DCT

1:25-cv-00967

Velos Media LLC v. ByteDance Ltd

Log In

I. Executive Summary and Procedural Information

  • Parties & Counsel:
  • Case Identification: 1:25-cv-00967, W.D. Tex., 06/24/2025
  • Venue Allegations: Venue is based on allegations that Defendant TikTok, Inc. maintains a regular and established place of business in Austin, Texas, and has committed acts of infringement in the district. The complaint emphasizes TikTok’s “Project Texas” initiative, involving data storage and operations with Austin-headquartered Oracle, as evidence of its ties to the state.
  • Core Dispute: Plaintiff alleges that Defendant’s TikTok video platform infringes six patents related to technologies for video encoding and decoding that are essential to the H.265/HEVC video compression standard.
  • Technical Context: The High-Efficiency Video Coding (HEVC/H.265) standard is a cornerstone of modern digital video, enabling the efficient compression and transmission of high-quality video content essential for streaming platforms like TikTok.
  • Key Procedural History: The complaint alleges that the asserted patents are essential to the H.265 standard and that their prior owners made commitments to license them on reasonable and non-discriminatory (RAND) terms. Plaintiff alleges it attempted to negotiate a license with Defendant, but the parties failed to reach an agreement, with Plaintiff characterizing Defendant’s negotiation conduct as being in bad faith.

Case Timeline

Date Event
2011-01-21 ’843 Patent Priority Date
2011-10-24 ’338 Patent Priority Date
2011-11-01 ’849 and ’962 Patents Priority Date
2012-01-20 ’184 and ’395 Patents Priority Date
2015-02-24 U.S. Patent No. 8,964,849 Issues
2015-04-14 U.S. Patent No. 9,008,184 Issues
2023-03-23 TikTok CEO testifies before Congress regarding "Project Texas"
2023-04-11 U.S. Patent No. 11,627,338 Issues
2024-09-10 U.S. Patent No. 12,088,843 Issues
2025-01-07 U.S. Patent No. 12,186,395 Issues
2025-06-24 U.S. Patent No. 12,341,962 Issues
2025-06-24 Complaint Filed

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

U.S. Patent No. 11,627,338 - "Significance map encoding and decoding using partition selection", issued April 11, 2023

The Invention Explained

  • Problem Addressed: In video compression, a "significance map" indicates the locations of important (non-zero) data points, or coefficients, within a block of video data. Encoding this map can consume a large portion of the total data (30-80%) and requires significant computational resources to manage the "contexts" used for efficient compression (Compl. ¶51; ’338 Patent, col. 2:21-40). Conventional methods assigned a unique context to each position in smaller blocks, leading to high computational demands (Compl. ¶52; ’338 Patent, col. 2:24-40).
  • The Patented Solution: The invention proposes a more efficient method by using a "non-spatially-uniform partitioning" of the significance map. Instead of assigning a unique context to every position, it strategically groups some positions to share a common context while leaving others with their own. This focuses the computational expense on the data positions that gain the most from having unique contexts, thereby improving overall compression efficiency (Compl. ¶53; ’338 Patent, col. 3:8-25). Figure 4 of the patent illustrates an example of this non-uniform partitioning for a 4x4 block, where different positions (e.g., P1, P2, P4) are assigned to a limited number of contexts (e.g., C0, C1, C3) (’338 Patent, Fig. 4).
  • Technical Importance: This approach improves video compression by making the signaling of significance maps more efficient, which in turn reduces data usage and computational load (Compl. ¶54).

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 (Compl. ¶83).
  • Claim 1 recites a method for decoding a bitstream to reconstruct a significance map for a 4x4 transform unit, comprising the steps of:
    • For each bit position in the significance map, determining a context based on a partition set.
    • The partition set assigns one of four unique contexts to each of the four bit positions in the upper-left quadrant.
    • It assigns a shared fifth context to two bit positions in one column of the upper-right quadrant, and a shared sixth context to two positions in the other column.
    • It assigns a shared seventh context to two positions in one row of the lower-left quadrant, and a shared eighth context to two positions in the other row.
    • It assigns a shared ninth context to three specific bit positions in the lower-right quadrant.
    • Decoding the encoded data based on the determined context to reconstruct a bit value.
    • Updating the context based on the reconstructed bit value.

U.S. Patent No. 9,008,184 - "Multiple sign bit hiding within a transform unit", issued April 14, 2015

The Invention Explained

  • Problem Addressed: After identifying the non-zero coefficients in a block of video data, the encoder must also signal whether each one is positive or negative. Signaling these "sign bits" can consume a substantial amount of data in the final bitstream (Compl. ¶58; ’184 Patent, col. 1:65-67). A technique called "sign bit hiding" existed, which used the mathematical parity (even or odd) of the sum of all coefficients in a block to implicitly signal the sign of one coefficient, but it was limited to being applied only to entire data blocks at a time (Compl. ¶59; ’184 Patent, col. 8:28-34).
  • The Patented Solution: The invention improves upon this technique by applying sign bit hiding to multiple, smaller "sets of coefficients" within a single larger transform block. By partitioning a block into these smaller sets, a sign bit can be hidden for each set, compounding the data-saving benefits of the technique (Compl. ¶59; ’184 Patent, col. 8:29-32). This allows for greater data reduction than was possible when the technique was applied only once to the entire block.
  • Technical Importance: This method enhances video compression efficiency by significantly reducing the amount of data required to encode the signs of transform coefficients (Compl. ¶56).

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 (Compl. ¶94).
  • Claim 1 recites a method for decoding a bitstream of encoded video, comprising the steps of:
    • Reconstructing coefficients for a transform unit where the bitstream encodes two or more sets of sign bits.
    • Each set corresponds to a respective non-overlapping set of coefficients.
    • For each of the sets, summing the absolute value of the coefficients to obtain a parity value.
    • Assigning a sign to one of the coefficients within that set based on whether the parity value is even or odd.

Multi-Patent Capsule: U.S. Patent No. 12,186,395

  • Patent Identification: U.S. Patent No. 12,186,395, "Multiple sign bit hiding within a transform unit", issued January 7, 2025.
  • Technology Synopsis: This patent, which is a continuation of the ’184 Patent, further refines the sign bit hiding technique. It introduces a method where sign bit hiding is applied to a subset of coefficients only if that subset meets a certain threshold, such as having a minimum number of non-zero coefficients. This allows for greater efficiency by selectively applying the data-saving technique only in cases where it is most beneficial (Compl. ¶62; ’395 Patent, col. 9:18-25).
  • Asserted Claims: At least independent claim 5 is asserted (Compl. ¶105).
  • Accused Features: The complaint accuses TikTok's encoders and decoders of infringing this patent in their processing of uploaded and streamed videos (Compl. ¶97).

Multi-Patent Capsule: U.S. Patent No. 8,964,849

  • Patent Identification: U.S. Patent No. 8,964,849, "Multi-level significance maps for encoding and decoding", issued February 24, 2015.
  • Technology Synopsis: This patent addresses the inefficiency of signaling significance maps, which indicate the location of non-zero data coefficients. The invention reduces the amount of data that must be sent by allowing certain significance flags to be inferred by the decoder under specific circumstances, rather than being explicitly signaled in the bitstream. This results in decoding high-quality video using less data (Compl. ¶¶65-66; ’849 Patent, col. 18:11-24).
  • Asserted Claims: At least independent claim 1 is asserted (Compl. ¶117).
  • Accused Features: The complaint accuses the decoders within the TikTok application and on TikTok's servers of infringing this patent (Compl. ¶108).

Multi-Patent Capsule: U.S. Patent No. 12,088,843

  • Patent Identification: U.S. Patent No. 12,088,843, "Method for deriving a motion vector", issued September 10, 2024.
  • Technology Synopsis: This patent addresses motion vector prediction, a process used to track movement between video frames. The invention seeks to solve the problem of efficiently creating a list of candidate motion vectors, where too many candidates increases memory requirements and too few leads to unreliable predictions. The patented solution determines candidate motion vectors for a current block based on the motion vectors of a co-located block in a previously decoded frame, reducing complexity while maintaining reliability (Compl. ¶69; ’843 Patent, col. 4:15-20).
  • Asserted Claims: At least independent claim 1 is asserted (Compl. ¶128).
  • Accused Features: The complaint accuses the encoders and decoders used in transcoding and video processing on the TikTok platform of infringement (Compl. ¶120).

Multi-Patent Capsule: U.S. Patent No. 12,341,962

  • Patent Identification: U.S. Patent No. 12,341,962, "Multi-level significance maps for encoding and decoding", issued June 24, 2025.
  • Technology Synopsis: This patent, related to the ’849 Patent, describes further improvements to efficiently encoding transform coefficient information. It provides methods for grouping significant-coefficient flags and using indirect signaling to convey information about a group of flags. By reducing the amount of data needed to send this transform information, the invention enables higher quality video to be transmitted using less bandwidth (Compl. ¶72).
  • Asserted Claims: At least independent claim 1 is asserted (Compl. ¶139).
  • Accused Features: The complaint accuses the encoders used in transcoding and video encoding within the TikTok platform of infringement (Compl. ¶131).

III. The Accused Instrumentality

Product Identification

The accused instrumentalities are the decoders and encoders within the TikTok video platform, which includes both the TikTok application distributed to users and the servers used to process video content (Compl. ¶¶13-15).

Functionality and Market Context

The complaint alleges that encoding and decoding video are central to TikTok's business model (Compl. ¶2). Functionally, users upload videos to TikTok's servers, which then use transcoders (containing decoders and encoders) to process the video for storage and subsequent streaming to its user base. The TikTok application on user devices also contains decoders and encoders for video playback and recording (Compl. ¶¶13-14). The complaint alleges these instrumentalities decode and encode video according to the H.265 standard, thereby infringing the asserted patents (Compl. ¶15).

IV. Analysis of Infringement Allegations

The complaint references, but does not include, claim chart exhibits that detail its infringement theories. The narrative infringement theory is summarized below.

The complaint’s core infringement theory is that the asserted patents are essential to the H.265/HEVC video coding standard (Compl. ¶32). It alleges that TikTok's encoders and decoders, by complying with this standard, necessarily practice the methods claimed in the asserted patents (Compl. ¶15). For each asserted patent, the complaint alleges that TikTok's making, using, selling, and importing of H.265-compliant decoders and/or encoders constitutes direct infringement (e.g., Compl. ¶¶75-76, 86-87). This standards-based infringement theory suggests that any product correctly implementing the relevant mandatory portions of the H.265 standard will infringe. The complaint provides a screenshot showing a video playing on the TikTok platform as evidence of the accused functionality in use (Compl. p. 22).

  • Identified Points of Contention:
    • Scope Questions: A primary question will be whether the asserted claims are truly essential to the H.265 standard. This raises the question of whether compliance with the standard requires practicing the claimed methods, or if the relevant portions of the standard can be implemented in a non-infringing manner.
    • Technical Questions: The analysis will depend on evidence demonstrating that TikTok's specific implementations of the H.265 standard on its servers and in its application practice each and every limitation of the asserted claims.

V. Key Claim Terms for Construction

For the ’338 Patent:

  • The Term: "partition set"
  • Context and Importance: The invention is premised on a "non-spatially-uniform partitioning" of a significance map. The definition of "partition set," and whether it is limited to the specific nine-context arrangement recited in claim 1 or can encompass other non-uniform arrangements, will be central to determining the claim's scope and its applicability to the H.265 standard's requirements.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The specification describes the general concept of a "non-spatially-uniform partitioning of the map into parts, wherein the bit positions within each part are associated with a given context" (’338 Patent, col. 3:10-14).
    • Evidence for a Narrower Interpretation: Claim 1 itself recites a highly detailed structure with exactly nine different contexts assigned to specific quadrants and rows/columns of a 4x4 transform unit (’338 Patent, col. 23:58 - col. 24:22). This detailed recitation may suggest the term is limited to this specific structure or ones with similar characteristics.

For the ’184 Patent:

  • The Term: "set of coefficients"
  • Context and Importance: The patent's contribution is applying sign bit hiding to multiple "sets" within a transform unit. The scope of what constitutes a "set"—whether it must align with predefined "coefficient groups" used for significance map coding or can be any arbitrary grouping of coefficients—is critical to the infringement analysis.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The patent abstract describes partitioning a transform unit into "two or more sets of coefficients" without further structural limitation (’184 Patent, Abstract). The detailed description also mentions dividing a transform unit "into sets of non-zero coefficients" more generally (’184 Patent, col. 8:29-31).
    • Evidence for a Narrower Interpretation: The specification suggests that larger transform blocks could be partitioned into "coefficient groups or 'sets of significant coefficient flags'," potentially linking the claimed "set" to these pre-existing structures from the H.265 standard (’184 Patent, col. 7:60-62).

VI. Other Allegations

  • Indirect Infringement: The complaint alleges induced infringement for all asserted patents. The theory is that by distributing the TikTok application and encouraging its use, TikTok induces its end-users to directly infringe the patents when they record, upload, or play back videos (e.g., Compl. ¶¶77, 88). The complaint alleges this inducement is supported by TikTok's affirmative acts, such as providing instructions and encouragement through prompts like "Get [the] App" (Compl. p. 22) and by requiring users to grant permissions for the application to access the camera and microphone to perform the infringing video recording functions (Compl. p. 23).
  • Willful Infringement: For each asserted patent, the complaint alleges that TikTok has had knowledge and notice of the patent and its infringement "since at least as of the date this lawsuit was filed and/or the date this Original Complaint was served" (e.g., Compl. ¶¶74, 85). This forms the basis for a claim of post-suit willful infringement.

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

This dispute appears to center on the licensing of patents allegedly essential to the H.265 video coding standard. The key questions for the case will likely be:

  • A core issue will be one of standard essentiality: Can Velos prove that the asserted patent claims are necessarily infringed by any device or service that properly implements the mandatory portions of the H.265/HEVC standard, or has TikTok implemented the standard in a way that avoids the claimed methods?
  • A central question will concern RAND licensing obligations: The case involves allegations of failed good-faith negotiations from both sides (Compl. ¶¶144-151, 155-158). A key determination for the court will be whether the Plaintiff's licensing offers were consistent with its RAND commitments and whether the Defendant failed to negotiate in good faith to take such a license.
  • The ultimate financial question will be one of royalty determination: If infringement and essentiality are found, the court will need to determine an appropriate RAND royalty. This will likely involve a complex analysis of the incremental value the patented technologies contribute to the H.265 standard and the TikTok platform, rather than a royalty based on the total revenue of the accused product.