| 摘要 |
<p>This invention relates to video coding, in particular a method of accurately estimating the bits generated by a video sequence to be encoded at a constant quality, utilising parameters associated with previously encoded video sequences. The bitstream characteristics from the video sequence encoded with two different perceptual qualities is used, including the average quantiser and bits generated per segment. These parameters are used in an algorithm to predict the bits produced per segment that will be generated from the video sequence encoded at a third perceptual quality. The invention yields data on bit sizes for each chunk without having to perform the actual encoding, and is thus much faster when used in conjunction with determining if a video sequence encoded at a given perceptual quality will meet certain practical transmission constraints.</p> |
