Michael Niedermayer
8428e2c5f3
avcodec/jpeg2000dwt: Print 1 digit less in the 9/7f DWT test
...
This avoids test failure due to differing rounding between 32 and 64bit x86
Found-by: jamrial
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 23:18:38 +02:00
Michael Niedermayer
b1fdf81c6e
avcodec/jpeg2000dwt: use 32x32->64 multiplies in the 9/7i DWT
...
This significantly improves the quality when the integer 9/7 transform
is used
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 02:27:50 +02:00
Michael Niedermayer
6c7b1597c7
avcodec/jpeg2000: Move H band scaling from wavelet into quantization code
...
This reduces the number of operations done and is equivalent except for
rounding
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 02:00:57 +02:00
Michael Niedermayer
1c495b0bf6
avcodec/jpeg2000: Move gainb handling into the quantization code
...
thats how the specification defines it, this also improves numerical
accuracy of the integer wavelet implementation. It otherwise should
be equivalent, in case of overflows this can be reverted.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 01:36:05 +02:00
Michael Niedermayer
3bcf61f495
avcodec/jpeg2000dwt: Fix value of F_LFTG_X and I_LFTG_X
...
Even if the jpeg2000 spec uses a wrong value this does not
make mathematics work this way, also this has been corrected in the 2004
version AFAIK
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-21 19:25:35 +02:00
Michael Niedermayer
4299f085f4
avcodec/jpeg2000dwt: also test 9/7 float DWT
...
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-21 15:52:53 +02:00
Michael Niedermayer
6528b90dea
avcodec/jpeg2000dwt: List differences for DWT test
...
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-21 13:43:02 +02:00
