xx to 2012-03 (execution date not stated in the fetched record) / recorded 2012-03-23 — Reel/frame not exposed in the fetched Google Patents event record; verify in Assignment Center

   • Conveyance: Assignment of assignors' interest ("ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
   • Assignor: Pierre Berthet; David Virette; Balazs Kovesi; Stéphane Ragot (the four named inventors)
   • Assignee: France Telecom SA, 6 place d'Alleray, Paris 75015, France
   • Correspondent: Not stated in the fetched record. In national-stage cases of this vintage the recording is normally filed by the PCT national-stage firm of record (typically the US associate handling the 371 entry); I cannot name it without the reel image.
   • Context: Routine inventor-to-employer assignment taken to perfect record title on national-stage entry (PCT/FR2010/051307, filed 2010-06-25; US 13/382,786). Not a transfer, not a monetization step.

2. (Likely) Executed 2013-07-02 / recorded 2017-12-08 — Reel 044625/0361 — France Télécom → Orange — UNVERIFIED FOR THIS PATENT

   • Conveyance: Change of name (see document for details)
   • Assignor: France Télécom — Assignee: Orange, 78 rue Olivier de Serres, Paris 75015
   • Correspondent: as surfaced, Daniel W. Juffernbruch, Patents and Licensing LLC, 28 Barrington Bourne, Barrington, IL 60010-9605
   • Source: this entry appears in a USPTO PTAB petition exhibit (petition 1555329) that attaches a bulk assignment run covering 52 properties in a different France Télécom family (assignors Joumay / Niger / Cauvin — "Management of Network Layer Packets in an Access Network of a Telecommunications Network"). It is a name-change record applied across an Orange portfolio, so it plausibly includes US 8,812,327, but I could not confirm this patent number appears on that reel. Treat as probable but unverified.
   • Context: Internal corporate rename (France Télécom rebranded to Orange). No change in beneficial ownership, no consideration, no monetization.

Adjacent context not in this chain: on the same exhibit, an assignment at Reel 044625/0315 moved properties of that other France Télécom family (52 properties, executed 2017-09-21, recorded 2017-12-08) to Transpacific IP Group Limited (PO Box 10008, Willow House, Cricket Square, Grand Cayman). This is a different patent family and must not be attributed to US 8,812,327. It is nonetheless a useful reminder to verify whether Orange's 2017 Transpacific divestment program touched this patent — I found no evidence that it did.

No other recorded assignments for US 8,812,327 were found. No security agreements, no licenses of record, no releases, no merger records.

Timeline diagram

timeline
    title Ownership of US 8812327
    2009 : FR priority filed by France Telecom
    2010 : PCT filed
         : US national stage entered
    2012 : Inventors assign to France Telecom
    2013 : France Telecom renamed Orange
    2014 : US patent issued
    2030 : Adjusted expiry active

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. The only confirmed assignee-in-interest is France Telecom/Orange, an operating carrier with a public SIREN registration and a physical Paris headquarters. No "IP/Holdings/Ventures" successor appears on the patent's docket. The Grand Cayman Transpacific IP Group entity that appears in the search results belongs to a different family (reel 044625/0315) and cannot be cited here.

2. Known asserter in the chain — NOT PRESENT. Searched against the named set (Acacia, Marathon, IV, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation, Spangenberg entities): none appears as assignor or assignee. France Telecom/Orange is a carrier, not on any NPE directory I can confirm.

3. Repeat correspondent across the chain — UNCLEAR / INSUFFICIENT DATA. The confirmed 2012-03-23 inventor-assignment entry in the fetched record does not state a correspondent, and the reel image was not retrievable this session, so I cannot test for recurrence. The one correspondent name that surfaced (Daniel W. Juffernbruch, Patents and Licensing LLC, Barrington IL) appears on the unverified France Télécom→Orange name change and is also the correspondent on the separate Transpacific Cayman assignment — but note that a name-change recording and a bulk divestment are exactly the kind of routine and bulk filings many IP-service firms handle; a single firm handling several France Télécom/Orange records is not by itself an NPE indicator. No recurrence within this patent's chain is established.

4. Cascading transfers — NOT PRESENT. Zero chained-LLC hops. One inventor assignment in 2012; nothing after.

5. Pre-litigation transfer — NOT PRESENT. No infringement suit naming US 8,812,327 was found, and there is no transfer within six months of any suit. There is no litigation-associated assignment at all.

6. Bankruptcy fire-sale — NOT PRESENT. France Télécom/Orange has not been through Chapter 7/11; the 2013 event is a change of name, not an asset sale in insolvency.

7. Privateering — NOT PRESENT. No operating-company→NPE transfer; no SEC 10-K/8-K disclosure found indicating Orange placed this patent with an assertion vehicle. Orange's own portfolio is asserted/licensed by Orange, if at all, in its own name.

8. Defensive aggregator — NOT PRESENT. Chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. This patent is not neutralized.

Note on a positive (inverse) signal: the fetched record lists the patent as Active with an adjusted expiration of 2030-11-15, i.e. maintenance fees are being paid and the asset is being kept alive by the original corporate owner — the ordinary profile of a held operating-company patent, not an abandoned shell asset.

Verdict

Insufficient data — only the original assignment is documented (inventor→France Telecom recorded 2012-03-23; no second link confirmed against a reel image in this session).

Justification: the entire documented chain is a single inventor-to-employer assignment to France Telecom SA, with the only subsequent event a change of name to Orange S.A. (plausibly reel 044625/0361, executed 2013-07-02 — unverified for this patent number); there are zero shell-entity transfers, no asserter in the chain, no cascading LLC hops, no bankruptcy sale, and no suit-triggered transfer. On this record the patent is a held operating-company asset, and every NPE signal above scores not-present or unclear — but because I could not pull the Assignment Center reel images directly, I am reporting this as insufficient data with an explicit no-NPE lean rather than certifying a clean chain. (Had the chain terminated at a defensive aggregator it would be "defensive/non-asserting"; had a suit been found with a clean operator plaintiff it would be "operating-company assertion" — neither is established.)

Verify at: USPTO Assignment Center (search patent 8812327 / application 13/382,786) and the legacy index at assignment.uspto.gov. Confirm whether reel 044625/0361 (France Télécom → Orange, change of name) lists US 8,812,327 among its 52 properties, and capture the correspondent of record on the 2012-03-23 entry.

I'll research this patent and its prior-art citations. Let me run several searches.

Let me research the key prior-art references in depth.

Prior-Art Analysis — US 8,812,327 B2

1. Confirmation of the patent (USPTO / database match)

Subject matter: a scalable/hierarchical audio codec in which the core layer uses an energy-criterion binary bit allocation, and the improvement layer instead allocates bits using a frequency masking threshold (normalised at the low/high band junction) combined into a "perceptual importance" that also subtracts the bits already spent by the core (and by any earlier enhancement stage), the residual then being coded with that allocation. Claim 6 ties it to TDAC enhancement of a G.729.1-type core.

Important framing note. Section 102 requires that one single reference disclose every element of a claim, arranged as claimed. None of the 18 references cited on the front page discloses the whole combination of claim 1 (masking threshold + normalisation at the band junction + perceptual importance computed as a function of the bits allocated for the core coding + a second binary allocation for the residual). Consequently the citations below are, with one partial exception, §103 obviousness candidates rather than true §102 anticipators. I flag the limited §102 exposure explicitly.

2. The 18 U.S. patent citations (as listed on the face of US 8,812,327)

Tier A — closest art

A1. US 2010/0121646 A1 — Ragot & Guillaume (France Telecom); pub. 2010‑05‑13; priority 2007‑02‑02; granted as US 8,543,389 B2 — "Coding/decoding of digital audio signals."
Discloses transform coding of adjacent first (low) and second (high) sub‑bands, determining a frequency masking threshold applied to the second/high sub‑band, and normalising that masking threshold to ensure spectral continuity between the two sub‑bands — explicitly normalising by the value at the last sub‑band of the low band, or the first sub‑band of the high band — with advantageous application to the high‑band TDAC of a G.729.1 hierarchical coder.

• Potential §102 exposure: the "calculation of a frequency masking threshold…" plus "…normalized by the value of the masking threshold at a last sub‑band of the low frequency band and/or a first sub‑band of the high frequency band" language of claims 1, 7, 9, 10, 11, 12 and the convolution/spreading‑function feature of claim 4.
• Why it does not anticipate claims 1/7 as a whole: it does not disclose (i) computing a perceptual importance that is a function of the number of bits allocated for the core coding, nor (ii) a second binary allocation of the residual/improvement layer. It is directed to perceptual weighting of the spectrum, not to masking‑driven bit allocation of an enhancement residual. This is the single most dangerous reference, but as a §103 primary reference combined with the G.729.1 allocation teaching, not as a clean §102 bar.

A2. US 2009/0326931 A1 — Ragot & Virette (France Telecom); pub. 2009‑12‑31; priority 2005‑07‑13; "Hierarchical encoding/decoding device."
Core layer (parametric, analysis‑by‑synthesis) + band‑extension layer widening to wideband + a transform‑coding quality‑enhancement layer that uses a spectral parameter from the band‑extension layer. Directly maps onto the architecture of claims 1/7 (core + higher‑bitrate improvement over a residual) and the G.729.1 context of claim 6.

• Potential §102/§103: relevant to the preambles of claims 1, 6, 7, 9, 10; does not disclose the masking‑threshold/perceptual‑importance allocation → no full anticipation.

A3. US 2006/0036435 A1 — Kovesi et al. (France Telecom); pub. 2006‑02‑16; priority 2003‑01‑08; "Method for encoding and decoding audio at a variable rate."
Variable‑rate (scalable) audio codec from the same applicant family; layered bit allocation across stages.

• Potential §103: bit‑allocation across hierarchical layers (claims 1, 3); no masking‑threshold normalisation, no PI‑vs‑core‑bits coupling → no §102.

A4. US 8,209,190 B2 — Ashley, Gibbs & Mittal (Motorola Mobility); pub. 2012‑06‑26; filed 2008‑08‑07; "Method and apparatus for generating an enhancement layer within an audio coding system."
Codes the error signal (input vs. locally‑decoded core) in the MDCT domain through a perceptual weighting matrix W derived from LP coefficients, with gain selection minimising a distortion metric.

• Potential §103: "improvement coding of a higher bitrate of a residual signal" and perceptual weighting of the enhancement (claims 1, 7, 9). Does not disclose a masking threshold normalised at the band junction nor a PI computed from core‑coded bit counts → no §102.

Tier B — scalable/layered coding context (preamble‑level relevance only)

Tier C — band‑extension / enhancement‑layer art (limited relevance)

3. Non‑patent literature cited

• Jin A. et al., "Scalable Audio Coder Based on Quantizer Units of MDCT Coefficients," ICASSP 1999, vol. 2, pp. 897–900 (XP01032846). MDCT‑domain scalable coding with bit allocation — §103 context for the transform/enhancement architecture (claims 1, 6).
• M. Bosi & R. Goldberg, Introduction to Digital Audio Coding and Standards, Kluwer, 2003 — cited by the applicant itself as the general teaching on frequency masking models and their application in transform coders (supports the claim‑1 masking‑threshold element as known in the art → relevant to §103).
• International Search Report & Written Opinion, PCT/FR2010/051307, 2010‑10‑06 — the search report for this family; useful to see the examiner's own relevance categories.

4. Bottom line

1. The most relevant prior art is US 2010/0121646 A1 (Ragot & Guillaume, France Telecom → US 8,543,389 B2). It is the only cited reference that expressly discloses a frequency masking threshold normalised by the value at the last low‑band sub‑band / first high‑band sub‑band — i.e., a key limitation of claim 1 (and mirror claims 7, 9, 10, 11, 12) and the convolution/spreading feature of claim 4.
2. No single cited reference anticipates claim 1, 7, 9 or 10 as a whole. The differentiating elements — computing a perceptual importance that is a function of the number of bits allocated for the core coding (and, per claim 3, for earlier energy‑based enhancement stages) and performing a second binary allocation over the residual accordingly — are not shown by any one citation. Expect §103 grounds, with Ragot '646 as primary and a G.729.1/energy‑allocation reference (Bosi & Goldberg, the G.729.1 standard itself, or Kovesi '435/Vasilache '986) as secondary, rather than a clean §102 rejection.
3. Claim 6 is the most vulnerable to a §102/§103 attack using the G.729.1 standard (ITU‑T G.729.1, May 2006) itself as the base, since it recites an improvement coding of TDAC type over a G.729.1‑type core.
4. The remaining references (Tier B/C) are preamble‑level or background art for scalable/layered coding and band extension and individually cannot anticipate any of claims 1–12.

Caveats on confidence: the citation metadata and titles above are taken from the patent's own front‑page citation table (Google Patents / FreePatentsOnline / Justia renderings of the official record). I was able to verify the substance of US 8,543,389/2010‑0121646 (Ragot), US 8,209,190 (Ashley) and US 2010/0292986 (Vasilache) in detail; for the remaining references I relied on their titles, dates and assignees as recorded, so the per‑reference descriptions of B/C‑tier documents should be treated as provisional and confirmed against full text before filing any §102 position. I did not obtain a live USPTO PatentCenter/PAIR file‑wrapper (prosecution history) copy, so no statement is made about examiner‑applied art versus applicant‑cited art beyond the face of the patent.

I'll ground this § 103 analysis by pulling the cited prior-art references and checking what they teach.

This is a rich result set. Let me dig into the most probative references to nail down exactly what they disclose.

Obviousness Analysis — U.S. Patent No. 8,812,327 B2 ("Coding/decoding of digital audio signals")

0. Framework and caveats

Legal standard. Because the application was filed June 25, 2010 (priority 2009‑07‑07), pre‑AIA 35 U.S.C. §§ 102/103 governs. Obviousness is assessed under Graham v. John Deere, 383 U.S. 1 (1966), and KSR Int'l v. Teleflex, 550 U.S. 398 (2007): the scope/content of the prior art, the differences between the art and the claims, PHOSITA's level of skill, and secondary considerations. Under KSR, a combination of known elements is obvious where it yields no more than predictable results, is a mere "design choice," or responds to a "known problem for which there was an obvious solution encompassed by the patent's claims."

PHOSITA. A speech/audio codec engineer with a bachelor's degree in EE/CS and 2–4 years' experience with standards‑based codecs (G.729, G.729.1, MPEG‑4), transform/MDCT coding, and psychoacoustic (masking) models.

Caveats (stated up front, per the operating rules):

• I have full claim text and substantial specification text for U.S. 8,543,389 B2 (the U.S. counterpart of US 2010/0121646 A1), which is on the face of US 8,812,327. For the other cited references I have abstracts, claims, and specification excerpts from the search results, not complete texts. Statements below are hedged where I could not verify a disclosure.
• Date check. US 8,543,389 B2 published as US 2010/0121646 A1 on 2010‑05‑13, i.e. after the '327 patent's 2009‑07‑07 priority date. However, its underlying disclosure has an FR priority of 2007‑02‑02 and published earlier as FR 2 912 249 A1 and WO 2008/104663 (2008). Those earlier publications are pre‑2009‑07‑07 and therefore are available as prior art under pre‑AIA §102(a)/(b); the U.S. patent is usable as evidence of what that art teaches. I flag this because it matters for how the reference is formally cited (I use the family's earlier publications as the §102 art and US 8,543,389 as corrborating disclosure of the same content).
• I assume the G.729.1 Recommendation (ITU‑T, approved May 2006) and Bosi & Goldberg, Introduction to Digital Audio Coding and Standards (Kluwer, 2003) — the latter cited by the '327 patent itself — are prior art printed publications. Both are.

The '327 patent's three independent‑claim families (claims 1, 7, 9/10, 11/12) are the same subject matter in method, decoder, apparatus, and CRM form, so the analysis for claim 1 carries to the rest.

1. Claim 1 decomposed, and what the cited art shows

Claim 1 (as granted) requires:

Reference A — US 8,543,389 B2 / WO 2008/104663 / FR 2 912 249 A1 (Ragot & Guillaume, France Telecom) — cited on the face of the '327 patent as US 2010/0121646 A1

This is extraordinarily close and, in my view, the single most dangerous reference. Its disclosed/claimed subject matter:

• C (masking + normalization): squarely disclosed. Claim 1: "determining at least one frequency masking threshold to be applied on the second sub‑band, the same threshold not being applied on the first sub‑band, and normalizing said masking threshold in order to ensure a spectral continuity between said first and second sub‑bands." The specification explicitly gives normfac = log₂[Σ_{j=9..17} σ²(j)·B(v₉−v_j)] (normalize by the last sub‑band of the low band) or normfac = log₂[Σ_{j=10..17} σ²(j)·B(v₁₀−v_j)] (normalize by the first sub‑band of the high band). That is verbatim the "and/or" limitation of '327 claim 1. (Source: https://patentimages.storage.googleapis.com/51/67/7f/ae4c43a7c584f2/US8543389.pdf)
• D (in part): disclosed. Claim 3 gives the very formula the '327 specification reproduces: ip(j) = ½·rms_index(j) for j in the first band, and ip(j) = ½[rms_index(j) − log_mask(j)], with log_mask(j) = log₂(M(j)) − normfac, where rms_index(j) are "quantized values originating from the coding of the envelope."
• E: disclosed. Claim 2: "the bit allocation for the second sub‑band at least is determined moreover as a function of a normalized masking curve computation."
• A/B context: disclosed. Claims 6, 8, 9: the transform coding is in an upper layer of a hierarchical coder; the first sub‑band comprises the core‑coded signal (indeed "a signal representing a difference between an original signal and a synthesis of this original signal"); and "the transform coding is of the TDAC type in an overall coder according to standard G.729.1… the second sub‑band is included in a high‑frequency band." The upper layer of G.729.1 is the energy‑criterion core.
• Claim‑4‑type element (spreading convolution): disclosed. Claim 11: masking threshold "defined by a convolution between… an expression of the spectral envelope, and a spread function involving a central frequency of said sub‑band."
• Claim‑5‑type element (tonality): disclosed. The spec discusses using the G.729.1 envelope‑coding‑mode bit (differential Huffman vs. direct natural binary) to decide whether to apply the masking, which is exactly claim 5 of '327.

Residual gap versus claim 1: US 8,543,389 computes ip(j) from the quantized envelope and the masking threshold; it does not expressly recite that the perceptual importance is a function of "the number of bits allocated for the core coding," nor does it frame the upper layer as coding a separate residual with its own second allocation distinct from the core allocation. Those are the two points of novelty.

Reference B — US 2006/0036435 A1 / WO 2004/070706 (Kovesi & Massaloux, France Telecom) — cited on the face of the '327 patent

This reference closes the gap on elements B and D:

• B (residual/enhancement coding): disclosed. Spec: the "second coding stage 2 … operates on the coding residual E of the first stage, supplied by a subtractor 3." So the enhancement layer codes the residual of the core.
• D (allocation as a function of core‑coded parameters + psychoacoustic criterion): disclosed. Family member CN 1735928 B claim 5: the ranking of the enhancement‑subset bits is determined, "according to the encoded parameters of the first subset, by means of at least one psychoacoustic criterion." Claim 6 adds: "a frequency masking curve is computed by applying a perceptual model to the estimated spectral envelope, and … the psychoacoustic criterion references the level of the estimated spectral envelope relative to the masking curve in each spectral band" — and the envelope is estimated from the core subset's coded parameters. Claim 14/spec confirms the "difference signal" and its energy levels are carried in the first subset.
  (CN 1735928 B: https://patentimages.storage.googleapis.com/a4/d0/7d/9f3e731283c78f/CN1735928B.pdf; US pub: https://patents.google.com/patent/US20060036435)

So Reference B teaches the precise conceptual linkage that the '327 patent asserts as its point of novelty: rank/allocate enhancement‑layer bits as a function of what the core layer already coded, using a masking‑curve‑based perceptual criterion.

Reference C — US 2003/0220783 A1 / US 7,277,849 B2 (Streich, Nokia) — cited on the face of the '327 patent

Perceptual scalable coding with a residual: claim 2 does the evaluation "individually for components … corresponding to said plurality of frequency bands"; claim 3 "compar[es] the residual signal to a masking threshold"; claim 15 accumulates "a core layer quality measure by summing together differences between masking thresholds and residual signal components … and us[es] the core layer quality measure when evaluating the residual signal." This is masking‑based evaluation of the residual, explicitly accounting for the quality the core layer already delivered. (https://patents.google.com/patent/US20030220783)

Reference D — US 8,209,190 B2 / US 8,219,408 B2 / US 8,200,496 B2 (Ashley et al., Motorola) — cited on the face of the '327 patent

These teach generating an enhancement layer that codes the error signal E = MDCT{W(s − s_c)} with W a perceptual weighting matrix, and using a "psychoacoustic model for improved coding of the enhancement layer error signal." They confirm that (i) residual‑coding enhancement layers and (ii) psychoacoustic modeling inside the enhancement layer were known. (https://patents.google.com/patent/US8209190)

Reference E — the base art the patent itself concedes

• G.729.1 (and the '327 patent's own background): core CELP + TDBWE + TDAC; the energy‑criterion "perceptual importance" ip(j)=½rms_index(j); bit allocation by dichotomy; the envelope‑coding‑mode bit (differential Huffman / direct natural binary). These supply element A and half of D's "energy criterion" antecedent.
• Bosi & Goldberg (2003), cited in the '327 patent's own "Drawbacks of the Prior Art," supplies the general teaching that transform audio coders use frequency/simultaneous masking models and that energy‑only allocation "is not optimal from a perceptual point of view, especially for coding music signals" — the stated problem.
• US 6,526,384 B1 (Siemens) and US 5,864,801 (Mitsubishi) — cited — provide background on perceptual/scaleable bit‑rate limitation and hierarchical audio coding respectively (cumulative).
• US 2003/0206558 A1 (Parkkinen), US 2003/0220783 and US 2008/0021712 (Fejzo) — cited — establish that perceptual masking + scalable layering were combined in the art.

2. The obviousness combinations

Ground 1 (strongest): G.729.1 + US 8,543,389 (Ragot '389) + US 2006/0036435 (Kovesi '/435), optionally + Bosi & Goldberg

Claim 1 mapping

Why a POSITA would combine them (motivation, KSR‑compliant):

1. Same field of endeavor, same problem, same assignee. All three references are in hierarchical/scaleable transform audio coding; the masking references are France Telecom documents in the same G.729.1 TDAC setting, addressing exactly the '327 patent's stated problem — that G.729.1's "energy criterion … is not optimal from a perceptual point of view, especially for coding music signals," and its super‑wideband extension "emphasizes the existing defects." Ragot '389 even flags the same motivation: extending the TDAC to 50 Hz–14 kHz.
2. The '327 patent's own background supplies the "known problem" and the "obvious solution." It concedes that "most audio coders rely on a transform coding using frequency masking models" and cites Bosi & Goldberg for masking models. Under KSR, "a known problem for which there was an obvious solution encompassed by the claims" supports obviousness.
3. Ragot '389 itself points to the combination: it expressly identifies the application as "perceptual weighting of the high‑frequency band in the TDAC transform coding of a hierarchical encoder according to standard G.729.1," and notes the low band is already perceptually weighted while the high band is not — the exact asymmetry the '327 patent purports to remedy.
4. The only remaining difference — making the enhancement‑layer importance a function of the core bit allocation — is not only suggested but taught by '/435 (enhancement bit ranking "as a function of" the core‑coded parameters, using a masking curve), and is, in any event, a predictable accounting step: an enhancement layer that codes the residual of the core should base its perceptual importance on what the core already captured, otherwise it double‑counts bits. KSR design‑choice/predictable‑result reasoning applies.

Ground 2: US 2003/0220783 (Streich) + G.729.1 + Ragot '389

Streich supplies the residual‑enhancement architecture with masking‑based evaluation and a "core layer quality measure" used when evaluating the residual (claims 2, 3, 15) — i.e., element D's core‑dependence. Ragot '389 supplies the G.729.1‑specific masking+normalization. Motivation: both are scaleable perceptual audio coders; combining a masking‑guided residual coder with a normalization scheme to preserve spectral continuity at the band junction addresses a recognized artifact‑avoidance need.

Ground 3: Ragot '389 + Motorola '190/'408/'496 + G.729.1

The Motorola references expressly build a psychoacoustic model inside the enhancement layer and code the error signal E = MDCT{W(s − s_c)}. Combining with Ragot '389's masking/normalization formula yields all elements of claim 1. Motivation: both target the music/non‑speech quality deficiency of CELP‑core codecs (Motorola: "there remains a need for improving performance of CELP type speech coders at low bit rates, especially for music").

3. Dependent claims

4. Secondary considerations

• No assertion/nexus evidence located. My earlier litigation search found no litigation on US 8,812,327 (owner Orange SA, an operating telco). There is therefore no commercial‑success or licensing evidence in the record tying any secondary consideration to the claims.
• The only forward citation is US 2020/… Tencent (US 12,002,484 B2) — not evidence of non‑obviousness.
• The patentee's likely rebuttal: the "subtract core bits" element (claim 2 and the corresponding clause in claim 1) is argued to be the inventive contribution. This is the only place the record supports a genuine non‑obviousness argument, and it is a weak one (routine accounting), given that Ragot '389 and Kovesi '/435 already place masking‑based, core‑parameter‑dependent allocation in the same G.729.1 enhancement setting.

5. Bottom line

1. Claim 1 is, in my assessment, obvious over G.729.1 (energy‑criterion core) in view of US 8,543,389 / WO 2008/104663 (masking threshold + normalization by the last low‑band sub‑band and/or first high‑band sub‑band; masking‑based perceptual importance ip(j)=½[rms_index−log_mask]) and US 2006/0036435 / WO 2004/070706 (residual enhancement coding with bit ranking that is a function of core‑coded parameters and a masking curve). Elements C, E, and the claim‑4/‑5/‑6‑type features are essentially disclosed by US 8,543,389; elements B and D are supplied by US 2006/0036435.
2. Claims 4, 5, 6 stand very weak — Ragot '389 (claim 11 for the convolution; claims 9–10 for the TDAC/G.729.1 application) discloses them almost verbatim; claim 5's tonal gate is disclosed by the G.729.1 envelope‑mode bit.
3. Claim 2 (and the "number of bits allocated for the core coding" language in claims 1/7) is the narrowest, most defensible claim. Absent a reference expressly reciting the subtraction of core‑allocated bits per coefficient, the best showing is "obvious modification / design choice," which a reasonable examiner or jury could accept under KSR but which is not airtight on the current record.
4. Recommended next step to make the §103 case airtight: verify whether US 7,835,904 B2 ("Perceptual, scalable audio compression") — which explicitly describes using "a psychoacoustic mask to guide residue coding in enhancement layer coders," with the mask computed at the base layer — is prior art to 2009‑07‑07. It is the strongest candidate for teaching masking‑guided residual allocation and, if earlier, would close the last gap in claim 1/2. (It was not in the "Prior Art" citation list on the '327 page, so I flag it as supplementary rather than part of the sanctioned set.)

Explicit uncertainty flags: (a) For every reference except US 8,543,389, I worked from abstracts/claims/excerpts returned by search, not complete specifications; specific disclosures (e.g., whether Kovesi '/435 ties the masking‑based ranking quantitatively to bit counts) should be confirmed against the full text. (b) The prior‑art status of US 8,543,389 in its U.S. patent form is date‑sensitive (US publication 2010‑05‑13 post‑dates the '327 priority); rely on the earlier WO 2008/104663 / FR 2 912 249 publications for the §102 predicate.