VVC Test Model

Description

The VVC Test Model (VTM) is the reference software implementation for the Versatile Video Coding (VVC) standard, also known as H.266 and MPEG-I Part 3. VVC was developed jointly by ITU-T Study Group 16 (VCEG) and ISO/IEC JTC 1/SC 29 (MPEG) under the Joint Video Experts Team (JVET). The VTM serves as the common test platform for evaluating new coding tools and for verifying the performance of the standard. It is not a product implementation but rather a research and development tool that embodies the normative and non-normative aspects of VVC.

In the context of 3GPP, VTM is referenced in specifications such as 26.927 and 26.955, which relate to video codec testing and performance evaluation for 5G media services. 3GPP adopts video codecs like VVC to enable efficient video streaming and immersive media over 5G networks. The VTM is used by 3GPP to assess the performance of VVC in terms of compression efficiency, computational complexity, and error resilience, ensuring that the codec meets the requirements of 5G use cases such as enhanced Mobile Broadband (eMBB) and ultra-reliable low-latency communication (URLLC).

The VTM software includes a complete encoder and decoder, along with tools for bitstream analysis and conformance testing. It implements all the coding tools defined in the VVC standard, including advanced intra and inter prediction, transform and quantization, loop filters, and entropy coding. The VTM is updated regularly as new tools are adopted into the standard. Each version of the VTM corresponds to a specific draft of the VVC specification, allowing researchers to experiment with proposed modifications and measure their impact.

Key components of the VTM include the encoder, which compresses raw video into a VVC-compliant bitstream, and the decoder, which reconstructs the video from the bitstream. The VTM also includes a common test conditions (CTC) framework that defines the test sequences, encoding configurations, and objective quality metrics (such as PSNR and BD-rate) used for performance evaluation. This framework ensures that results from different laboratories are comparable. The VTM is typically written in C++ and is available as open source.

VTM is important because it provides a neutral, standardized platform for developing and benchmarking video coding technology. Without a common test model, it would be difficult to compare different coding tools and to reach consensus on the best techniques. The VTM has been instrumental in achieving the significant bitrate reductions of VVC compared to its predecessor HEVC (H.265). For 3GPP, the VTM helps ensure that VVC is suitable for 5G media delivery, including 4K/8K streaming, cloud gaming, and extended reality (XR). As video traffic continues to dominate mobile networks, the VTM's role in optimizing video compression remains critical.

Purpose & Motivation

The VTM exists to provide a common reference implementation for the development and evaluation of VVC coding tools. During the standardization process, experts from various companies and research institutions propose new coding tools. To assess these proposals fairly, a common test model is needed that implements the current state of the standard. The VTM fulfills this role by providing a consistent platform for experiments, ensuring that results are reproducible and comparable across different organizations.

Another purpose of the VTM is to serve as a basis for conformance testing. Once the VVC standard is finalized, implementations must be tested to ensure they comply with the specification. The VTM can be used as a reference to generate test bitstreams and to verify decoder behavior. This helps guarantee interoperability between different VVC products.

Historically, similar test models have been used for previous video coding standards, such as the HEVC Test Model (HM) and the Advanced Video Coding Test Model (JM). The VTM continues this tradition, incorporating lessons learned and adapting to the more complex tools of VVC. It addresses the need for a rigorous, collaborative development environment that can handle the increasing complexity of modern video codecs.