This speeds up a normal YT style 1080P clip decode by ~1% on nexus 7.

Change-Id: Ied7fa0d8bc941b2adb4db9382f549ee4d5654f3a

Change-Id: I7fb3a78ed0e0bc940d8b4a57c470302f8369782f

While CONFIG_INTERNAL_STATS=1, PSNR is calculated while encoding.
The aligned width/height were used mistakenly in the calculation.
This patch fixed it, and used the orignal image width/height.

Change-Id: Iad5334f8693af761b71ebb78f2587db8a3404ecf

* changes:
  vp9_fdct8x8_quant_ssse3: quiet a static analysis warning
  vp9_fdct8x8_quant_sse2: quiet a static analysis warning
  vp9_mv_pred: quiet a static analysis warning

add explicit returns in cases where ASSERT_* can't be used due to the
function returning a value; retain the EXPECT_* for reporting purposes.

Change-Id: I1f514728537fee42a99277d3aba538e832d3b65b

document resolution assumptions with a few asserts

Change-Id: Ia4ab738fd3e0a1ba0ed30a57facd2658c2c1fd60

add an assert to validate 'in' array size

Change-Id: Ie5a24275c066d9dd59714f6104510abbd4850dc5

add an assert to validate 'in' array size

Change-Id: Ib72946a86f34e1ce8a69954e8e3e4fe1a0f18a91

add an assert to validate pred_mv array size

Change-Id: I532b882b71e2baff3ac76e07ed133ec5a11bd0fc

* changes:
  Remove last remnants of obj_int_extract
  Initialize "corrupted" to zero

Factor in segment#2 and skip blocks into the postencode estimated bits,
and increase somewhat the aggressiveness of the refresh.

PSNR/SSIM Metrics on RTC set go up by ~0.8/0.5%.

Change-Id: I5d4e7cb00a3aefb25d18c88b6b24118b72dc5d51

This commit makes the encoder to explicitly calculate the SAD
associated with the LAST_FRAME motion vector and compare it to
that of the GOLDEN_FRAME given by integral projection motion
estimation. It skips the expensive sub-pixel motion search over
GOLDEN_FRAME when the LAST_FRAME can provide fairly good motion
compensated prediction quality.

For dark720p speed -6 single thread goes from
33304 b/f, 40.070 dB, 18156 ms ->
33319 b/f, 40.061 dB, 17611 ms

Change-Id: I01bc94b9b598075567a392111046b97a9bc30efe

Change-Id: I3e275544bff478849c1b5f3dcd5de950ee330d14

Change-Id: Icc7da6027763b5ed7cbfe70ffe271103ead59fe1

Because the call to vpx_codec_control at line 928 is now guarded by
!frame_parallel, 'corrupted' may not be set.

Change-Id: Id166bd8a8cdb5e5120fca1640011a3545f6e178a

Make the C and SSE2 versions consistent.

Change-Id: I03c405d22a36bd1a97480efb96dc5af230667424

Use force_split to constrain the partition selection.
This is used because in the top-down approach to variance partition,
a block size may be selected even though one of its subblocks may have
high variance.

In this patch the selection of the 64x64 block size will only
be allowed if the variance of all the 32x32 subblocks are also below the threshold.

Stil testing, but some visual improvement for areas near slow moving boundary
can be seen. Metrics for RTC set increase by about ~0.5%.

Change-Id: Iab3e7b19bf70f534236f7a43fd873895a2bb261d

Change-Id: I9101cb1132aa580fc32fd47e8ac9d91d5b09e963

Code refactoring.

Change-Id: I73b6fcc0444155ee46c1efa5253c1d608c6439cb

Change-Id: I5169155b20ea3676a6ce58ec77d6aeba07db29d9

Shut off all the metric checks for golden reference frame, if we
decide that it is unlikely to be selected for reference.

Change-Id: Ie457cc1fd43935584403b4982659aed80fb9909c

Change-Id: I0ba02d501d65a2f219fa1bb87b8eb81f7e6ab332

These were uncovered by the chromoting perftest.

Change-Id: Ia5a90fd1718ff757c1484decf3861295260e6722

The mvs are clamped in the vp9_find_best_ref_mvs() already.

Change-Id: I9bea5e35aef6007466fe7fca4bc2dc5c17e74222