Decoding a block of video samples
A method of decoding a block of video samples from a video bitstream is provided in which the method includes decoding a DQPRP value, deriving a PQP value for a block having a non-zero residual, decoding a DDQP value for the block, deriving a SDQP value for the block from the DDQP value and the DQPRP value, deriving a block QP value for the block from the SDQP value and the PQP value, decoding at least one non-zero coefficient of the block, deriving values of a residual block for the block by first scaling the at least one non-zero coefficient using the QP value and thereafter applying an inverse transform, deriving values of a prediction block spatial or temporal prediction, and deriving decoded sample values for the block by adding the values of the residual block and the values of the prediction block.
1. A method of decoding a block of video samples having non-zero coefficients from a video bitstream, the method being performed by a video decoder and comprising:
decoding a Delta Quantization Parameter Resolution Parameter, DQPRP, value from the video bitstream;
deriving a Predicted Quantization Parameter, PQP, value for a block having a non-zero residual;
decoding a Decoded Block Delta Quantization Parameter, DDQP, value for the block;
deriving a Scaled Block Delta Quantization Parameter, SDQP, value for the block from the DDQP value and the DQPRP value;
deriving a block Quantization Parameter, QP, value for the block from the SDQP value and the PQP value;
decoding at least one non-zero coefficient of the block from the video bitstream;
deriving values of a residual block for the block by first scaling the at least one non-zero coefficient using the QP value and thereafter applying an inverse transform;
deriving values of a prediction block for the block by spatial or temporal prediction; and
deriving decoded sample values for the block by adding the values of the residual block and the values of the prediction block.
2. The method according to claim 1 , wherein the deriving a SDQP value for the block comprises deriving the SDQP value as SDQP=DDQP*(DQPRP+1).
3. The method according to claim 1 , wherein the scaling the at least one of the non-zero coefficients, coeff, using the QP value comprises computing:
coeff*scaling_factor*levelScale[QP mod n ]<<(QP/ n ),
where scaling_factor is a scaling factor that depends on a size of the inverse transform, levelScale is an array with n entries, and QP is the derived block QP value.
4. The method according to claim 1 , further comprising rounding the derived PQP value such that it only can take values that the deltaQP with reduced resolution can have.
5. The method according to claim 1 , wherein the DQPRP value depends on a size of the block, such that a resolution represented by the DQPRP value is coarser for smaller blocks than for larger blocks.
6. The method according to claim 1 , wherein the DQPRP value depends on any one of: a QP value signaled in a Picture Parameter Set (PPS), a QP value associated with a picture comprising the block, a QP value associated with a slice comprising the block, and the PQP.
7. The method according to claim 1 , wherein the DQPRP value depends on a layer identifier of a layer comprising the block.
8. The method according to claim 1 , wherein the DQPRP value depends on a type of a slice comprising the block.
9. A video decoder for decoding a block of video samples having non-zero coefficients from a video bitstream, the video decoder comprising a processor and a memory comprising instructions which, when executed by the processor, cause the video decoder to:
decode a Delta Quantization Parameter Resolution Parameter, DQPRP, value from the video bitstream;
derive a Predicted Quantization Parameter, PQP, value for a block having a non-zero residual;
decode a Decoded Block Delta Quantization Parameter, DDQP, value for the block,
derive a Scaled Block Delta Quantization Parameter, SDQP, value for the block from the DDQP value and the DQPRP value;
derive a block Quantization Parameter, QP, value for the block from the SDQP value and the PQP value;
decode at least one non-zero coefficient of the block from the video bitstream;
derive values of a residual block for the block by first scaling the at least one non-zero coefficient using the QP value and thereafter applying an inverse transform;
derive values of a prediction block for the block by spatial or temporal prediction; and
derive decoded sample values for the block by adding the values of the residual block and the values of the prediction block.
10. The video decoder according to claim 9 , wherein the instructions cause the video decoder to derive a SDQP value for the block by deriving the SDQP value as SDQP=DDQP*(DQPRP+1).
11. The video decoder according to claim 9 , wherein the instructions cause the video decoder to scale the at least one of the non-zero coefficients, coeff, using the QP value by computing:
coeff*scaling_factor*levelScale[QP mod n ]<<(QP/ n ),
where scaling_factor is a scaling factor that depends on a size of the inverse transform, levelScale is an array with n entries, and QP is the derived block QP value.
12. The video decoder according to claim 9 , wherein the instructions further cause the video decoder to round the derived PQP value such that it only can take values that the deltaQP with reduced resolution can have.
13. The video decoder according to claim 9 , wherein the DQPRP value depends on a size of the block, such that a resolution represented by the DQPRP value is coarser for smaller blocks than for larger blocks.
14. The video decoder according to claim 9 , wherein the DQPRP value depends on any one of: a QP value signaled in a Picture Parameter Set (PPS), a QP value associated with a picture comprising the block, a QP value associated with a slice comprising the block, and the PQP.
15. The video decoder according to claim 9 , wherein the DQPRP value depends on a layer identifier of a layer comprising the block.
16. The video decoder according to claim 9 , wherein the DQPRP value depends on a type of a slice comprising the block.
17. A non-transitory computer readable storage medium storing a computer program comprising computer-executable instructions for causing a processing unit of a video decoder to perform a method of decoding a block of video samples having non-zero coefficients from a video bitstream, the method comprising:
decoding a Delta Quantization Parameter Resolution Parameter, DQPRP, value from the video bitstream;
deriving a Predicted Quantization Parameter, PQP, value for a block having a non-zero residual;
decoding a Decoded Block Delta Quantization Parameter, DDQP, value for the block;
deriving a Scaled Block Delta Quantization Parameter, SDQP, value for the block from the DDQP value and the DQPRP value;
deriving a block Quantization Parameter, QP, value for the block from the SDQP value and the PQP value;
decoding at least one non-zero coefficient of the block from the video bitstream;
deriving values of a residual block for the block by first scaling the at least one non-zero coefficient using the QP value and thereafter applying an inverse transform;
deriving values of a prediction block for the block by spatial or temporal prediction; and
deriving decoded sample values for the block by adding the values of the residual block and the values of the prediction block.
18. The method according to claim 2 , wherein the scaling the at least one of the non-zero coefficients, coeff, using the QP value comprises computing:
coeff*scaling_factor*levelScale[QP mod n ]<<(QP/ n ),
where scaling_factor is a scaling factor that depends on a size of the inverse transform, levelScale is an array with n entries, and QP is the derived block QP value.
19. The method according to claim 2 , further comprising rounding the derived PQP value such that it only can take values that the deltaQP with reduced resolution can have.
20. The video decoder according to claim 10 , wherein the instructions cause the video decoder to scale the at least one of the non-zero coefficients, coeff, using the QP value by computing:
coeff*scaling_factor*levelScale[QP mod n ]<<(QP/ n ),
where scaling_factor is a scaling factor that depends on a size of the inverse transform, levelScale is an array with n entries, and QP is the derived block QP value.