Change-Id: I9d18f351abe7614107f34f47eeb38a234a9937c9

Change-Id: Ib27fc4f3dbe01fe8adfa04a61aaba21b3480e75c

Change-Id: Ia7f640ca395e8deaac5986f19d11ab18d85eec2d

Moving because q_index is used only by encoder.

Change-Id: I0b96175614ed4fd3d76ee56a0ba36258e1e896f6

Match any whitespace instead of individual spaces. The macro
definitions in vp9/common/arm/neon/vp9_short_idct32x32_1_add_neon.asm
triggered this and treated spaces as arguments leading to lines like:
$8vld1$8.$88$8           {$8q8$8}, [$$89$8], $$8stride$8

Change-Id: I2d5718aba4614e4fd7b702e15c2a1bd80e656bd2

Change-Id: I60e02fa3de930ff1f969687ab5af93dee40d86ad

These changes are to support automated regressions of vpx on android
	new file:   test/android/Android.mk
	new file:   test/android/README
	new file:   test/android/get_files.py

Change-Id: I52c8e9daf3676a3561badbe710ec3a16fed72abd

Change-Id: I4b5106bdc08fd4551339b968c13428a8f43122e2

As Jim suggested, 1D array was used to store filter levels instead
of 2D array. This used shift_y in setup_mask directly, and saved
few cycles.

Change-Id: If61ab298784861f1806b1cd396d4e4e2e0f097b9

1920x1080

vp90-2-08-tile-4x4.webm
vp90-2-08-tile-4x4.webm.md5

vpxenc blue_sky_1080p25 --frame-parallel=0 \
--tile-columns=2 --tile-rows=2 \
--lossless=0 --error-resilient=0 \
--min-q=0 --max-q=30 --kf-min-dist=0 --kf-max-dist=9999 \
--cpu-used=0 --codec=vp9 -p 2 --good --fps=25/1 --limit=5

vp90-2-08-tile-4x1.webm
vp90-2-08-tile-4x1.webm.md5

vpxenc blue_sky_1080p25 --frame-parallel=0 \
--tile-columns=0 --tile-rows=2 \
--lossless=0 --error-resilient=0 \
--min-q=0 --max-q=30 --kf-min-dist=0 --kf-max-dist=9999 \
--cpu-used=0 --codec=vp9 -p 2 --good --fps=25/1 --limit=5

Change-Id: I77d0ccf9b721ae5cfa72781fe0d3c534ea43b756

iOS doesn't recognize B:
bad instruction `B idct32_pass_loop'

Change-Id: I3cf6aede4639f1d9efa97f7962fa287ba6feaaef

Implements scan order to band map with arrays in both the encoder
and decoder to remove conditional statements.

Encoding seems to be about 1% faster at speed 0, tested on football.
Decoding seems to be about 0.5-1% faster on a set of 25 videos.

Change-Id: Idb233ca0b9e0efd790e30880642e8717e1c5c8dd

Removing foreach_predicted_block_visitor and calling build_inter_predictors
directly.

Change-Id: I11bb3c872b99b47c2680b01b0dbcc01c558c4a2b

This commit enables the dual buffer rate-distortion optimization
and encoding scheme. It stacks the original transform coefficients,
quantized levels, and reconstructed coefficients, in the rate-
distortion optimization search process, hence eliminates the need
to re-run residual generation, forward transform, and quantization
in the encoding stage.

Change-Id: I011bfad3a59a380a869ee552e91dae0394ec492e

xd->mi_8x8 is assigned inside set_offsets() for each prediction block.

Change-Id: I20e5974a9eaf105e5a04fc7f99b7a93bd50e3d0a

Change-Id: Iaee740ec3bfb2b5328c24f4641c285e5a4a046dc

Added loop filter mask checking, and made the caller function
ready for implementation of parallel loopfiltering in horizontal
direction.

Next, we need to go through the loopfilter functions (both c and
optimized versions), and provide 16-byte wide loopfiltering for
each filter type.

Change-Id: Ifef47e7ef9086ebc2fd6ca7ede8f27c9bbf79e66

Change-Id: I6b245ef5017c49373ba33e30f8fbf3d1b97409b5

Allocate memory space of dual buffer sets that store the coeff, qcoeff,
dqcoeff, and eobs. Connect the pointers of macroblock_plane and
macroblockd_plane to the actual buffer in use accordingly.

Change-Id: I2f0b5f482ca879fae39095013eaf8901db20a5a4