| 摘要 |
Video signal processing apparatus in which at least two input video signals are combined in proportions determined by a pixel key signal to generate an output video signal for compression, at least one of the input video signals each having respective associated compression parameters from a data compression process applied to that video signal; comprises: means for estimating the quantity of data which will be produced by compression of a current image of the output video signal; means for comparing the detected quantity of data with a target quantity of data, to determine whether a data overflow is expected; means, operable in the event that a data overflow is expected, for determining which blocks of the output video signal can be compressed by re-using compression parameters from corresponding blocks of one of the input video signals, the determining means comprising: (f) means for detecting blocks of the output image derived entirely from a single one of the input video signals, whereby blocks of the output image not derived entirely from a single one of the input video signals are compressed using newly derived compression parameters; (g) means for detecting an expected amount of data overflow and deriving a threshold value from the expected amount of overflow in accordance with a predetermined algorithm; and (h) means for detecting whether a quantization parameter associated with the compression parameters of a block under test, that block being derived entirely from a single one of the input video signals, lies within a proportion representing the least-harshly quantizing quantization parameters across the respective input image, the proportion being defined by the threshold value, blocks for which the quantization parameter lies within the said proportion being compressed by re-using the compression parameters associated with the corresponding block of the input video. |
