Commit Graph

1635 Commits

Author SHA1 Message Date
Scott LaVarnway
c34d91a84e Merge "Packing bitstream on-the-fly with delayed context updates" 2012-03-01 06:20:02 -08:00
Yunqing Wang
aabae97e57 vpxenc: fix time and fps calculation in 2-pass encoding
When we do 2-pass encoding, elapsed time is accumulated through
whole 2-pass process, which gives incorrect time and fps results
for second pass. This change fixed that by resetting the time
accumulator for second pass.

Change-Id: Ie6cbf0d0e66e6874e7071305e253c6267529cf20
2012-02-29 15:44:56 -05:00
Attila Nagy
52cf4dcaea Packing bitstream on-the-fly with delayed context updates
Produce the token partitions on-the-fly, while processing each MB.
Context is updated at the beginning of each frame based on the
previoud frame's counters. Optimally encoder outputs partitions in
separate buffers. For frame based output, partitions are concatenated
internally.

Limitations:
    - enabled just in combination with realtime-only mode
    - number of encoding threads has to be equal or less than the
    number of token partitions. For this reason, by default the encoder
    will do 8 token partitions.
    - vpxenc supports partition output (-P) just in combination with
    IVF output format (--ivf)

Performance:
    - Realtime encoder can be up to 13% faster (ARM) depending on the number
    of threads and bitrate settings. Constant gain over the 5-16 speed
    range.
    - Token buffer reduced from one frame to 8 MBs

Quality:
    - quality is affected by the delayed context updates. This again
    dependents on input material, speed and bitrate settings. For VC
    style input the loss seen is up to 0.2dB. If error-resilient=2
    mode is used than the effect of this change is negligible.

Example:
./configure --enable-realtime-only --enable-onthefly-bitpacking
./vpxenc --rt --end-usage=1 --fps=30000/1000 -w 640 -h 480
--target-bitrate=1000 --token-parts=3 --static-thresh=2000
--ivf -P -t 4 -o strm.ivf tanya_640x480.yuv

Change-Id: I127295cb85b835fc287e1c0201a67e378d025d76
2012-02-29 12:13:37 -05:00
Scott LaVarnway
ce328b855f Merge changes Ifb450710,I61c4a132
* changes:
  Eliminated reconintra_mt.c
  Eliminated vp8mt_build_intra_predictors_mbuv_s
2012-02-28 11:42:45 -08:00
Scott LaVarnway
aab70f4d7a Merge "Removed duplicate code in threading.c" 2012-02-28 11:25:43 -08:00
Scott LaVarnway
bcba86e2e9 Eliminated reconintra_mt.c
Reworked the code to use vp8_build_intra_predictors_mby_s,
vp8_intra_prediction_down_copy, and vp8_intra4x4_predict_d_c
functions instead.  vp8_intra4x4_predict_d_c is a decoder-only
version of vp8_intra4x4_predict.  Future commits will fix this
code duplication.

Change-Id: Ifb4507103b7c83f8b94a872345191c49240154f5
2012-02-28 14:12:30 -05:00
Scott LaVarnway
9a4052a4ec Removed duplicate code in threading.c
Change-Id: Id7e44950ceda67b280e410e541510106ef02f1da
2012-02-28 14:00:32 -05:00
Yunqing Wang
b1bfd0ba87 Merge "Only do uv intra-mode evaluation when intra mode is checked" 2012-02-28 10:11:24 -08:00
Yunqing Wang
019384f2d3 Only do uv intra-mode evaluation when intra mode is checked
When we encode slide-show clips, for the majority of the time,
only ZEROMV mode is checked, and all other modes are skipped.
This change delayed uv intra-mode evaluation until intra mode is
actually checked. This gave big performance gain for slide-show
video encoding (2nd pass gain: 18% to 28%). But, this change
doesn't help other types of videos.

Also, zbin_mode_boost is adjusted in mode-checking loop, which
causes bitstream mismatch before/after this change when --best
or --good with --cpu-used=0 are used.

Change-Id: I582b3e69fd384039994360e870e6e059c36a64cc
2012-02-28 13:08:17 -05:00
James Berry
e2c6b05f9a bugfix: use oxcf width/height for reinit check
use oxcf instead of common in check to Reinit the
lookahead buffer if the frame size changes
prior behavior would cause assertion fail/crash

first observed in:
support changing resolution with vpx_codec_enc_config_set

Change-Id: Ib669916ca9b4f206d4cc3caab5107e49d39a36aa
2012-02-27 16:10:45 -05:00
Yunqing Wang
61c5e31ca1 Merge "Fix skippable evaluation in mode decision" 2012-02-27 11:06:13 -08:00
John Koleszar
ad1216151d Merge "vpxenc: initial implementation of multistream support" 2012-02-27 09:59:14 -08:00
John Koleszar
02a31e6b3c Merge "decoder: reset segmentation map on keyframes" 2012-02-27 09:58:29 -08:00
Yunqing Wang
84be08b07f Fix skippable evaluation in mode decision
Yaowu fixed the skippable evaluation by correcting 2nd order
block's eob.

Change-Id: Id47930cbc74a90a046c0c0e324efb03477639ee0
2012-02-27 12:45:12 -05:00
James Berry
313bfbb6a2 Merge "Add unit tests for idctllm_test and idctllm_mmx" 2012-02-23 08:50:36 -08:00
Jim Bankoski
2089f26b08 Merge "Remove the frame rate factor for key frame size." 2012-02-23 08:38:44 -08:00
Marco Paniconi
507ee87e3e Remove the frame rate factor for key frame size.
When temporal layers is used (i.e., number_of_layers > 1),
we don't use the frame rate boost for setting the key
frame target size. The factor was forcing the target size to be
always at its minimum (2* per_frame_bandwidth) for low frame rates
(i.e., base layer frame rate).

Generally we should modify or remove this frame rate factor;
for now we turn if off for number_of_layers > 1.

Change-Id: Ia5acf406c9b2f634d30ac2473adc7b9bf2e7e6c6
2012-02-22 15:25:32 -08:00
Scott LaVarnway
f2bd11faa4 Eliminated vp8mt_build_intra_predictors_mbuv_s
Reworked the code to use vp8_build_intra_predictors_mbuv_s
instead.  This is WIP with the goal of eliminating all
functions in reconintra_mt.h

Change-Id: I61c4a132684544b24a38c4a90044597c6ec0dd52
2012-02-21 14:59:05 -05:00
James Berry
0c1cec2205 Add unit tests for idctllm_test and idctllm_mmx
add unit tests for vp8_short_idct4x4llm_c

Change-Id: I472b7c0baa365ba25dc99a3f6efccc816d27c941
2012-02-21 14:52:36 -05:00
John Koleszar
dadc9189ed Merge changes I0341554f,I64e110c8
* changes:
  Consolidate C version of token packing functions
  Multithreaded encoder, late sync loopfilter
2012-02-21 10:09:23 -08:00
Scott LaVarnway
f05feab7b9 Merge "Remove redundant init of segment_counts in vp8_encode_frame" 2012-02-21 09:51:02 -08:00
John Koleszar
02360dd2c2 Merge "Update encoder mb_skip_coeff and prob_skip_false calculation" 2012-02-21 09:48:26 -08:00
Johann
b0a12a2880 Refine offset pattern
When compiling with -ggdb3 the output includes an extraneous EQU from
vpx_ports/asm_offsets.h

https://trac.macports.org/ticket/33285

Change-Id: Iba93ddafec414c152b87001a7542e7a894781231
2012-02-17 12:28:13 -08:00
John Koleszar
b5ce9456db Merge changes Idf1a05f3,If227b29b,Iac784d39
* changes:
  vpxenc: factor out input open/close
  vpxenc: add warning()/fatal() helpers
  vpxenc: factor out global config options
2012-02-17 11:14:17 -08:00
Johann
e6047a17a9 Merge "OS X shell is incompatible with echo -n" 2012-02-17 10:53:19 -08:00
Yunqing Wang
f93b1e7be1 Merge "Fix incorrect use of uv eobs in intra modes" 2012-02-17 10:43:05 -08:00
Yunqing Wang
04b9e0d787 Fix incorrect use of uv eobs in intra modes
In vp8_rd_pick_inter_mode(), if total of eobs is zero, rate needs
to be adjusted since there are no non-zero coefficients for
transmission. The uv intra eobs calculated in
rd_pick_intra_mbuv_mode() need to be saved before they are
overwritten by inter-mode eobs.

Change-Id: I41dd04fba912e8122ef95793d4d98a251bc60e58
2012-02-17 09:15:08 -05:00
Attila Nagy
ce42e79abc Update encoder mb_skip_coeff and prob_skip_false calculation
mode_info_context->mbmi.mb_skip_coeff has to always reflect the
existence or not of coeffs for a certain MB. The loopfilter needs this
info.
mb_skip_coeff is either set by the vp8_tokenize_mb or has to be set to
1 when the MB is skipped by mode selection. This has to be done
regardless of the mb_no_coeff_skip value.

prob_skip_false is needed just when mb_no_coeff_skip is 1. No need to
keep count of both skip_false and skip_true as they are complementary
(skip_true+skip_false = total_mbs)

Change-Id: I3c74c9a0ee37bec10de7bb796e408f3e77006813
2012-02-17 14:27:40 +02:00
Attila Nagy
565d0e6feb Remove redundant init of segment_counts in vp8_encode_frame
segment_counts was zero init twice in the beginning of vp8_encode_frame.

Change-Id: Ibc29f6896dabd9aab1d0993f3941cf6876022e70
2012-02-17 09:51:24 +02:00
Johann
6b151d436d Clarify 'max_sad' usage
Depending on implementation the optimized SAD functions may return early
when the calculated SAD exceeds max_sad.

Change-Id: I05ce5b2d34e6d45fb3ec2a450aa99c4f3343bf3a
2012-02-16 15:17:44 -08:00
Johann
5f0b303c28 OS X shell is incompatible with echo -n
Built in echo in 'sh' on OS X does not support -n (exclude trailing
newline). It's not necessary so just leave it off. Fixes issue 390.

Build include guard using 'symbol' so that it is more likely to be
unique.

Change-Id: I4bc6aa1fc5e02228f71c200214b5ee4a16d56b83
2012-02-16 14:20:44 -08:00
Fritz Koenig
3653fb473a Include path fix for building against Android NDK.
cpu-features.h is not in the common paths, add
to the cflags for Android.

Change-Id: Icbafc7600d72f6b59ffb030f6ab80ee6860332bb
2012-02-16 12:38:17 -08:00
John Koleszar
9e50ed7f27 vpxenc: initial implementation of multistream support
Add the ability to specify multiple output streams on the command line.
Streams are delimited by --, and most parameters inherit from previous
streams.

In this implementation, resizing streams is still not supported. It
does not make use of the new multistream support in the encoder either.
Two pass support runs all streams independently, though it's
theoretically possible that we could combine firstpass runs in the
future. The logic required for this is too tricky to do as part of this
initial implementation. This is mostly an effort to get the parameter
passing and independent streams working from the application's
perspective, and a later commit will add the rescaling and
multiresolution support.

Change-Id: Ibf18c2355f54189fc91952c734c899e5c072b3e0
2012-02-16 12:30:01 -08:00
John Koleszar
732cb9a643 vpxenc: factor out input open/close
Simplify some of the file I/O for later commits which will add multistream
support

Change-Id: Idf1a05f3a29c95331d0c4a6ea5960904e4897fd4
2012-02-16 12:30:00 -08:00
John Koleszar
c535025c12 vpxenc: add warning()/fatal() helpers
Cosmetic. Allows exiting with an error message without opening a new
scope.

Change-Id: If227b29b825f0241acea79dd38f19e524552ee18
2012-02-16 12:26:58 -08:00
John Koleszar
e8223bd250 decoder: reset segmentation map on keyframes
Refactoring some of the mode decoding logic introduced a bug where
the segmentation maps would not be properly reset on keyframes.

http://code.google.com/p/webm/issues/detail?id=378

The text of the bug is somewhat misleading as I initially read it to
imply the bug was present in v0.9.7-p1 (Cayuga), but note the text
"master", which indicates this was something subsequent. This issue
bisects back to v0.9.7-p1-84-ga99c20c, so unfortunately it was broken
during the Duclair release.

Thanks to Alexei Leonenko for investigating the root cause.

Change-Id: I9713c9f070eb37b31b3b029d9ef96be9b6ea2def
2012-02-16 12:22:18 -08:00
Makoto Kato
7989bb7fe7 Support Android x86 NDK build
On Android NDK, rand() is inlined function.  But, on our SSE optimization,
we need symbol for rand()

Change-Id: I42ab00e3255208ba95d7f9b9a8a3605ff58da8e1
2012-02-16 12:03:30 -08:00
Scott LaVarnway
6776bd62b5 Simplify mb_to_x_edge calculation during mode decoding
Change-Id: Ibcb35c32bf24c1d241090e24c5e2320e4d3ba901
2012-02-16 13:36:46 -05:00
Scott LaVarnway
a5879f7c81 Merge "decodemv cleanup/improvements" 2012-02-16 09:33:59 -08:00
Scott LaVarnway
12ee845ee7 decodemv cleanup/improvements
Removed unnecessary variables, unrolled functions, eliminated
unnecessary mv bounds checks and branches.

Change-Id: I02d034c70cd97b65025d59dd67c695e1db529f0b
2012-02-16 11:38:33 -05:00
Attila Nagy
d02e74a073 Consolidate C version of token packing functions
Replace inner loops of pack_mb_row_tokens_c and
pack_tokens_into_partitions_c with a call to pack_tokens_c.

Change-Id: I0341554fb154a14a5dadb63f8fc78010724c2c33
2012-02-16 14:11:28 +02:00
Attila Nagy
78071b3b97 Multithreaded encoder, late sync loopfilter
Second shot at this...

Sync with loopfilter thread as late as possible, usually just at the
beginning of next frame encoding. This returns control to application
faster and allows a better multicore scaling.

When PSNR packets are generated the final filtered frame is needed
imediatly so we cannot delay the sync. Same has to be done when
internal frame is previewed.

Change-Id: I64e110c8b224dd967faefffd9c93dd8dbad4a5b5
2012-02-16 12:26:39 +02:00
John Koleszar
efd54f8f41 vpxenc: factor out global config options
This is a first step towards specifying multiple output streams
with one command line.

Change-Id: Iac784d3911bf553694d024bbd0c3d547261e914b
2012-02-15 16:11:35 -08:00
Fritz Koenig
8144132866 Fix rtcd build process for Android.mk
Add a dependency so ndk-build will
generate the needed vpx_rtcd.h file.

Change-Id: I92c82e0996943dd0403c9956e1ba60e92e2837a9
2012-02-15 15:23:04 -08:00
John Koleszar
e6df50031e Merge "support changing resolution with vpx_codec_enc_config_set" 2012-02-10 16:18:00 -08:00
Johann
169823428f Missed some variance casts
Change-Id: I9fb510f9421fb3c317a8e32e3058cee977ddf9fa
2012-02-10 11:07:33 -08:00
Johann
12d45f62f6 Merge "max_sad check is not always implemented" 2012-02-10 10:28:00 -08:00
Johann
8c50a70a95 max_sad check is not always implemented
As an optimization some architectures use the max_sad argument to break
out early from the SAD. Pass in INT_MAX instead of 0 to prevent this.

Change-Id: I653c476834b97771578d63f231233d445388629d
2012-02-09 16:19:10 -08:00
Johann
fea3556e20 Fix variance overflow
In the variance calculations the difference is summed and later squared.
When the sum exceeds sqrt(2^31) the value is treated as a negative when
it is shifted which gives incorrect results.

To fix this we cast the result of the multiplication as unsigned.

The alternative fix is to shift sum down by 4 before multiplying.
However that will reduce precision.

For 16x16 blocks the maximum sum is 65280 and sqrt(2^31) is 46340 (and
change).

PPC change is untested.

Change-Id: I1bad27ea0720067def6d71a6da5f789508cec265
2012-02-09 12:38:31 -08:00
John Koleszar
2e0d55314c Merge "Add OS/2 supports" 2012-02-08 11:00:55 -08:00