Datascription LLC v. Meta Platforms Inc

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Datascription, LLC (Delaware)
  • Defendant: Meta Platforms, Inc. (Delaware)
  • Plaintiff's Counsel: Folio Law Group PLLC
• Case Identification: 7:26-cv-00183, W.D. Tex., 07/23/2026
• Venue Allegations: Plaintiff alleges venue is proper in the Western District of Texas because Defendant Meta Platforms, Inc. conducts substantial business in the district, operates a datacenter campus in Temple, Texas, employs individuals in the district, and has previously consented to personal jurisdiction in the district in other litigation.
• Core Dispute: Plaintiff alleges that Defendant's automated systems for transcribing and normalizing audio in media on its platforms, including Facebook and Instagram, infringe a patent related to the automated extraction of metadata from non-transcribed audio and video streams.
• Technical Context: The lawsuit concerns technology for the large-scale, automated analysis of audio-visual media, a foundational capability for modern social media and content platforms that handle vast quantities of user-generated video and live streams.
• Key Procedural History: The complaint notes that Defendant has previously agreed that it is subject to personal jurisdiction in the Western District of Texas in separate, unrelated cases.

Case Timeline

Date	Event
2013-07-10	'547 Patent Priority Date
2016-01-05	'547 Patent Issue Date
2026-07-23	Complaint Filing Date

II. Technology and Patent(s)-in-Suit Analysis

• Patent Identification: U.S. Patent No. 9,230,547, "Metadata extraction of non-transcribed video and audio streams", issued January 5, 2016 (the "'547 Patent").

The Invention Explained

• Problem Addressed: The patent's background describes the process of analyzing and extracting metadata from media without a pre-existing script as a historically labor-intensive, human-driven, and error-prone task. Existing automated methods were often slow and required a written script for comparison, which is unavailable for most user-generated or live content. '547 Patent, col. 1:18-34 '547 Patent, col. 1:53-61
• The Patented Solution: The '547 Patent discloses a server-based system that automates this process. The system extracts an audio stream, generates a time-aligned textual transcription using speech recognition, and, critically, normalizes the audio stream to a "single, normalized, universal amplitude scale." This normalization is achieved by generating an audio histogram, identifying the loudest and softest frames within the audio stream, assigning them normalized minimum and maximum values, and then mapping every other frame to a value within that normalized scale. '547 Patent, abstract '547 Patent, col. 2:20-38 '547 Patent, FIG. 5 This enables all processed media to be searched and compared on a consistent basis, not just by text but also by audio characteristics. '547 Patent, col. 11:18-24
• Technical Importance: The invention sought to provide an automated, scalable method for making large volumes of unstructured multimedia content searchable by textual, visual, and aural metadata, including loudness. Compl. ¶13

Key Claims at a Glance

• The complaint asserts infringement of at least independent claim 1. Compl. ¶19
• The essential elements of independent claim 1 include:
  • Extracting an audio stream from a non-transcribed source media.
  • Using speech recognition to create a time-aligned textual transcription.
  • Extracting time-aligned audio frames.
  • Processing the audio frames to extract and assign a numerical value to aural amplitudes.
  • Generating an audio histogram of the audio stream.
  • Normalizing the audio stream to a universal scale by determining a loudest and a softest frame.
  • Assigning normalized minimum and maximum amplitude values to the softest and loudest frames, respectively.
  • Comparing each frame to the loudest and softest frames using the audio histogram to assign a normalized amplitude value.
  • Processing the transcription to extract time-aligned textual metadata.
  • Storing the transcription, audio frames, aural amplitudes, textual metadata, and normalized amplitude values in a database.
• The complaint states that the provided infringement example is non-limiting and reserves the right to allege infringement of other claims. Compl. ¶21

III. The Accused Instrumentality

Product Identification

• The complaint identifies the "Accused Instrumentalities" as Meta's use of the "Wav2vec algorithm for data extraction and transcription from non-transcribed source media such as Facebook live, an uploaded video, Instagram Live, and Reels" within its datacenters. Compl. ¶6

Functionality and Market Context

• The complaint alleges that Meta's platforms, including Facebook and Instagram, employ automated systems to generate captions for video content. This service is allegedly powered by AI, specifically using Automated Speech Recognition (ASR) algorithms like Wav2vec to analyze speech and generate transcriptions from raw audio data. (Compl. ¶¶23; Compl. ¶25; Compl. ¶27). A screenshot from a Facebook help page describes how the system "automatically creates captions for you to review." Compl. p. 8
• The complaint further alleges that Meta's services perform loudness normalization to ensure a consistent volume level for users across different videos. This is allegedly accomplished using technologies such as the xHE-AAC codec and methods consistent with industry standards like ITU-R BS.1770, which involve measuring loudness levels (e.g., in LUFS) and adjusting them to a target value. (Compl. ¶31; Compl. ¶32).

IV. Analysis of Infringement Allegations

'547 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
[a] extracting an audio stream from the non-transcribed source media by a processor-based server;	Meta's datacenter servers extract raw audio data from uploaded media (e.g., Facebook Live, Reels) for processing.	¶24	col. 9:5-6
[b] speech recognition processing of the audio stream to transcribe the audio stream into a time-aligned textual transcription by a speech recognition engine to provide a time-aligned machine transcribed media;	Meta's systems use ASR algorithms, including Wav2vec, to analyze the extracted audio and generate time-aligned text transcriptions (i.e., auto-captions).	¶26	col. 9:7-12
[c] extracting time-aligned audio frames from the audio stream by an audio frame engine;	Meta's speech recognition allegedly utilizes corresponding time-aligned audio frames to associate recognized speech with temporal locations within the audio stream.	¶28	col. 9:16-20
[d] processing the time-aligned audio frames to extract audio amplitudes by a timed interval, to measure aural amplitudes...and assign a numerical value...to provide time-aligned aural amplitudes...;	Meta's systems allegedly process audio frames to measure loudness (e.g., LUFS) and assign these numerical values to provide time-aligned aural amplitudes as part of its loudness normalization process.	¶31	col. 9:30-34
[e] generating an audio histogram of the audio stream by the server processor;	Meta's loudness normalization process allegedly analyzes loudness information across the audio stream, which the complaint alleges involves generating and utilizing an audio histogram.	¶36	col. 9:46-49
[f] normalizing the audio stream to a single, normalized, universal amplitude scale by determining a loudest frame with a loudest sound and a softest frame with a softest sound within the audio stream...;	Meta's systems allegedly normalize audio to a target loudness level (e.g., -16 LUFS), which the complaint alleges includes determining the loudest and softest portions of the stream.	¶40	col. 9:50-54
[g] assigning a normalized minimum amplitude value to the softest frame of the audio stream and a normalized maximum amplitude value to the loudest frame of the audio stream;	The complaint alleges that Meta's normalization process assigns normalized loudness values, including minimum and maximum values corresponding to the softest and loudest frames.	¶43	col. 9:54-62
[h] comparing each frame of the audio stream to the loudest frame and the softest frame by utilizing the audio histogram and assigning a normalized amplitude value between the normalized minimum amplitude value and the normalized maximum amplitude value...;	Meta allegedly compares each frame relative to the loudest and softest frames by utilizing the generated audio histogram information to apply normalized loudness values throughout the stream.	¶45	col. 9:62-67
[i] processing the time-aligned machine transcribed media...to extract time-aligned textual metadata associated with the source media; and	Meta's systems allegedly extract and generate time-aligned textual metadata, such as timestamps and timing information, associated with the auto-generated captions.	¶47	col. 11:55-63
[j] storing the...media, the...audio frames, the...aural amplitudes, time-aligned textual metadata and the normalized amplitude value of each frame...in a database.	Meta's datacenters allegedly store the information generated during transcription and loudness normalization, including the transcription, timing information, and loudness data.	¶50	col. 23:7-12

• Identified Points of Contention:
  • Scope Questions: The complaint's theory appears to equate modern loudness normalization (targeting a specific integrated loudness value like -16 LUFS) with the patent's specific, multi-step normalization method. A central dispute may be whether targeting an average loudness level (as is common practice with LUFS) is the same as the claimed method of "determining a loudest frame...and a softest frame," assigning absolute min/max values to them, and "comparing each frame" to those two poles. The complaint provides an annotated diagram from an audio engineering publication to illustrate its theory of how these concepts map. Compl. p. 46
  • Technical Questions: A key technical question will be what constitutes an "audio histogram" as required by claim 1. The complaint alleges Meta's process utilizes "histogram-based processing" in connection with loudness measurement Compl. ¶37, but the defense may argue that its use of industry-standard algorithms (e.g., ITU-R BS.1770) does not generate or utilize an "audio histogram" in the specific manner required by the patent's normalization scheme.

V. Key Claim Terms for Construction

• The Term: "generating an audio histogram"

  • Context and Importance: This term is critical because the claimed normalization method requires "utilizing the audio histogram" to compare frames. The viability of the infringement case may depend on whether Meta's alleged loudness analysis process, which the complaint links to tools like FFmpeg and the xHE-AAC codec, can be shown to perform this specific step.
  • Intrinsic Evidence for a Broader Interpretation: The patent specification does not appear to provide a detailed definition, stating only that the server "generates audio histogram of the audio file/data" as part of the normalization process. '547 Patent, col. 9:46-49 A party could argue this encompasses any process that creates a data structure representing the distribution of audio amplitude levels for subsequent analysis.
  • Intrinsic Evidence for a Narrower Interpretation: The term is presented as a distinct step (417) in the process flow of FIG. 5, which is then used as an input for the comparison step (part of 600). A party could argue that, in the context of the overall claimed process, it implies a specific type of data structure necessary for the subsequent comparison to the "loudest frame and the softest frame," not just any loudness measurement data. '547 Patent, FIG. 5 '547 Patent, col. 9:62-65

• The Term: "normalizing the audio stream to a single, normalized, universal amplitude scale by determining a loudest frame...and a softest frame"

  • Context and Importance: This phrase describes the core of the patented normalization method. The central dispute will be whether Meta's practice of adjusting audio to a target LUFS (Loudness Units Full Scale) value is equivalent to the patent's method, which is defined by specific steps.
  • Intrinsic Evidence for a Broader Interpretation: A party might focus on the stated purpose of enabling a "universal search based on the sound level," arguing that any method achieving cross-file comparability by mapping sound levels to a consistent scale falls within the claim's scope. '547 Patent, col. 10:1-3
  • Intrinsic Evidence for a Narrower Interpretation: A party would argue that the claim language itself provides a narrow definition. The term is not just "normalizing," but normalizing by performing a specific sequence of actions: "determining a loudest frame" and "a softest frame," then assigning min/max values, then comparing other frames to them. '547 Patent, claim 1(f)-(h) They may argue that systems using integrated or average loudness measurements do not perform this specific pole-based comparison.

VI. Other Allegations

• Indirect Infringement: The complaint does not plead specific facts to support separate counts of induced or contributory infringement.
• Willful Infringement: The complaint does not contain allegations of pre- or post-suit knowledge of the patent sufficient to support a claim for willful infringement.

VII. Analyst's Conclusion: Key Questions for the Case

• A core issue will be one of technical equivalence: Does Meta's use of modern, standards-based loudness normalization (which typically relies on integrated or gated loudness measurements like LUFS) perform the same specific function as the patent's described method of identifying the absolute loudest and softest frames, creating an "audio histogram," and comparing each frame to those two poles?
• The case will also likely turn on a definitional question: Can the term "audio histogram," as used in the '547 Patent, be construed broadly to cover the loudness distribution data generated by algorithms like ITU-R BS.1770, or is it limited to a more specific data structure used in the precise manner detailed in the patent's process flow? The resolution of this construction may be dispositive for the infringement analysis.