Method and apparatus for encoding video signal and method and apparatus for decoding video signal
Provided are a method and apparatus for encoding a video signal and a method and apparatus for decoding a video signal. The method of encoding the video signal includes: determining an optimum scaling bit depth of the video signal in consideration of a maximum bit number of an intermediate value obtained during transformation of the video signal and a maximum bit precision of the video signal; scaling the video signal according to a local shift value corresponding to a difference between the determined optimum scaling bit depth and an original bit depth of the video signal; and transforming the scaled video signal.
1. A method for decoding a video, the method comprising:
obtaining, by at least one processor, quantized transform coefficients of a transform block from a bitstream;
generating, by the at least one processor, scaled transform coefficients of the transform block by performing inverse-quantization and scaling on the quantized transform coefficients;
generating, by the at least one processor, intermediate sample values by performing vertical inverse-transformation on the scaled transform coefficients;
generating, by the at least one processor, residual values by performing horizontal inverse-transformation on the intermediate sample values;
generating, by the at least one processor, a prediction block by performing intra prediction or inter prediction; and,
restoring, by the at least one processor, sample values using the residual values of the transform block and the prediction block,
wherein the scaling on the quantized transform coefficients comprises:
determining an offset value based on a scaling variable; and
adding the inverse-quantized transform coefficients to the offset value and bit shifting on the added transform coefficients by the scaling variable,
wherein the scaling variable is generated by using a size of the transform block and the bit depth of samples,
wherein the vertical inverse-transformation is performed by multiplying a first transform matrix by the scaled transform coefficients, and
wherein the horizontal inverse-transformation is performed by multiplying a second transform matrix by the intermediate sample values.
2. The method of claim 1 , wherein each of the vertical inverse-transformation and the horizontal inverse-transformation includes linear transformation that is a discrete cosine transform (DCT) operation or a discrete wavelet transform (DWT) operation.