Reduced encoder time complexity using enhanced adaptive multiple transform selection in versatile video coding

Abstract

When compressing a video, the architecture is provided with the width and height of the frame along with the video as input. The multiple transform selection (MTS) techniques are used to compress video. Traditional discrete cosine transforms (DCT) and other trigonometric (sine and cosine) transforms, such as DCT-8 and DST-7, are included in the pool of transforms. For the best outcomes, two additional transforms-DST-2 and DCT-5-were also utilized to compress any remaining video frames. To improve coding effectiveness, peak signal-to-noise ratio (PSNR), and delay, 2D Transforms are broken into 1D Directional Transforms for block sizes ranging from 4×4 to 64×64. The temporal complexity was greatly decreased by adding DST-2 to the existing video test model (VTM). The MTS’s adaptive multiple transform (AMT) scheme already included a pool of transforms, but DST-2 was introduced as a fourth transform. Encoder time complexity was reduced by 22% on average without compromising PSNR and bitrate.