Prediction using extra-buffer samples for intra block copy in video coding
A method of visual media processing is provided. The method includes determining a size of a buffer to store reference samples for prediction in an intra block copy mode, and performing a conversion between a current video block of visual media data and a bitstream of the current video block. The conversion is performed using the reference samples stored in the buffer, and in the intra block copy (IBC) mode. The IBC mode is based on motion information related to a reconstructed block located in same video region with the current video block without referring to a reference picture.
1. A method of processing video data, comprising:
determining, for a conversion between a current video block in a first video region of visual media data and a bitstream of the current video block, a virtual intra block copy (IBC) buffer that stores reference samples for prediction in an IBC mode, wherein the conversion is performed in the IBC mode which is based on motion information related to a reconstructed block located in the first video region without referring to a reference picture;
resetting the virtual IBC buffer before performing the conversion for the current video block;
generating reconstructing samples for a first video block of the first video region;
storing the reconstructing samples in the virtual IBC buffer without applying a filtering operation on the reconstructing samples;
for a first reference sample, of the reference samples, spatially located at location (x0, y0) of the current video block and having a block vector (BVx, BVy), computing a first reference location (P, Q), wherein the first reference location (P, Q) is determined using the block vector (BVx, BVy) and a location (x0, y0);
upon determining that the first reference location (P, Q) lies outside the virtual IBC buffer, re-computing a second reference sample of the reference samples;
deriving a block vector for the current video block; and
generating, based on the virtual IBC buffer and the block vector, prediction samples for the current video block,
wherein a virtual IBC buffer derived from the first video region is disabled for a second video region which is different from the first video region,
wherein the first video region and the second video region correspond to two different coding tree unit (CTU) rows, wherein the two different CTU rows are in a common picture, and
wherein the conversion includes encoding the current video block into the bitstream or decoding the current video block from the bitstream.
2. The method of claim 1 , wherein the first reference location (P, Q) is determined as P=x0+BVx and Q=y0+BVy.
3. The method of claim 1 , wherein a location of the second reference sample in the virtual IBC buffer is represented as P mod M, clip (Q, 0, N−1)) where “mod” is modulo operation, “clip” is clipping operation, and M and N are integers representing horizontal and vertical dimensions of the virtual IBC buffer.
4. The method of claim 1 , wherein a location of the second reference sample in the virtual IBC buffer is represented as (clip (P, 0, M−1), Q mod N) where “mod” is modulo operation, “clip” is clipping operation, and M and N are integers representing horizontal and vertical dimensions of the virtual IBC buffer.
5. The method of claim 2 , wherein Q is constrained to lie within a pre-defined range [0, . . . , N−1], where N is an integer representing a vertical dimension of the virtual IBC buffer.
6. The method of claim 2 , wherein P is constrained to lie within a pre-defined range [0, . . . , M−1], where M is an integer representing a horizontal dimension of the virtual IBC buffer.
7. The method of claim 2 , wherein P and Q are constrained to lie within respective pre-defined ranges [0, . . . , M−1] and [0, . . . , N−1], respectively, and where M and N are integers representing x and y dimensions of the virtual IBC buffer.
8. The method of claim 1 , further comprising:
upon determining that the block vector (BVx, BVy) lies outside the virtual IBC buffer, padding a sample value to the second reference sample.
9. The method of claim 8 , wherein the sample value is a predefined value.
10. The method of claim 8 , wherein the sample value, denoted as 1, is represented as 1<<(Bitdepth-1), where Bitdepth is a bit-depth of the current video block.
11. The method of claim 10 , wherein the sample value equals 512 when the bit-depth of the current video block is 10.
12. The method of claim 10 , wherein the sample value equals 128 when the bit-depth of the current video block is 8.
13. The method of claim 8 , wherein the sample value, denoted as 1, is represented as 1=forwardLUT(m), where m=1<<(Bitdepth-1), where Bitdepth is a bit-depth of the current video block and forwardLUT( ) is a lookup-table operation.
14. The method of claim 8 , wherein the sample value is included as a field in the bitstream.
15. The method of claim 8 , wherein the sample value corresponds to a nearest sample value of a sample in the virtual IBC buffer.
16. An apparatus for processing video data comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to:
determine, for a conversion between a current video block in a first video region of visual media data and a bitstream of the current video block, a virtual intra block copy (IBC) buffer that stores reference samples for prediction in an IBC mode, wherein the conversion is performed in the IBC mode which is based on motion information related to a reconstructed block located in the first video region without referring to a reference picture;
reset the virtual IBC buffer before performing the conversion for the current video block;
generate reconstructing samples for a first video block of the first video region;
store the reconstructing samples in the virtual IBC buffer without applying a filtering operation on the reconstructing samples;
for a first reference sample, of the reference samples, spatially located at location (x0, y0) of the current video block and having a block vector (BVx, BVy), compute a first reference location (P, Q), wherein the first reference location (P, Q) is determined using the block vector (BVx, BVy) and a location (x0, y0);
upon determining that the first reference location (P, Q) lies outside the virtual IBC buffer, re-compute a second reference sample of the reference samples;
derive a block vector for the current video block; and
generate, based on the virtual IBC buffer and the block vector, prediction samples for the current video block,
wherein a virtual IBC buffer derived from the first video region is disabled for a second video region which is different from the first video region,
wherein the first video region and the second video region correspond to two different coding tree unit (CTU) rows, wherein the two different CTU rows are in a common picture, and
wherein the conversion includes encoding the current video block into the bitstream or decoding the current video block from the bitstream.
17. The apparatus of claim 16 , wherein the first reference location (P, Q) is determined as P=x0+BVx and Q=y0+BVy.
18. A non-transitory computer-readable recording medium storing a bitstream of a video which is generated by a method performed by a video processing apparatus, wherein the method comprises:
determining, for generating the bitstream based on a current video block in a first video region of visual media data, a virtual intra block copy (IBC) buffer that stores reference samples for prediction in an IBC mode, wherein the generating is performed in the IBC mode which is based on motion information related to a reconstructed block located in the first video region without referring to a reference picture;
resetting the virtual IBC buffer before generating the bitstream for the current video block;
generating reconstructing samples for a first video block of the first video region;
storing the reconstructing samples in the virtual IBC buffer without applying a filtering operation on the reconstructing samples;
for a first reference sample, of the reference samples, spatially located at location (x0, y0) of the current video block and having a block vector (BVx, BVy), computing a first reference location (P, Q), wherein the first reference location (P, Q) is determined using the block vector (BVx, BVy) and a location (x0, y0); and
upon determining that the first reference location (P, Q) lies outside the virtual IBC buffer, re-computing a second reference sample of the reference samples;
deriving a block vector for the current video block; and
generating, based on the virtual IBC buffer and the block vector, prediction samples for the current video block,
wherein a virtual IBC buffer derived from the first video region is disabled for a second video region which is different from the first video region,
wherein the first video region and the second video region correspond to two different coding tree unit (CTU) rows, wherein the two different CTU rows are in a common picture, and
wherein the generating the bitstream includes encoding the current video block into the bitstream.
19. The apparatus of claim 16 , wherein the first reference location (P, Q) is determined as P=x0+BVx and Q=y0+BVy.
20. The apparatus of claim 19 , wherein P and Q are constrained to lie within respective pre-defined ranges [0, . . . , M−1] and [0, . . . , N−1], respectively, and where M and N are integers representing x and y dimensions of the virtual IBC buffer.