[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
/*
|
|
|
|
|
Copyright (c) 2012 The WebM project authors. All Rights Reserved.
|
|
|
|
|
|
|
|
|
|
Use of this source code is governed by a BSD-style license
|
|
|
|
|
that can be found in the LICENSE file in the root of the source
|
|
|
|
|
tree. An additional intellectual property rights grant can be found
|
|
|
|
|
in the file PATENTS. All contributing project authors may
|
|
|
|
|
be found in the AUTHORS file in the root of the source tree.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <cstdio>
|
|
|
|
|
#include <cstdlib>
|
|
|
|
|
#include <string>
|
|
|
|
|
#include "third_party/googletest/src/include/gtest/gtest.h"
|
|
|
|
|
#include "test/codec_factory.h"
|
|
|
|
|
#include "test/encode_test_driver.h"
|
|
|
|
|
#include "test/i420_video_source.h"
|
|
|
|
|
#include "test/util.h"
|
|
|
|
|
#include "test/md5_helper.h"
|
|
|
|
|
extern "C" {
|
|
|
|
|
#include "vpx_mem/vpx_mem.h"
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
namespace {
|
|
|
|
|
class TileIndependenceTest : public ::libvpx_test::EncoderTest,
|
|
|
|
|
public ::libvpx_test::CodecTestWithParam<int> {
|
|
|
|
|
protected:
|
|
|
|
|
TileIndependenceTest() : EncoderTest(GET_PARAM(0)), n_tiles_(GET_PARAM(1)),
|
|
|
|
|
md5_fw_order_(), md5_inv_order_() {
|
|
|
|
|
init_flags_ = VPX_CODEC_USE_PSNR;
|
|
|
|
|
vpx_codec_dec_cfg_t cfg;
|
2013-02-07 00:30:21 +01:00
|
|
|
cfg.w = 704;
|
|
|
|
|
cfg.h = 144;
|
[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
cfg.threads = 1;
|
|
|
|
|
fw_dec_ = codec_->CreateDecoder(cfg, 0);
|
|
|
|
|
inv_dec_ = codec_->CreateDecoder(cfg, 0);
|
2013-03-27 19:22:20 +01:00
|
|
|
inv_dec_->Control(VP9_INVERT_TILE_DECODE_ORDER, 1);
|
[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
virtual ~TileIndependenceTest() {
|
|
|
|
|
delete fw_dec_;
|
|
|
|
|
delete inv_dec_;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
virtual void SetUp() {
|
|
|
|
|
InitializeConfig();
|
2013-03-01 21:43:10 +01:00
|
|
|
SetMode(libvpx_test::kTwoPassGood);
|
[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
virtual void PreEncodeFrameHook(libvpx_test::VideoSource *video,
|
|
|
|
|
libvpx_test::Encoder *encoder) {
|
|
|
|
|
if (video->frame() == 1) {
|
|
|
|
|
encoder->Control(VP9E_SET_TILE_COLUMNS, n_tiles_);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void UpdateMD5(::libvpx_test::Decoder *dec, const vpx_codec_cx_pkt_t *pkt,
|
|
|
|
|
::libvpx_test::MD5 *md5) {
|
2013-05-06 20:45:03 +02:00
|
|
|
const vpx_codec_err_t res =
|
|
|
|
|
dec->DecodeFrame(reinterpret_cast<uint8_t*>(pkt->data.frame.buf),
|
|
|
|
|
pkt->data.frame.sz);
|
|
|
|
|
if (res != VPX_CODEC_OK) {
|
|
|
|
|
abort_ = true;
|
|
|
|
|
ASSERT_EQ(VPX_CODEC_OK, res);
|
|
|
|
|
}
|
[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
const vpx_image_t *img = dec->GetDxData().Next();
|
|
|
|
|
md5->Add(img);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
virtual void FramePktHook(const vpx_codec_cx_pkt_t *pkt) {
|
|
|
|
|
UpdateMD5(fw_dec_, pkt, &md5_fw_order_);
|
|
|
|
|
UpdateMD5(inv_dec_, pkt, &md5_inv_order_);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
private:
|
|
|
|
|
int n_tiles_;
|
|
|
|
|
protected:
|
|
|
|
|
::libvpx_test::MD5 md5_fw_order_, md5_inv_order_;
|
|
|
|
|
::libvpx_test::Decoder *fw_dec_, *inv_dec_;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
// run an encode with 2 or 4 tiles, and do the decode both in normal and
|
|
|
|
|
// inverted tile ordering. Ensure that the MD5 of the output in both cases
|
|
|
|
|
// is identical. If so, tiles are considered independent and the test passes.
|
|
|
|
|
TEST_P(TileIndependenceTest, MD5Match) {
|
|
|
|
|
const vpx_rational timebase = { 33333333, 1000000000 };
|
|
|
|
|
cfg_.g_timebase = timebase;
|
|
|
|
|
cfg_.rc_target_bitrate = 500;
|
|
|
|
|
cfg_.g_lag_in_frames = 25;
|
|
|
|
|
cfg_.rc_end_usage = VPX_VBR;
|
|
|
|
|
|
2013-02-07 00:30:21 +01:00
|
|
|
libvpx_test::I420VideoSource video("hantro_collage_w352h288.yuv", 704, 144,
|
[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
timebase.den, timebase.num, 0, 30);
|
|
|
|
|
ASSERT_NO_FATAL_FAILURE(RunLoop(&video));
|
|
|
|
|
|
|
|
|
|
const char *md5_fw_str = md5_fw_order_.Get();
|
|
|
|
|
const char *md5_inv_str = md5_inv_order_.Get();
|
|
|
|
|
|
|
|
|
|
// could use ASSERT_EQ(!memcmp(.., .., 16) here, but this gives nicer
|
|
|
|
|
// output if it fails. Not sure if it's helpful since it's really just
|
|
|
|
|
// a MD5...
|
|
|
|
|
ASSERT_STREQ(md5_fw_str, md5_inv_str);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
VP9_INSTANTIATE_TEST_CASE(TileIndependenceTest,
|
2013-02-07 00:30:21 +01:00
|
|
|
::testing::Range(0, 2, 1));
|
[WIP] Add column-based tiling.
This patch adds column-based tiling. The idea is to make each tile
independently decodable (after reading the common frame header) and
also independendly encodable (minus within-frame cost adjustments in
the RD loop) to speed-up hardware & software en/decoders if they used
multi-threading. Column-based tiling has the added advantage (over
other tiling methods) that it minimizes realtime use-case latency,
since all threads can start encoding data as soon as the first SB-row
worth of data is available to the encoder.
There is some test code that does random tile ordering in the decoder,
to confirm that each tile is indeed independently decodable from other
tiles in the same frame. At tile edges, all contexts assume default
values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode),
and motion vector search and ordering do not cross tiles in the same
frame.
t log
Tile independence is not maintained between frames ATM, i.e. tile 0 of
frame 1 is free to use motion vectors that point into any tile of frame
0. We support 1 (i.e. no tiling), 2 or 4 column-tiles.
The loopfilter crosses tile boundaries. I discussed this briefly with Aki
and he says that's OK. An in-loop loopfilter would need to do some sync
between tile threads, but that shouldn't be a big issue.
Resuls: with tiling disabled, we go up slightly because of improved edge
use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf,
~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5%
on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is
concentrated in the low-bitrate end of clips, and most of it is because
of the loss of edges at tile boundaries and the resulting loss of intra
predictors.
TODO:
- more tiles (perhaps allow row-based tiling also, and max. 8 tiles)?
- maybe optionally (for EC purposes), motion vectors themselves
should not cross tile edges, or we should emulate such borders as
if they were off-frame, to limit error propagation to within one
tile only. This doesn't have to be the default behaviour but could
be an optional bitstream flag.
Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 18:35:28 +01:00
|
|
|
|
|
|
|
|
} // namespace
|