~60-70% faster depending on the block size

Change-Id: Icdbaa9977a91a63cbcc6ead0cf19d5a2af7f27e1

Change-Id: Iba03852ce778c956200818e3473cfb2b48cf8d8e

Removed unused tx_select_threshes and tx_select_diff.

Change-Id: I5e9e7ad170056efe14b5f071e94d0c5a36e4a34c

This commit replaces vp9_idct.h with txfm_common.h in many SIMD
implementation files for precise file dependency.

Change-Id: If73dd726bb16537e7494f28538b0a169810f9756

Separate the common coefficient constant into vpx_dsp/txfm_common.h.
Move the SSE2 macro definitions to vpx_dsp/x86/txfm_common_sse2.h.
This clears the use case of vp9_idct.h in vpx_dsp folder.

Change-Id: I319735a2abf42888e5080ac14cfbcde34be7b121

Change-Id: I283d364a4e65ca9bf6ff581da1d0b498433c5402

Refactor and clean up the msa transform related code layout.

Change-Id: Ic5048bd3d62a6046589817da745370ea89448e44

Avoid scaling the references if they have already been scaled.
Change only affects 1 pass non-svc mode for now.

Change-Id: I204f4079c026cba7adce7a7f855d072f6139ccec

The RD and load save/code grabs it as groups of four. In practice there
is no change to physical allocations becaquse this is backed by a 16-byte
memalign.

Change-Id: I01e89769872300e23227e03dd24a6e229f482025

Change-Id: I5368f5fc7283420c38d5bd85e3077b761d94ace6

Add vpx_dsp_rtcd.h to the header file list. The od_bin_fdct8x8()
here depends on forward 8x8 2D-DCT.

Change-Id: I1d71edc71f07069808823d2445c1cafd285e1b94

This commit factors out common macro definitions from the forward
and inverse transform implementations into vpx_dsp. It removes
the duplicate macro definitions from encoder and decoder folders.

Change-Id: I92301acbd3317075e9c5f03328a25abb123bca78

The msa macros are locally used and should not be named with VP9
prefix.

Change-Id: I2c9c746c4027383c16b9ab12b77b4e70e7e7d206

This commit factors the 4x4, 8x8, and 16x16 2D-DCT forward
transform operations into vpx_dsp folder.

Change-Id: I084b117b79c0925edcbcabb93f62b9f4bf8dbe7d

Change-Id: Ie4d958b26e586db218f8ee95d5df4bf11f2345a1

Change-Id: I0ac7beaa160a6c2a60a019f6b8ce85e6537bed7d

changes prefix vp9_ to vpx_ for non codec specific functions and
data structures.

Change-Id: I91a21548e39bd24d2c7caaaa223ae47240bb78c8

change prefix vp9_ to vpx_ for non codec specific functions and data
structures.

Change-Id: I97c7e6422eceea99212b93f4942bc2187763a07c

Replace vp9_ in names to vpx_ as they are not codec specific.

Change-Id: I2e583aa63dee769353ada4b42417aa15c4074ebb

Remove redundant file dependency.

Change-Id: I4708218157617dabe00e2e33e237be2838c16603

Remove the redundant function declarations therein.

Change-Id: I27731fb70bb1abce63da761a5812f518c62f590f

Change-Id: I963f08f1023481712c6f9ed624ddf05e5bac6321

This commit limits the scope of 1-D DCT and ADST functions within
vp9_dct.c and makes them static. This largely clears out the cross
referencing issue between vp9_dct.c and the SIMD optimizations.

Change-Id: If7cac478b11bb32328ccf70a9f60b709dad43d7f

Change-Id: I07647c7482b9ec498fbad3a9c9901f72b2336500

Change-Id: Id27e0007a0feac821ca66bcecbf3a723305da82d

The SSE2 version high bit-depth forward hybrid transforms are
essentially using the C functions via cross referencing to 1-D
functions in vp9_dct.c. This commit unifies the two versions and
removes the unnecessary dependency.

Change-Id: Ib4d0702a138f8daf7d0bd97c141ee7088f293765

Change-Id: Ib919a8ec2ec66d460d2f8a26d72aabc09dcbbd72

Separate the hybrid transform case from 2D-DCT case. This will
allow us to clear up cross dependency between c and SIMD
implementations later.

Change-Id: Iaa499e8b096850a1c5a0c50a3b6e63e15d0184bf

Change-Id: I0426126d0a65f13f9250983e44cc366b1b1a9c4a

The following quantization functions were moved:
vp9_quantize_b
vp9_quantize_b_32x32
vp9_highbd_quantize_b
vp9_highbd_quantize_b_32x32

vp9_quantize_dc
vp9_quantize_dc_32x32
vp9_highbd_quantize_dc
vp9_highbd_quantize_dc_32x32

The purpose of doing that was to allow these functions to be shared
by multiple codecs.

Change-Id: Id8ab939f283353cdd07bd930d47db3d932a5d87f

Change-Id: Ib1cb1fbe92a39ff5312cee069559be6d3ea458d0

Remove the use of drop_frames_water_mark, as this is used for
frame dropping control. Use fixed threshold for now on buffer underflow.

Change-Id: If0ddda9f7f6fa96067cdcb0eccb42e17bda37c32

In aq-mode=3 under a resizing action (i.e., resize_pending != 0),
force an update of the golden reference frame.

Change-Id: I14806f6db71b5f8c827678cc5e1fc913c138a9a4

Fix bug in setting this flag for animated content.
The bug did cause quality to increase because far
more frames are not boosted than boosted.

However, the speed trade off to gain is a lot less
favorable and the behavior was not as intended.

Change-Id: I89fb70419c88b26f40b3534de0481730a1b3fcfa

Use drop_frames_water_mark for threshold on buffer underflow,
and change threshold for resize down.

Change-Id: I2de19adce50abe9bcdc0b107528cec8cc1857fcc

BUG=https://code.google.com/p/webm/issues/detail?id=1023

Change-Id: I212a1d67b23ce3b5ce08800de369b25b9e375e7d

The fast scaling for 1 pass mode was being used only on the
first frame after resizing event (because resize_scale_num/den
is set to 1 and only changed for first frame following resize event).

Change-Id: I723b63e21823eb858f25f5662d2bbe4f1842e61f

Proper use/update of resize_state and resize_pending to constrain
the total amount of downsizing to be at most one scale down, for now.

Change-Id: Id18fc32499f2fbdbec16728dcdc9e4eac09098f0

From Change  Ibf0c30b72074b3f71918ab278ccccc02a95a70a0
There is still an issue relating to one animated test clip with repeat
patterns where this change effectively increase the default  maximum
arf interval by +1. This can be examined seperately.

Change-Id: Idd01d5480fc45202d8a059a0c3afc0997cc5bdd1

This commit simplifies the intra block boundary condition logic.
It removes the block index from the argument set.

Change-Id: If00142512eb88992613d6609356dfd73ba390138