A Video Encoder, a Video Decoder and Corresponding Methods One Flag That Saves Sending the Same List Twice

This patent describes a signalling shortcut in video compression: a single flag that tells a decoder the second reference picture list is identical to the first, so the encoder does not have to transmit it. It comes from Huawei Technologies Co., Ltd., one of the largest contributors to modern video coding standards.

The Problem

Video is enormous. Even a short clip contains far more data than a network with limited bandwidth can comfortably carry, or than a device would want to store, which is why essentially all video is compressed before it is transmitted and decompressed when it arrives. As networks fill up and viewers expect higher resolutions, every percentage point of compression efficiency gained without visible quality loss is worth having.

A large share of that compression comes from inter-frame prediction: rather than describing each frame from scratch, the encoder describes it as a set of changes from other frames it has already sent. To do that, both encoder and decoder must agree on exactly which previously decoded frames are available as references, and in what order. That agreement is maintained through reference picture lists, and video standards generally use two of them, list 0 and list 1. The lists have to be signalled in the bitstream, and in a great many sequences the candidate entries for the two lists are the same. Sending an identical list twice wastes bits in the header of every relevant unit, which adds up across a long video.

What This Invention Does

The method gives the encoder a way to say “the second one is the same as the first”. A decoder parses a flag. It then parses the candidates for the first reference picture list structure. If the flag has a first value, the decoder determines that candidates for the second reference picture list structure are not present in the bitstream at all, and infers that they are the same as the candidates for the first list. If the flag has a second value, the decoder determines that the second list’s candidates are present and reads them normally.

Either way, the decoder then generates a reference picture list from whichever set of candidates applies, and performs inter-prediction based on that list to reconstruct the picture. The saving is small per instance and large in aggregate: one bit replaces a full duplicated list structure wherever the two lists coincide. The disclosure also covers improved signalling for reference picture marking, the bookkeeping that decides which decoded frames are kept in the buffer for future reference, tying it directly to the reference picture lists rather than maintaining it separately.

Key Features

• Single controlling flag. One parsed flag determines whether the second reference picture list structure is transmitted or inferred.
• Inference of the second list. When the flag takes the first value, the second list’s candidates are inferred to match the first list’s.
• Explicit signalling retained. When the flag takes the second value, the second list’s candidates are read from the bitstream as usual.
• Reference list generation. A reference picture list is generated from whichever candidate set applies.
• Inter-prediction. The generated list is used for inter-prediction to reconstruct the picture.
• List-based reference marking. Reference picture marking is derived directly from the reference picture lists.

Who Is Behind It

The applicant is Huawei Technologies Co., Ltd. of Shenzhen, China. The named inventors are Fnu Hendry and Ye-Kui Wang, both long-standing contributors to international video coding standardisation. The application is a divisional of Australian application 2024200359, with priority from September 2018, placing the work in the development window of the Versatile Video Coding standard.

Why It Matters

Video accounts for the majority of internet traffic, so compression research has an unusually direct effect on infrastructure costs and on what quality of streaming is possible over a constrained connection. Australia’s geography and its mix of fixed and mobile access make bandwidth efficiency more than an abstract concern. Patents like this one also sit inside the standards ecosystem: techniques adopted into a standard such as VVC become standard-essential patents, which are licensed to device and platform makers on defined terms, so protecting them in each significant market is routine practice rather than a signal of a specific product launch.

Related Concepts

• Video coding format – the standards context this technique belongs to.
• Inter frame – the prediction method that relies on reference picture lists.
• Versatile Video Coding – the standard generation this work coincides with.
• Data compression – the broader discipline behind the bit saving.
• Bitstream – the encoded data structure the flag is written into.
• Standard-essential patent – the licensing framework surrounding codec patents.

---

AU 2026202038 was published in the Australian Official Journal of Patents on 2 April 2026 and is open for public inspection. Patent applications represent inventions that are sought to be protected and do not necessarily reflect commercially available products.