在(原创)基于ZedBoard的Webcam设计(四):MJPG编码和AVI封装 中，我们在zed上完成了视频的采集和存储，但是MJPG只是对一帧一帧图片进行编码，编码效率只有10%。x264-snapshot提供了开源x264源代码，已经广泛用于嵌入式解决方案中。本小节将介绍zed上x264-snapshot源码的配置，并在zed上使用High版本、main版本对一个33M的测试视频进行编码，测试编码效率。
更多更新请关注我的博客：@超群天晴 http://www.cnblogs.com/surpassal/
相关阅读：
(原创)基于ZedBoard的Webcam设计(一)：USB摄像头(V4L2接口)的图片采集
(原创)基于ZedBoard的Webcam设计(二)：USB摄像头图片采集+QT显示
(原创)基于ZedBoard的Webcam设计(三)：视频的采集和动态显示
(原创)基于ZedBoard的Webcam设计(四)：MJPG编码和AVI封装
(原创)基于ZedBoard的Webcam设计(五)：x264编码在zedboard上的实现(软编码)
硬件平台： Digilent ZedBoard
开发环境： Windows XP 32 bit + Wmare 8.0 + Ubuntu 10.04 +Qt+ arm-linux-xilinx-gnueabi交叉编译环境
Zedboard linux: Digilent OOB Design
一、x264开源包获取
x264-snapshot提供了开源x264源代码，已经在X86和ARM架构下均已实现。linux下可以使用git获得最新的代码包
git clone git:// git.videolan.org/x264.git
也可以通过网页下载：
http://download.videolan.org/pub/videolan/x264/snapshots/
二、配置、安装x264-snapshot
下载到源码后，对其进行配置，以便在后续make时编译出zed上可以运行的可执行程序
1、解压源码包后，进入源码目录
./configure --host=arm-xilinx-linux-gnueabi --prefix=/opt/rcq --enable-shared --disable-asm
congfig之后，设定了host为arm-xilinx-linux-gnueabi，install路径是/opt/rcq
2、修改config.mak文件
针对zed的gcc，需要修改config.mak文件
1 SRCPATH=. 2 prefix=/opt/rcq 3 exec_prefix=${prefix} 4 bindir=${exec_prefix}/bin 5 libdir=${exec_prefix}/lib 6 includedir=${prefix}/include 7 ARCH=ARM 8 SYS=LINUX 9 CC=arm-xilinx-linux-gnueabi-gcc 10 CFLAGS=-Wshadow -O3 -fno-fast-math -Wall -I. -I$(SRCPATH) -std=gnu99 -fPIC -fomit-frame-pointer -fno-tree-vectorize 11 DEPMM=-MM -g0 12 DEPMT=-MT 13 LD=arm-xilinx-linux-gnueabi-gcc -o 14 LDFLAGS= -lm -lpthread 15 LIBX264=libx264.a 16 AR=arm-xilinx-linux-gnueabi- ar rc17 RANLIB=arm-xilinx-linux-gnueabi-ranlib 18 STRIP=arm-xilinx-linux-gnueabi-strip 19 AS=20 ASFLAGS= -DHAVE_ALIGNED_STACK=1 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 21 RC=22 RCFLAGS=23 EXE=24 HAVE_GETOPT_LONG=1 25 DEVNULL=/dev/null 26 PROF_GEN_CC=-fprofile-generate 27 PROF_GEN_LD=-fprofile-generate 28 PROF_USE_CC=-fprofile-use 29 PROF_USE_LD=-fprofile-use 30 default : cli 31 install: install-cli 32 SOSUFFIX=so 33 SONAME=libx264.so.129 34 SOFLAGS=-shared -Wl,-soname,$(SONAME) -Wl,-Bsymbolic 35 default : lib-shared 36 install: install-lib-shared 37 LDFLAGSCLI =38 CLI_LIBX264 = $(LIBX264)
其中9、13、16、17、18行为修改后的。
3、make
完运行库的安装和成执行文件的生成
三、在zed上测试x264
将可执行x264和测试rcq.yuv文件拷贝到zed上，使用默认运行命令
./x264 rcq.yuv --input-res 640x480 -o rcq.flv
其中--input-res 640x480设定输入yuv视频帧的大小，rcq.flv是输出文件，可以设定为.mkv/.mp4/H264等。从控制台可以看到
使用的profile是最高等级High level 3.0，但是的编码速率只有1.40fps ……情何以堪……
编码结束后：
生成的rcq.flv文件大小348k，相比较原来的rcq.yuv文件大小33M，压缩率在1%。
可以使用最快编码方式，命令和参数如下
./x264 rcq.yuv --input-res 640x480 -o rcq.flv --no-8x8dct --aq-mode0 --b-adapt0 --bframes0 --no-cabac --no-deblock --no-mbtree --me dia --no-mixed-refs --partitions none --rc-lookahead0 --ref 1 --scenecut0 --subme0 --trellis0
编码速率能有22fps ，这样还凑合。生成的rcq.flv文件大小787K，压缩率在5%。
四、x264命令和参数(fullhelp)
Syntax: x264 [options] -
o outfile infile
in which
case resolution
is
required), *
.y4m),
if
compiled with support (no). * formats
if
compiled with lavf support (no) or ffms support (no).
is
selected by filename:
264 ->
Raw bytestream ->
Matroska ->
Flash Video -> MP4
if
compiled with GPAC support (no)
8
(configured at compile time) -h, --
help List basic options --
longhelp List more options --
fullhelp List all options --crf
24 -o
pass with a bitrate of 1000kbps: --pass
1 --bitrate
1000 -o 2 --bitrate
1000 -o 0 -o 2 second-
buffer: --vbv-bufsize
2000 --bitrate
1000 -o string > Force the limits of an H.
264
profile -
baseline: --no-8x8dct --bframes
0 --no-
cabac --cqm flat --weightp
0
main: --no-8x8dct --
cqm flat -
high: -
high10:
for bit depth
8 -
10
. -
high422:
for bit depth
8 -
10
.
for
4 :
2 :
0 /
4 :
2 :
2
chroma subsampling. -
high444:
for bit depth
8 -
10
.
for
4 :
2 :
0 /
4 :
2 :
2 /
4 :
4 :
4
chroma subsampling. --preset <
string > Use a preset to
select
encoding settings [medium] -
ultrafast: --no-8x8dct --aq-mode
0 --b-adapt
0
0 --no-cabac --no-
deblock --no-mbtree --me dia --no-mixed-
refs --partitions none --rc-lookahead
0 --
ref
1
0 --subme
0 --trellis
0
0
superfast: --no-mbtree --me dia --no-mixed-
refs --partitions i8x8,i4x4 --rc-lookahead
0
ref
1 --subme
1 --trellis
0 --weightp
1
veryfast: --no-mixed-refs --rc-lookahead
10
ref
1 --subme
2 --trellis
0 --weightp
1
faster: --no-mixed-refs --rc-lookahead
20
ref
2 --subme
4 --weightp
1
fast: --rc-lookahead
30 --
ref
2 --subme
6
1
medium: -
slow: --b-adapt
2 --direct auto --
me umh --rc-lookahead
50 --
ref
5 --subme
8
slower: --b-adapt
2 --direct auto --
me umh --partitions all --rc-lookahead
60
ref
8 --subme
9 --trellis
2
veryslow: --b-adapt
2 --bframes
8 --
direct auto --me umh --merange
24 --
partitions all --
ref
16 --subme
10 --trellis
2
60
placebo: --bframes
16 --b-adapt
2 --
direct auto --slow-firstpass --no-fast-
pskip --me tesa --merange
24 --
partitions all --rc-lookahead
60 --
ref
16 --subme
11
2
string > Tune the settings
for
a particular type of source -
film (psy tuning): --deblock -
1 :-
1 --psy-rd :
0.15
animation (psy tuning): --bframes {+
2 } --deblock
1 :
1
0.4 : --aq-strength
0.6
ref {Double
if >
1
else
1
} -
grain (psy tuning): --aq-strength
0.5 --no-dct-
decimate --deadzone-inter
6 --deadzone-intra
6
2 :-
2 --ipratio
1.1
1.1 --psy-rd :
0.25
0.8
stillimage (psy tuning): --aq-strength
1.2 --deblock -
3 :-
3
2.0 :
0.7
psnr (psy tuning): --aq-mode
0 --no-
psy -
ssim (psy tuning): --aq-mode
2 --no-
psy -
fastdecode: --no-cabac --no-deblock --no-
weightb --weightp
0
zerolatency: --bframes
0 --force-cfr --no-
mbtree --sync-lookahead
0 --sliced-
threads --rc-lookahead
0
'
t force these faster settings with --pass 1:
partitions none --
ref
1 --subme {
2
if >
2
else
unchanged} --trellis
0 --fast-
pskip -
type options: -I, --keyint "
infinite
" > Maximum GOP size [
250
] -i, --min-keyint
Minimum GOP size [auto] --no-scenecut Disable adaptive I-
frame decision --scenecut How aggressively to insert extra I-frames [
40
] --intra-
refresh Use Periodic Intra Refresh instead of IDR frames -b, --bframes Number of B-frames between I and P [
3
] --b-adapt Adaptive B-frame decision method [
1
] -
0
: Disabled -
1
: Fast -
2 : Optimal (slow with high --
bframes) --b-bias Influences how often B-frames are used [
0
] --b-pyramid <
string > Keep some B-frames
as
references [normal] -
none: Disabled -
strict: Strictly hierarchical pyramid - normal: Non-strict (not Blu-
ray compatible) --open-
gop Use recovery points to close GOPs -
frames --no-
cabac Disable CABAC -r, --
ref Number of reference frames [
3
] --no-
deblock Disable loop filter -f, --deblock Loop filter parameters [
0 :
0
] --slices
Number of slices per frame; forces rectangular
is
overridden by other slicing options --slice-max-size Limit the size of each slice
in
bytes --slice-max-mbs Limit the size of each slice
in
macroblocks --
tff Enable interlaced mode (top field first) --
bff Enable interlaced mode (bottom field first) --constrained-
intra Enable constrained intra prediction. --pulldown <
string >
Use soft pulldown to change frame rate - none,
22 ,
32 ,
64 ,
double
, triple, euro (requires cfr input) --fake-interlaced Flag stream
as
interlaced but encode progressive. -
Ray
in
interlaced mode. --frame-packing
For stereoscopic videos define frame arrangement -
0 : checkerboard - pixels are alternatively
from
L and R -
1 : column alternation -
L and R are interlaced by column -
2 : row alternation -
L and R are interlaced by row -
3 : side by side - L
is
on the left, R on the right -
4 : top bottom - L
is
on top, R on bottom -
5 : frame alternation -
one view per frame -q, --qp Force constant QP (
0 -
69 ,
0 =
lossless) -B, --bitrate Set bitrate (kbit/
s) --crf <
float > Quality-based VBR (
0 -
51 ) [
23.0
] --rc-lookahead Number of frames
for frametype lookahead [
40
] --vbv-maxrate Max local bitrate (kbit/s) [
0
] --vbv-bufsize Set size of the VBV buffer (kbit) [
0
] --vbv-init <
float > Initial VBV buffer occupancy [
0.9
] --crf-max <
float > With CRF+VBV, limit RF to
this
value !
Set min QP [
0
] --qpmax Set max QP [
69
] --qpstep Set max QP step [
4
] --ratetol <
float > Tolerance of ABR ratecontrol and VBV [
1.0
] --ipratio <
float > QP factor between I and P [
1.40
] --pbratio <
float > QP factor between P and B [
1.30
] --chroma-qp-offset QP difference between chroma and luma [
0
] --aq-mode AQ method [
1
] -
0
: Disabled -
1
: Variance AQ (complexity mask) -
2 : Auto-
variance AQ (experimental) --aq-strength <
float > Reduces blocking and blurring
in
flat and
1.0
] -p, --pass
Enable multipass ratecontrol -
1
: First pass, creates stats file -
2
: Last pass, does not overwrite stats file -
3
: Nth pass, overwrites stats file --stats <
string > Filename
for
2 pass stats [
"
x264_2pass.log
"
] --no-mbtree Disable mb-
tree ratecontrol. --qcomp <
float > QP curve compression [
0.60
] --cplxblur <
float > Reduce fluctuations
in QP (before curve compression) [
20.0
] --qblur <
float > Reduce fluctuations
in QP (after curve compression) [
0.5
] --zones //
... Tweak the bitrate of regions of the video
is
of the form ,,
where
is
either =
(force QP) =<
float >
(bitrate multiplier) --qpfile <
string > Force frametypes and QPs
for
some or all frames
is
optional (none lets x264 choose). Frametypes: I,i,K,P,B,b. = depending on open-
gop setting /
qpmax. -A, --partitions <
string > Partitions to consider [
"
p8x8,b8x8,i8x8,i4x4
"
] -
p8x8, p4x4, b8x8, i8x8, i4x4 -
none, all --
8x8dct.) --direct <
string > Direct MV prediction mode [
"
spatial
"
] -
none, spatial, temporal, auto --no-weightb Disable weighted prediction
for B-
frames --weightp Weighted prediction
for P-frames [
2
] -
0
: Disabled -
1
: Weighted refs -
2 : Weighted refs +
Duplicates --me <
string > Integer pixel motion estimation method [
"
hex
"
] - dia: diamond search, radius
1
(fast) - hex: hexagonal search, radius
2
hexagon search -
esa: exhaustive search -
tesa: hadamard exhaustive search (slow) --merange Maximum motion vector search range [
16
] --mvrange Maximum motion vector length [-
1
(auto)] --mvrange-thread <
int > Minimum buffer between threads [-
1
(auto)] -m, --subme Subpixel motion estimation and mode decision [
7
] -
0
: fullpel only (not recommended) -
1
: SAD mode decision, one qpel iteration -
2
: SATD mode decision -
3 -
5
: Progressively more qpel -
6 : RD mode decision
for I/P-
frames -
7 : RD mode decision
for
all frames -
8 : RD refinement
for I/P-
frames -
9 : RD refinement
for
all frames -
10 : QP-RD - requires trellis=
2 , aq-mode>
0
11
: Full RD: disable all early terminations --psy-rd <
float :
float > Strength of psychovisual optimization [
"
1.0:0.0
"
]
1 : RD (requires subme>=
6
)
2
: Trellis (requires trellis, experimental) --no-
psy Disable all visual optimizations that worsen --no-mixed-refs Don
'
t decide references on a per partition basis
in
motion estimation --no-
8x8dct Disable adaptive spatial transform size -t, --trellis Trellis RD quantization. [
1
] -
0
: disabled -
1
: enabled only on the final encode of a MB -
2
: enabled on all mode decisions --no-fast-pskip Disables early SKIP detection on P-
frames --no-dct-decimate Disables coefficient thresholding on P-
frames --nr Noise reduction [
0
] --deadzone-inter <
int > Set the size of the inter luma quantization deadzone [
21
] --deadzone-intra <
int > Set the size of the intra luma quantization deadzone [
11
]
in the range
0 -
32
. --cqm <
string > Preset quant matrices [
"
flat
"
] -
jvt, flat --cqmfile <
string > Read custom quant matrices
from a JM-
compatible file --cqm*
options. --cqm4
Set all 4x4 quant matrices -separated list of
16
integers. --cqm8
Set all 8x8 quant matrices -separated list of
64
integers. --cqm4i, --cqm4p, --cqm8i, --cqm8p
for
details. Use at your own risk. --overscan <
string > Specify crop overscan setting [
"
undef
"
] -
undef, show, crop --videoformat <
string > Specify video format [
"
undef
"
] -
component, pal, ntsc, secam, mac, undef --range <
string > Specify color range [
"
auto
"
] -
auto, tv, pc --colorprim <
string > Specify color primaries [
"
undef
"
] -
undef, bt709, bt470m, bt470bg --transfer <
string > Specify transfer characteristics [
"
undef
"
] -
undef, bt709, bt470m, bt470bg, linear, --colormatrix <
string > Specify color matrix setting [
"
???
"
] -
undef, bt709, fcc, bt470bg --chromaloc Specify chroma sample location (
0 to
5 ) [
0
] --nal-hrd <
string > Signal HRD information (requires vbv-
bufsize) - none, vbr, cbr (cbr not allowed
in
.mp4) --pic-
struct Force pic_struct
in
Picture Timing SEI --crop-rect <
string > Add
'
left,top,right,bottom
' to the bitstream-
level /
Output: -o, --output <
string >
Specify output file --muxer <
string > Specify output container format [
"
auto
"
] -
auto, raw, mkv, flv --demuxer <
string > Specify input container format [
"
auto
"
] -
auto, raw, y4m --input-fmt <
string >
Specify input file format (requires lavf support) --input-csp <
string > Specify input colorspace format
for
raw input - valid csps
for `raw
'
demuxer:
i420, yv12, nv12, i422, yv16, nv16, i444, yv24, bgr, bgra, rgb --output-csp <
string > Specify output colorspace [
"
i420
"
] -
i420, i422, i444, rgb --input-depth Specify input bit depth
for
raw input --input-range <
string > Specify input color range [
"
auto
"
] -
auto, tv, pc --input-res
Specify input resolution (width x height) --index <
string > Filename
for
input index file --
sar width:height Specify Sample Aspect Ratio --fps <
float |rational>
Specify framerate --seek
First frame to encode --frames
Maximum number of frames to encode --level <
string > Specify level (
as
defined by Annex A) --bluray-compat Enable compatibility hacks
for Blu-
ray support -v, --verbose Print stats
for
each frame --no-progress Don
'
t show the progress indicator while encoding
quiet Quiet Mode --log-level <
string > Specify the maximum level of logging [
"
info
"
] -
none, error, warning, info, debug --
psnr Enable PSNR computation --
ssim Enable SSIM computation --threads
Force a specific number of threads --lookahead-threads
Force a specific number of lookahead threads --sliced-threads Low-latency but lower-
efficiency threading --thread-input Run Avisynth
in
its own thread --sync-lookahead Number of buffer frames
for
threaded lookahead --non-
deterministic Slightly improve quality of SMP, at the cost of repeatability --cpu-
independent Ensure exact reproducibility across different cpus,
as opposed to letting them
select
different algorithms --asm
Override CPU detection --no-
asm Disable all CPU optimizations --
visualize Show MB types overlayed on the encoded video --dump-yuv <
string >
Save reconstructed frames --sps-id Set SPS and PPS id numbers [
0
] --
aud Use access unit delimiters --force-
cfr Force constant framerate timestamp generation --tcfile-
in <
string >
Force timestamp generation with timecode file --tcfile-
out <
string > Output timecode v2 file
from
input timestamps --timebase <
int /
int >
Specify timebase numerator and denominator Specify timebase numerator
for
input timecode file
for
other input --dts-
compress Eliminate initial delay with container DTS hack --vf, --video-filter //
... Apply video filtering to the input file
in :=
format.
from
the edges of the frame
in
the pattern
select
a frame
//
avisynth.org/mediawiki/Select#SelectEvery
五、关于YUV420的格式说明
YUV格式有两大类：planar和packed。 对于planar的YUV格式，先连续存储所有像素点的Y，紧接着存储所有像素点的U，随后是所有像素点的V。 对于packed的YUV格式，每个像素点的Y,U,V是连续交 *存储的。 一般的录像程序是先从摄像头得到yuv420 planar的数据，然后编码成h264格式 帧，最后存储成3gp /mpeg等格式的视频文件。 以720×488大小图象YUV420 planar为例，其存储格式是： 共大小为( 720 ×480 ×3 >>1 )字节，分为三个部分:Y,U和V Y分量： ( 720 ×480 )个字节 U(Cb)分量：( 720 ×480 >>2 )个字节 V(Cr)分量：( 720 ×480 >>2 )个字节 三个部分内部均是行优先存储，三个部分之间是Y,U,V 顺序存储。 即YUV数据的0－－ 720 ×480字节是Y分量值， 720 ×480 －－720 ×480 ×5 /4字节是U分量 720 ×480 ×5 /4 －－720 ×480 ×3 /2字节是V分量。 一副w *h的yuv420图像数据的大小为w*h*1.5 。 4 ：2 ：2 和4：2 ：0 转换： 最简单的方式： YUV4: 2 :2 ---> YUV4:2 :0 Y不变，将U和V信号值在行(垂直方向)在进行一次隔行抽样。 YUV4: 2 :0 ---> YUV4:2 :2 Y不变，将U和V信号值的每一行分别拷贝一份形成连续两行数据。
六、测试视频
=====================================================
0、x264-snapshots:http://download.videolan.org/pub/videolan/x264/snapshots/
1、zed可执行文件：x264.rar
2、测试yuv文件：rcq.part1.rar rcq.part2.rar
3、其他通用测试yuv序列：http://www.cipr.rpi.edu/resource/sequences/sif.html
