Intra prediction modes for lossy coding when transform is skipped
A video coder applies a residual differential pulse code modulation technique to a residual data of a block coded using lossy coding. The block may be coded without application of a transform to the residual data.
1. A method of decoding video data, the method comprising:
receiving a bitstream that comprises an encoded representation of a picture of the video data; and
reconstructing the picture based at least in part on syntax elements in the bitstream, wherein reconstructing the picture comprises:
using an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes including intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
generating a residual block, wherein the residual block is a block for which transform application is skipped, the residual block consists of reconstructed residual values, and generating the residual block comprises generating the reconstructed residual values in the residual block such that the reconstructed residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and reconstructed sample values, wherein, for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block, and Q(r i,j ) is a reconstructed residual value in the residual block for a residual value r i,j at position (i,j):
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q({tilde over (r)} i,j ) denotes a quantized version of a residual value {tilde over (r)} i,j , the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and Q(r i−1,j ) is a reconstructed residual value for a residual value one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
and
Q(r i,j−1 ) is a reconstructed residual value for a residual value one column left of the residual value r i,j , and the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
;
and
for 0≦i≦(M−1) and 0≦j≦(N−1), adding the reconstructed residual value Q(r i,j ) to a prediction value of the predictive block to reconstruct one of the reconstructed sample values.
2. The method of claim 1 , further comprising obtaining, from the bitstream, one or more syntax elements indicating the quantized residual value Q({tilde over (r)} i,j ).
3. The method of claim 1 , wherein for a diagonal intra prediction mode in the plurality of intra prediction modes, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
Q
(
r
~
i
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed residual value for a residual value one column left of and one row above the residual value r i,j .
4. The method of claim 3 , wherein, for the diagonal intra prediction mode, the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed residual value for a residual value r i−1,j−1 that is one column left and one row above the residual value r i,j .
5. The method of claim 1 , wherein the residual value r i,j is a bit-shifted residual value.
6. A method of encoding video data, the method comprising:
generating an encoded representation of a picture of the video data; and
generating a bitstream that comprises the encoded representation of the picture,
wherein generating the encoded representation of the picture comprises:
using an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes including intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
generating a residual block, wherein the residual block is a block for which transform application is skipped, the residual block consists of residual values, and generating the residual block comprises:
generating the residual values in the residual block such that the residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and original sample values; and
after generating the residual values:
for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block:
determining a residual value {tilde over (r)} i,j for a residual value r i,j , wherein:
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q(r (i−1),j ) denotes a reconstructed residual value for a residual value r i−1,j one row above the residual value r i,j ,
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
and
Q(r i,(j−1) ) denotes a reconstructed residual value for a residual value r i,j−1 one column left of the residual value r i,j ; and
quantizing the residual value {tilde over (r)} i,j to produce a quantized residual value Q({tilde over (r)} i,j ).
7. The method of claim 6 , wherein generating the bitstream comprises signaling the quantized residual value Q({tilde over (r)} i,j ) in the bitstream.
8. The method of claim 6 , wherein:
for a vertical intra prediction mode in the plurality of intra prediction modes, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and
for a horizontal intra prediction mode in the plurality of intra prediction modes, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
.
9. The method of claim 8 , further comprising:
adding Q(r i,j ) to a prediction value to reconstruct a sample value; and
using the sample value for intra prediction or inter prediction of other blocks.
10. The method of claim 6 , wherein for the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and
for the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
.
11. The method of claim 6 , wherein for a diagonal intra prediction mode in the plurality of intra prediction modes, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed residual value for a residual value r i−1,j−1 that is one column left and one row above the residual value r i,j .
12. The method of claim 11 , wherein:
for the diagonal intra prediction mode, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
Q
(
r
~
i
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed version of a residual value one column left of and one row above the residual value r i,j .
13. The method of claim 6 , wherein the residual value r i,j is a bit-shifted residual value.
14. A video decoding apparatus comprising:
a memory configured to store video data; and
one or more processors configured to:
receive a bitstream that comprises an encoded representation of a picture of the video data; and
use an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes including intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
reconstruct the picture based at least in part on syntax elements in the bitstream, wherein the one or more processors are configured such that, as part of reconstructing the picture, the one or more processors:
generate a residual block, wherein the residual block is a block for which transform application is skipped, the residual block consists of reconstructed residual values, and to generate the residual block, the one or more processors generate the reconstructed residual values in the residual block such that the reconstructed residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and reconstructed sample values, wherein, for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block, and Q(r i,j ) is a reconstructed residual value Q(r i,j ) in the residual block for a residual value r i,j at position (i,j):
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and
Q({tilde over (r)} i,j ) denotes a quantized version of a residual value {tilde over (r)} i,j , the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and Q(r i−1,j ) is a reconstructed residual value for a residual value one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
Q(r i,j−1 ) is a reconstructed residual value for a residual value one column left of the residual value r i,j , and the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
;
and
for 0≦i≦(M−1) and 0≦j≦(N−1), add the reconstructed residual value Q(r i,j ) to a prediction value of the predictive block to reconstruct one of the reconstructed sample values.
15. The video decoding apparatus of claim 14 , wherein the one or more processors are configured to obtain, from a bitstream, one or more syntax elements indicating the quantized residual value Q({tilde over (r)} i,j ).
16. The video decoding apparatus of claim 14 , wherein for a diagonal intra prediction mode in the plurality of intra prediction modes, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
Q
(
r
~
i
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed residual value for a residual value one column left of and one row above the residual value r i,j .
17. The video decoding apparatus of claim 16 , wherein, for the diagonal intra prediction mode, the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
.
18. The video decoding apparatus of claim 14 , wherein the residual value r i,j is a bit-shifted residual value.
19. The video decoding apparatus of claim 14 , wherein the video decoding apparatus comprises at least one of:
an integrated circuit;
a microprocessor; and
a wireless communication device that includes the one or more processors.
20. A video encoding apparatus comprising:
a memory configured to store video data; and
one or more processors configured to:
generate an encoded representation of a picture of the video data; and
generate a bitstream that comprises the encoded representation of the picture,
wherein the one or more processors are configured such that, as part of generating the encoded representation of the picture, the one or more processors:
use an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes includes intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
generate a residual block, wherein the residual block is a block for which transform application is skipped and the residual block consists of residual values, and
wherein the one or more processors are configured such that, as part of generating the residual block, the one or more processors:
generate the residual values in the residual block such that the residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and original sample values; and
after generating the residual values:
for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of a block and N is the width of a block:
determine a residual value {tilde over (r)} i,j for a residual value r i,j , wherein:
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q(r (i−1),j ) denotes a reconstructed residual value for a residual value r i−1,j one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
and
Q(r i,(j−1) ) denotes a reconstructed residual value for a residual value r i,j−1 one column left of the residual value r i,j ; and
quantize the residual value {tilde over (r)} i,j to produce a quantized residual value Q({tilde over (r)} i,j ).
21. The video encoding apparatus of claim 20 , further comprising signaling the quantized residual value Q({tilde over (r)} i,j ) in a bitstream.
22. The video encoding apparatus of claim 20 , wherein
for a vertical intra prediction mode in the plurality of intra prediction modes, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and
for a horizontal intra prediction mode in the plurality of intra prediction modes, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
.
23. The video encoding apparatus of claim 22 , wherein the one or more processors are configured to:
add the reconstructed residual value Q(r i,j ) to a prediction value to reconstruct a sample value; and
use the sample value for intra prediction or inter prediction of other blocks.
24. The video encoding apparatus of claim 20 , wherein for the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, the quantized residual value Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and
for the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
.
25. The video encoding apparatus of claim 20 , wherein for a diagonal intra prediction mode in the plurality of intra prediction modes, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed residual value for a residual value r i−1,j−1 that is one column left and one row above the residual value r i,j .
26. The video encoding apparatus of claim 25 , wherein, for the diagonal intra prediction mode, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
Q
(
r
~
i
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
j
=
0
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
-
1
)
,
1
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
wherein Q(r i−1,j−1 ) is a reconstructed version of a residual value one column left of and one row above the residual value r i,j ; and
the one or more processors are configured to add the reconstructed residual value Q(r i−1,j−1 ) to a prediction value to reconstruct a sample value.
27. The video encoding apparatus of claim 20 , wherein the residual value r i,j is a bit-shifted residual value.
28. The video encoding apparatus of claim 20 , wherein the video encoding apparatus comprises at least one of:
an integrated circuit;
a microprocessor; and
a wireless communication device that includes the one or more processors.
29. A video decoding apparatus comprising:
means for receiving a bitstream that comprises an encoded representation of a picture of video data; and
means for reconstructing the picture based at least in part on syntax elements in the bitstream, wherein the means for reconstructing the picture comprises:
means for using an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes including intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
means for generating a residual block, wherein the residual block is a block for which transform application is skipped, the residual block consists of reconstructed residual values, and the means for generating the residual block comprises means for generating the reconstructed residual values in the residual block such that the reconstructed residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and reconstructed sample values, wherein, for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block, and Q(r i,j ) is a reconstructed residual value in the residual block for a residual value r i,j at position (i,j):
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
-
1
,
j
)
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q({tilde over (r)} i,j ) denotes a quantized version of a residual value {tilde over (r)} i,j , the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and
Q(r i−1,j ) is a reconstructed residual value for a residual value one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
and
Q(r i,j−1 ) is a reconstructed residual value for a residual value one column left of the residual value r i,j , and the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
;
and
for 0≦i≦(M−1) and 0≦j≦(N−1), means for adding the reconstructed residual value Q(r i,j ) to a prediction value of the predictive block to reconstruct one of the reconstructed sample values.
30. A video encoding apparatus comprising:
means for generating an encoded representation of a picture of video data; and
means for generating a bitstream that comprises the encoded representation of the picture,
wherein the means for generating the encoded representation of the picture comprises:
means for using an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes includes intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
means for generating a residual block, wherein the residual block is a block for which transform application is skipped, the residual block consists of residual values, and the means for generating the residual block comprises:
means for generating the residual values in the residual block such that the residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and original sample values; and
for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block:
means for determining, after generating the residual values, a residual value {tilde over (r)} i,j for a residual value r i,j , wherein:
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q(r (i−1),j ) denotes a reconstructed residual value for a residual value r i−1,j one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
Q(r i,(j−1) ) denotes a reconstructed residual value for a residual value r i,j−1 one column left of the residual value r i,j ; and
means for quantizing the residual value {tilde over (r)} i,j to produce a quantized residual value Q({tilde over (r)} i,j ).
31. A non-transitory computer-readable data storage medium having instructions stored thereon that, when executed, cause a video decoding apparatus to:
receive a bitstream that comprises an encoded representation of a picture of video data; and
reconstruct the picture based at least in part on syntax elements in the bitstream, wherein as part of causing the video decoding apparatus to reconstruct the picture, the instructions cause the video decoding apparatus to:
use an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes including intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
generate a residual block, wherein the residual block is a block for which transform application is skipped, the residual block consists of reconstructed residual values, and as part of causing the video decoding apparatus to generate the residual block, the instructions cause the video decoding apparatus to generate the reconstructed residual values in the residual block such that the reconstructed residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and reconstructed sample values, wherein, for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block, and Q(r i,j ) is a reconstructed residual value in the residual block for a residual value r i,j at position (i,j):
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, Q(r i,j ) is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q({tilde over (r)} i,j ) denotes a quantized version of a residual value {tilde over (r)} i,j , the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
and Q(r i−1,j ) is a reconstructed residual value for a residual value one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, Q(r i,j ) is defined as:
Q
(
r
i
,
j
)
=
{
Q
(
r
~
i
,
j
)
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
,
Q
(
r
~
i
,
j
)
+
Q
(
r
i
,
j
-
1
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
Q(r i,j−1 ) is a reconstructed residual value for a residual value one column left of the residual value r i,j , and the residual value {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
;
and
for 0≦i≦(M−1) and 0≦j≦(N−1), add the reconstructed residual value Q(r i,j ) to a prediction value of the predictive block to reconstruct one of the reconstructed sample values.
32. A non-transitory computer-readable data storage medium having instructions stored thereon that, when executed, cause a video encoding apparatus to:
generate an encoded representation of a picture of video data; and
generate a bitstream that comprises the encoded representation of the picture,
wherein as part of causing the video encoding apparatus to generate the encoded representation of the picture, the instructions cause the video encoding apparatus to:
use an intra prediction mode from a plurality of intra prediction modes to generate a predictive block for a prediction unit, the plurality of intra prediction modes includes intra prediction modes having prediction directions between prediction directions of intra prediction modes 22 and 30 and prediction directions between prediction directions of intra prediction modes 6 and 14 as defined in High Efficiency Video Coding (HEVC) text specification draft 10;
generate a residual block for, wherein the residual block is a block for which transform application is skipped, the residual block consists of residual values, and as part of causing the video encoding apparatus to generate the residual block, the instructions cause the video encoding apparatus to:
generate the residual values in the residual block such that the residual values in the residual block indicate differences between sample values of the predictive block for the prediction unit and original sample values; and
after generating the residual values:
for 0≦i≦(M−1) and 0≦j≦(N−1), where M is a height of the residual block and N is the width of the residual block:
determine a residual value {tilde over (r)} i,j for a residual value r i,j , wherein:
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 22 and 30, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
i
=
0
,
0
≤
j
≤
(
N
-
1
)
r
i
,
j
-
Q
(
r
(
i
-
1
)
,
j
)
,
1
≤
i
≤
(
M
-
1
)
,
0
≤
j
≤
(
N
-
1
)
,
Q(r (i−1),j ) denotes a reconstructed residual value for a residual value r i−1,j one row above the residual value r i,j , and
for each of the intra prediction modes having prediction directions between the prediction directions of intra prediction modes 6 and 14, {tilde over (r)} i,j is defined as:
r
~
i
,
j
=
{
r
i
,
j
,
0
≤
i
≤
(
M
-
1
)
,
j
=
0
r
i
,
j
-
Q
(
r
i
,
(
j
-
1
)
)
0
≤
i
≤
(
M
-
1
)
,
1
≤
j
≤
(
N
-
1
)
,
Q(r i,(j−1) ) denotes a reconstructed residual value for a residual value r i,j−1 one column left of the residual value r i,j ; and
quantize the residual value {tilde over (r)} i,j to produce a quantized residual value Q({tilde over (r)} i,j ).
33. The video decoding device of claim 14 , further comprising a display configured to display decoded video data.
34. The video encoding device of claim 20 , further comprising a camera configured to capture the video data.