virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-12 19:18:44 +02:00
Code Issues Releases Activity
99,027 Commits 37 Branches 400 Tags 463 MiB
5fcd83a244
Commit Graph

8566 Commits

Author SHA1 Message Date
Paul B Mahol
4dbb75c437 avfilter/vf_xfade: check that fps between inputs are valid 2020-08-11 10:57:04 +02:00
Ting Fu
91efc41a69 dnn/native: add native support for avg_pool 
Not support pooling strides in channel dimension yet.

Signed-off-by: Ting Fu <ting.fu@intel.com>
Reviewed-by: Guo, Yejun <yejun.guo@intel.com>
 2020-08-10 16:37:39 +08:00
Mingyu Yin
fab00b0ae0 dnn_backend_native_layer_mathunary: add floor support 
It can be tested with the model generated with below python script:

import tensorflow as tf
import os
import numpy as np
import imageio
from tensorflow.python.framework import graph_util
name = 'floor'

pb_file_path = os.getcwd()
if not os.path.exists(pb_file_path+'/{}_savemodel/'.format(name)):
    os.mkdir(pb_file_path+'/{}_savemodel/'.format(name))

with tf.Session(graph=tf.Graph()) as sess:
    in_img = imageio.imread('detection.jpg')
    in_img = in_img.astype(np.float32)
    in_data = in_img[np.newaxis, :]
    input_x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
    y_ = tf.math.floor(input_x*255)/255
    y = tf.identity(y_, name='dnn_out')
    sess.run(tf.global_variables_initializer())
    constant_graph = graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])

    with tf.gfile.FastGFile(pb_file_path+'/{}_savemodel/model.pb'.format(name), mode='wb') as f:
        f.write(constant_graph.SerializeToString())

    print("model.pb generated, please in ffmpeg path use\n \n \
    python tools/python/convert.py {}_savemodel/model.pb --outdir={}_savemodel/ \n \nto generate model.model\n".format(name,name))

    output = sess.run(y, feed_dict={ input_x: in_data})
    imageio.imsave("out.jpg", np.squeeze(output))

    print("To verify, please ffmpeg path use\n \n \
    ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.pb:input=dnn_in:output=dnn_out:dnn_backend=tensorflow -f framemd5 {}_savemodel/tensorflow_out.md5\n  \
    or\n \
    ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.pb:input=dnn_in:output=dnn_out:dnn_backend=tensorflow {}_savemodel/out_tensorflow.jpg\n \nto generate output result of tensorflow model\n".format(name, name, name, name))

    print("To verify, please ffmpeg path use\n \n \
    ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.model:input=dnn_in:output=dnn_out:dnn_backend=native -f framemd5 {}_savemodel/native_out.md5\n  \
    or \n \
    ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model={}_savemodel/model.model:input=dnn_in:output=dnn_out:dnn_backend=native {}_savemodel/out_native.jpg\n \nto generate output result of native model\n".format(name, name, name, name))

Signed-off-by: Mingyu Yin <mingyu.yin@intel.com>
 2020-08-07 10:34:22 +08:00
Mingyu Yin
9fbdd5454b dnn_backend_native_layer_mathunary: add ceil support 
It can be tested with the model generated with below python script:

import tensorflow as tf
import os
import numpy as np
import imageio
from tensorflow.python.framework import graph_util
name = 'ceil'

pb_file_path = os.getcwd()
if not os.path.exists(pb_file_path+'/{}_savemodel/'.format(name)):
    os.mkdir(pb_file_path+'/{}_savemodel/'.format(name))

with tf.Session(graph=tf.Graph()) as sess:
    in_img = imageio.imread('detection.jpg')
    in_img = in_img.astype(np.float32)
    in_data = in_img[np.newaxis, :]
    input_x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
    y = tf.math.ceil( input_x, name='dnn_out')
    sess.run(tf.global_variables_initializer())
    constant_graph = graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])

    with tf.gfile.FastGFile(pb_file_path+'/{}_savemodel/model.pb'.format(name), mode='wb') as f:
        f.write(constant_graph.SerializeToString())

    print("model.pb generated, please in ffmpeg path use\n \n \
    python tools/python/convert.py ceil_savemodel/model.pb --outdir=ceil_savemodel/ \n \n \
    to generate model.model\n")

    output = sess.run(y, feed_dict={ input_x: in_data})
    imageio.imsave("out.jpg", np.squeeze(output))

    print("To verify, please ffmpeg path use\n \n \
    ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model=ceil_savemodel/model.pb:input=dnn_in:output=dnn_out:dnn_backend=tensorflow -f framemd5 ceil_savemodel/tensorflow_out.md5\n \n \
    to generate output result of tensorflow model\n")

    print("To verify, please ffmpeg path use\n \n \
    ./ffmpeg -i detection.jpg -vf format=rgb24,dnn_processing=model=ceil_savemodel/model.model:input=dnn_in:output=dnn_out:dnn_backend=native -f framemd5 ceil_savemodel/native_out.md5\n \n \
    to generate output result of native model\n")

Signed-off-by: Mingyu Yin <mingyu.yin@intel.com>
Reviewed-by: Guo, Yejun <yejun.guo@intel.com>
 2020-08-04 19:56:54 +08:00
Andreas Rheinhardt
8129c32e48 avcodec, avfilter, avformat: Remove redundant avpriv_align_put_bits 
flush_put_bits() already fills the bitstream with zeroes, so it is
unnecessary to align the bitstream before.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2020-08-01 18:56:32 +02:00
Paul B Mahol
17f6bd6e58 avfilter/vf_xfade: add hblur transition 2020-07-21 23:29:55 +02:00
Paul B Mahol
c4c989c7ca avfilter/vf_yaepblur: fix naming of filter class 2020-07-20 16:12:56 +02:00
Paul B Mahol
fa8345cf05 avfilter/vf_bilateral: simplify code a little 
Make alpha_ calculation faster.
 2020-07-18 10:22:12 +02:00
Paul B Mahol
3a37aa597f avfilter/avf_showwavespic: add filter mode 2020-07-17 14:26:48 +02:00
Paul B Mahol
d363afb30e avfilter/vf_tinterlace: fix mergex2, first frame is always considered odd 2020-07-17 13:53:55 +02:00
Paul B Mahol
24fea4d09b avfilter/vf_tinterlace: use frame counter from lavfi 
Remove internal counter.
 2020-07-17 13:53:55 +02:00
leozhang
fe591393cd avfilter/vf_bilateral: remove useless memcpy 
Signed-off-by: leozhang <leozhang@qiyi.com>
 2020-07-17 13:53:22 +02:00
Paul B Mahol
241cdded0f avfilter/vf_bilateral: stop using sigmaS as percent of width/height 2020-07-17 13:53:22 +02:00
James Almer
320694ff84 x86/vf_blend: fix warnings about trailing empty parameters 
Finishes fixing ticket #8771

Signed-off-by: James Almer <jamrial@gmail.com>
 2020-07-12 11:30:23 -03:00
Jun Zhao
04037e2966 lavfi/setpts: fix setpts/asetpts option dump error 
fix the command ffmpeg -h filter=setpts/asetpts both dump the expr
option with "FVA" flags.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Jun Zhao <barryjzhao@tencent.com>
 2020-07-12 08:11:42 +08:00
Ben Clayton
4dab04622a libavfilter/glslang: Remove unused header 
The <glslang/Include/revision.h> include was not used, and revision.h has
been removed from glslang master.
See: https://github.com/KhronosGroup/glslang/pull/2277
 2020-07-11 13:01:33 +01:00
Paul B Mahol
6f84e92172 avfilter/vf_chromanr: move thres calculation to filter_frame() 2020-07-10 23:09:19 +02:00
Limin Wang
f9277cd796 avfilter/vf_showinfo: add dump_s12m_timecode() helper function 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-07-08 23:14:04 +08:00
Limin Wang
3ede8acba6 avfilter/vf_showinfo: check sd->size before reference the sd->data 
Or it'll cause null pointer dereference if size < sizeof(uint32_t), also
in case tc[0] > 3, the code will report error directly.

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-07-08 23:12:48 +08:00
Paul B Mahol
6cdddb773f avfilter: add chromanr video filter 2020-07-08 15:23:43 +02:00
Valery Kot
855d51bf48 avfilter/vf_edgedetect: properly implement double_threshold() 
Important part of this algorithm is the double threshold step: pixels
above "high" threshold being kept, pixels below "low" threshold dropped,
pixels in between (weak edges) are kept if they are neighboring "high"
pixels.

The weak edge check uses a neighboring context and should not be applied
on the plane's border. The condition was incorrect and has been fixed in
the commit.

Signed-off-by: Andriy Gelman <andriy.gelman@gmail.com>
Reviewed-by: Andriy Gelman <andriy.gelman@gmail.com>
 2020-07-06 23:20:53 -04:00
Reimar Döffinger
584f396132 dnn_backend_native: Add overflow check for length calculation. 
We should not silently allocate an incorrect sized buffer.
Fixes trac issue #8718.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Reviewed-by: Guo, Yejun <yejun.guo@intel.com>
 2020-07-06 20:22:30 +08:00
Ting Fu
c0cdeea0ee dnn_backend_native_layer_mathunary: add atanh support 
It can be tested with the model generated with below python script:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')

please uncomment the part you want to test

x_sinh_1 = tf.sinh(x)
x_out = tf.divide(x_sinh_1, 1.176) # sinh(1.0)

x_cosh_1 = tf.cosh(x)
x_out = tf.divide(x_cosh_1, 1.55) # cosh(1.0)

x_tanh_1 = tf.tanh(x)
x__out = tf.divide(x_tanh_1, 0.77) # tanh(1.0)

x_asinh_1 = tf.asinh(x)
x_out = tf.divide(x_asinh_1, 0.89) # asinh(1.0/1.1)

x_acosh_1 = tf.add(x, 1.1)
x_acosh_2 = tf.acosh(x_acosh_1) # accept (1, inf)
x_out = tf.divide(x_acosh_2, 1.4) # acosh(2.1)

x_atanh_1 = tf.divide(x, 1.1)
x_atanh_2 = tf.atanh(x_atanh_1) # accept (-1, 1)
x_out = tf.divide(x_atanh_2, 1.55) # atanhh(1.0/1.1)

y = tf.identity(x_out, name='dnn_out') #please only preserve the x_out you want to test

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
 2020-07-06 12:45:14 +08:00
Ting Fu
cd2e3a864d dnn_backend_native_layer_mathunary: add acosh support 
Signed-off-by: Ting Fu <ting.fu@intel.com>
 2020-07-06 12:45:14 +08:00
Ting Fu
9d14b38d9d dnn_backend_native_layer_mathunary: add asinh support 
Signed-off-by: Ting Fu <ting.fu@intel.com>
 2020-07-06 12:45:14 +08:00
Ting Fu
ea71e731f4 dnn_backend_native_layer_mathunary: add tanh support 
Signed-off-by: Ting Fu <ting.fu@intel.com>
 2020-07-06 12:45:14 +08:00
Ting Fu
62fc7e3035 dnn_backend_native_layer_mathunary: add cosh support 
Signed-off-by: Ting Fu <ting.fu@intel.com>
 2020-07-06 12:45:14 +08:00
Ting Fu
91b4037101 dnn_backend_native_layer_mathunary: add sinh support 
Signed-off-by: Ting Fu <ting.fu@intel.com>
 2020-07-06 12:45:14 +08:00
Limin Wang
49054fe94c FATE: fix colorbalance fate test failed on x86_32 
floating point precision will cause rgb*max generate different value on
x86_32 and x86_64. have pass fate test on x86_32 and x86_64 by using
lrintf to get the nearest integral value for rgb * max before av_clip.

Reviewed-by:   Paul B Mahol <onemda@gmail.com>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-07-02 21:12:37 +08:00
Guo, Yejun
9bcf2aa477 vf_dnn_processing.c: add dnn backend openvino 
We can try with the srcnn model from sr filter.
1) get srcnn.pb model file, see filter sr
2) convert srcnn.pb into openvino model with command:
python mo_tf.py --input_model srcnn.pb --data_type=FP32 --input_shape [1,960,1440,1] --keep_shape_ops

See the script at https://github.com/openvinotoolkit/openvino/tree/master/model-optimizer
We'll see srcnn.xml and srcnn.bin at current path, copy them to the
directory where ffmpeg is.

I have also uploaded the model files at https://github.com/guoyejun/dnn_processing/tree/master/models

3) run with openvino backend:
ffmpeg -i input.jpg -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=openvino:model=srcnn.xml:input=x:output=srcnn/Maximum -y srcnn.ov.jpg
(The input.jpg resolution is 720*480)

Also copy the logs on my skylake machine (4 cpus) locally with openvino backend
and tensorflow backend. just for your information.

$ time ./ffmpeg -i 480p.mp4 -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=tensorflow:model=srcnn.pb:input=x:output=y -y srcnn.tf.mp4
…
frame=  343 fps=2.1 q=31.0 Lsize=    2172kB time=00:00:11.76 bitrate=1511.9kbits/s speed=0.0706x
video:1973kB audio:187kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.517637%
[aac @ 0x2f5db80] Qavg: 454.353
real    2m46.781s
user    9m48.590s
sys     0m55.290s

$ time ./ffmpeg -i 480p.mp4 -vf format=yuv420p,scale=w=iw*2:h=ih*2,dnn_processing=dnn_backend=openvino:model=srcnn.xml:input=x:output=srcnn/Maximum -y srcnn.ov.mp4
…
frame=  343 fps=4.0 q=31.0 Lsize=    2172kB time=00:00:11.76 bitrate=1511.9kbits/s speed=0.137x
video:1973kB audio:187kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.517640%
[aac @ 0x31a9040] Qavg: 454.353
real    1m25.882s
user    5m27.004s
sys     0m0.640s

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>
 2020-07-02 09:56:55 +08:00
Guo, Yejun
ff37ebaf30 dnn: add openvino as one of dnn backend 
OpenVINO is a Deep Learning Deployment Toolkit at
https://github.com/openvinotoolkit/openvino, it supports CPU, GPU
and heterogeneous plugins to accelerate deep learning inferencing.

Please refer to https://github.com/openvinotoolkit/openvino/blob/master/build-instruction.md
to build openvino (c library is built at the same time). Please add
option -DENABLE_MKL_DNN=ON for cmake to enable CPU path. The header
files and libraries are installed to /usr/local/deployment_tools/inference_engine/
with default options on my system.

To build FFmpeg with openvion, take my system as an example, run with:
$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/deployment_tools/inference_engine/lib/intel64/:/usr/local/deployment_tools/inference_engine/external/tbb/lib/
$ ../ffmpeg/configure --enable-libopenvino --extra-cflags=-I/usr/local/deployment_tools/inference_engine/include/ --extra-ldflags=-L/usr/local/deployment_tools/inference_engine/lib/intel64
$ make

Here are the features provided by OpenVINO inference engine:
- support more DNN model formats
It supports TensorFlow, Caffe, ONNX, MXNet and Kaldi by converting them
into OpenVINO format with a python script. And torth model
can be first converted into ONNX and then to OpenVINO format.

see the script at https://github.com/openvinotoolkit/openvino/tree/master/model-optimizer/mo.py
which also does some optimization at model level.

- optimize at inference stage
It optimizes for X86 CPUs with SSE, AVX etc.

It also optimizes based on OpenCL for Intel GPUs.
(only Intel GPU supported becuase Intel OpenCL extension is used for optimization)

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
Signed-off-by: Pedro Arthur <bygrandao@gmail.com>
 2020-07-02 09:36:34 +08:00
Paul B Mahol
cca982ee01 avfilter/vf_colorbalance: remove wrong addition 2020-06-29 14:52:37 +02:00
Limin Wang
12c42c709e avfilter/vf_showinfo: add a \n for end of ERROR and WARNNING log 
Note for info level, one extra \n will be print after the log.

Reviewed-by:   Paul B Mahol <onemda@gmail.com>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-06-28 09:00:28 +08:00
exwm
32d6fe23b6 avfilter/zoompan: add in_time variable 
Currently, the zoompan filter exposes a 'time' variable (missing from docs) for use in
the 'zoom', 'x', and 'y' expressions. This variable is perhaps better named
'out_time' as it represents the timestamp in seconds of each output frame
produced by zoompan. This patch adds aliases 'out_time' and 'ot' for 'time'.

This patch also adds an 'in_time' (alias 'it') variable that provides access
to the timestamp in seconds of each input frame to the zoompan filter.
This helps to design zoompan filters that depend on the input video timestamps.
For example, it makes it easy to zoom in instantly for only some portion of a video.
Both the 'out_time' and 'in_time' variables have been added in the documentation
for zoompan.

Example usage of 'in_time' in the zoompan filter to zoom in 2x for the
first second of the input video and 1x for the rest:
    zoompan=z='if(between(in_time,0,1),2,1):d=1'

V2: Fix zoompan filter documentation stating that the time variable
would be NAN if the input timestamp is unknown.

V3: Add 'it' alias for 'in_time. Add 'out_time' and 'ot' aliases for 'time'.
Minor corrections to zoompan docs.

Signed-off-by: exwm <thighsman@protonmail.com>
 2020-06-25 10:27:07 +02:00
Ting Fu
13f5613e68 dnn_backend_native_layer_mathunary: add atan support 
It can be tested with the model generated with below python script:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.atan(x)
x2 = tf.divide(x1, 3.1416/4) # pi/4
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-25 08:41:50 +08:00
Ting Fu
461485feac dnn_backend_native_layer_mathunary: add acos support 
It can be tested with the model generated with below python script:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.acos(x)
x2 = tf.divide(x1, 3.1416/2) # pi/2
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-25 08:41:50 +08:00
Ting Fu
486c0c419d dnn_backend_native_layer_mathunary: add asin support 
It can be tested with the model generated with below python script:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.asin(x)
x2 = tf.divide(x1, 3.1416/2) # pi/2
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-25 08:41:50 +08:00
Paul B Mahol
ce297b44d3 avfilter/vf_v360: do not ignore return value of allocate_plane() 2020-06-23 21:55:40 +02:00
Paul B Mahol
00a5df71ad avfilter/vf_v360: add orthographic projection support 2020-06-23 16:00:02 +02:00
Paul B Mahol
44ce333f03 avfilters/vf_v360: add equisolid projection support 2020-06-22 14:41:36 +02:00
Andreas Rheinhardt
3f2be5372e avfilter/vf_showpalette: Don't pretend disp_palette can fail 
It can't fail, yet it returns an int and other code checks whether it
failed; yet if it did fail, an AVFrame would leak. One could of course
add an av_frame_free for this (that compilers could optimize away), yet
it is easier to simply stop pretending that disp_palette could fail.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2020-06-22 13:52:01 +02:00
Paul B Mahol
fdac3c80ac avfilter/af_ladspa: check return value of getenv() 2020-06-21 21:35:40 +02:00
Paul B Mahol
683a1599d4 avfilter/af_ladspa: add latency compensation 2020-06-21 21:35:40 +02:00
Paul B Mahol
842bc312ad avfilter/af_ladspa: check another directory for plugins 2020-06-21 14:48:27 +02:00
Limin Wang
548ef7a12b avfilter: add D2TS, TS2D, TS2T as a common macro in internal.h 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-06-19 23:12:49 +08:00
Limin Wang
dacae40a4b avfilter/vf_overlay: add yuv420p10 and yuv422p10 10bit format support 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-06-19 07:14:46 +08:00
Limin Wang
4d787c16e8 avfilter/vf_overlay: support for 8bit and 10bit overlay with macro-based function 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-06-19 07:14:46 +08:00
Guo Yejun
0b3bd001ac dnn_backend_native: check operand index 
it fixed the issue in https://trac.ffmpeg.org/ticket/8716
 2020-06-17 13:42:52 +08:00
Guo Yejun
fc932195ab dnn_backend_native.c: refine code for fail case 2020-06-17 13:42:52 +08:00
Limin Wang
567d571b20 avfilter/vf_showinfo: display H.26[45] user data unregistered sei message 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-06-15 07:19:55 +08:00
Paul B Mahol
c0e7164ba6 avfilter/vf_vaguedenoiser: fix small typo in option explanation 2020-06-13 00:41:16 +02:00
Paul B Mahol
e65d76fb94 avfilter/af_rubberband: adjust nb_samples after every command 2020-06-13 00:21:07 +02:00
Ting Fu
22d0860c13 dnn_backend_native_layer_mathunary: add tan support 
It can be tested with the model generated with below python scripy

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.multiply(x, 0.78)
x2 = tf.tan(x1)
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-11 11:10:51 +08:00
Ting Fu
88fb494f42 dnn_backend_native_layer_mathunary: add cos support 
It can be tested with the model generated with below python scripy

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.multiply(x, 1.5)
x2 = tf.cos(x1)
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-11 11:10:51 +08:00
Ting Fu
0b6d3f0d83 dnn_backend_native_layer_mathunary: add sin support 
It can be tested with the model file generated with below python scripy:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.multiply(x, 3.14)
x2 = tf.sin(x1)
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-11 11:10:51 +08:00
Anton Khirnov
c7d8d8d8d9 vf_spp: switch to child_class_iterate() 2020-06-10 12:36:44 +02:00
Anton Khirnov
6bfac4ee6f vf_scale: switch to child_class_iterate() 2020-06-10 12:36:44 +02:00
Anton Khirnov
344149cf01 framesync: switch to child_class_iterate() 2020-06-10 12:36:44 +02:00
Anton Khirnov
aba98de6b8 avfilter: switch to child_class_iterate() 2020-06-10 12:36:44 +02:00
Anton Khirnov
342230a537 af_resample: switch to child_class_iterate() 2020-06-10 12:36:44 +02:00
Anton Khirnov
3dd324427a af_aresample: switch to child_class_iterate() 2020-06-10 12:36:44 +02:00
Anton Khirnov
0d6b4351c6 Remove unnecessary use of avcodec_close(). 
Replace it with avcodec_free_context() or drop it completely as
appropriate.
 2020-06-10 11:31:16 +02:00
Michael Niedermayer
c5079bf3bc Bump minor versions after branching 4.3 
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-06-08 22:49:04 +02:00
Michael Niedermayer
0a8a96c251 Bump minor versions to separate 4.3 from master 
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-06-08 22:49:04 +02:00
Paul B Mahol
bd6336b970 avfilter/vf_vaguedenoiser: add new type of threshold 2020-06-07 15:20:25 +02:00
Paul B Mahol
6c57b0d63a avfilter/vf_vaguedenoiser: remove excessive code from soft thresholding 2020-06-07 15:20:11 +02:00
Paul B Mahol
7826fbfeaa avfilter/avf_showspectrum: properly handle EOF case 2020-06-06 19:49:14 +02:00
Paul B Mahol
1c32d7dfcf avfilter/asrc_anoisesrc: switch to activate 
Allows to set EOF timestamp.
 2020-06-06 15:53:07 +02:00
Wu Zhiwen
b6d7c4c1d4 dnn/native: fix typo for definition of DOT_INTERMEDIATE 
Signed-off-by: Wu Zhiwen <zhiwen.wu@intel.com>
Reviewed-by: Guo Yejun <yejun.guo@intel.com>
 2020-06-03 09:57:22 +08:00
Andreas Rheinhardt
317b722c51 avfilter/vf_lut3d: Fix mixed declaration and code 
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
 2020-06-01 15:21:40 +02:00
Mark Reid
a1221b96d8 avfilter/vf_lut3d: prelut support for 3d cinespace luts 
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-05-31 00:55:12 +02:00
Paul B Mahol
1329db8cfb avfilter/af_aiir: simplify polynomial evaluation 2020-05-30 18:04:14 +02:00
Paul B Mahol
aac16abd92 avfilter/af_aiir: use correct size when allocating in zp2tf 2020-05-30 18:04:14 +02:00
Paul B Mahol
726dbc57f8 avfilter: add dblur video filter 2020-05-30 18:04:14 +02:00
Jun Zhao
018cd437f8 lavfi/aiir: Refine the pad/vpad related operation 
move the pad/vpad related operation with more natural
coding style.

Signed-off-by: Jun Zhao <barryjzhao@tencent.com>
 2020-05-30 19:02:43 +08:00
Jun Zhao
ff8329a730 lavfi/afir: fix vpad.name leak 
Fix vpad.name leak in error path, move the vpad related operation
only if enabled show IR frequency response.

Signed-off-by: Jun Zhao <barryjzhao@tencent.com>
 2020-05-30 19:02:34 +08:00
Paul B Mahol
6485b54477 Revert "avfilter/af_aiir: move response drawing as last step" 
This reverts commit ca7095a907.
 2020-05-30 10:05:19 +02:00
Paul B Mahol
3fc7b01c52 avfilter/af_aiir: improve response calculation with zp coefficients 2020-05-30 10:05:19 +02:00
Paul B Mahol
e2e8121eaa avfilter/af_aiir: add S-plane support 2020-05-30 10:05:19 +02:00
Paul B Mahol
327b52412d avfilter/af_aiir: make it clear that transfer function is digital one 2020-05-30 10:05:19 +02:00
Paul B Mahol
1206a10d9c avfilter/af_biquads: implement 1st order allpass 2020-05-30 09:57:04 +02:00
Lynne
83fa39eb06
lavfi/vulkan: use av_get_random_seed instead of rand 
We need at least a few bits of entropy to determine the start index of each
queue, in order to let filters run in parallel as much as possible, and
rand() is not thread safe and disrupts any external API's usage of rand,
so instead replace it with av_get_random_seed.
While it has more overhead than rand, we only run it once per filter upon init.
 2020-05-29 13:10:58 +01:00
Ting Fu
f73cc61bf5 dnn_backend_native_layer_mathunary: add abs support 
more math unary operations will be added here

It can be tested with the model file generated with below python scripy:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpeg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.subtract(x, 0.5)
x2 = tf.abs(x1)
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Ting Fu <ting.fu@intel.com>
Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
 2020-05-28 11:04:21 +08:00
Paul B Mahol
d5f87f8567 avfilter/vf_blend: add support for float formats 2020-05-26 23:27:02 +02:00
Lynne
c0344cbfb0
lavfi/vulkan: fix queue counts and set indices 2020-05-26 10:52:11 +01:00
Lynne
fb49d5c0b1
lavfi/vulkan: use dedicated allocation for buffers when necessary 2020-05-26 10:52:11 +01:00
Lynne
727cac88b8
lavfi/vulkan: use all enabled queues in the queue family 
This should significantly improve the performance with certain
filterchains.
 2020-05-23 19:07:50 +01:00
Lynne
fac17fd46f
lavfi/vulkan: fix 2 minor memory leaks 2020-05-23 19:07:48 +01:00
Nicolas George
88567a2e52 lavfi: add untile filter. 2020-05-23 15:52:27 +02:00
Nicolas George
2a52f19a91 lavfi/framesync: use av_gcd_q(). 2020-05-23 15:51:45 +02:00
Nicolas George
beb98c0181 lavfi/tests/formats: reindent. 2020-05-23 15:50:20 +02:00
Nicolas George
d5e5c6862b lavfi/formats: remove dead code. 
Move the contents of all_channel_layouts.inc directly into
libavfilter/tests/formats.c.
 2020-05-23 15:50:20 +02:00
Nicolas George
df123590f0 lavfi/vf_crop: use ff_formats_pixdesc_filter(). 2020-05-23 15:50:20 +02:00
Nicolas George
563e1df5d6 lavfi/formats: add ff_formats_pixdesc_filter(). 2020-05-23 15:50:20 +02:00
Mark Reid
04f67dcccf avfilter/vf_lut3d: initial float pixel format support 
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-05-23 00:26:38 +02:00
Anton Khirnov
f30a41a608 Stop hardcoding align=32 in av_frame_get_buffer() calls. 
Use 0, which selects the alignment automatically.
 2020-05-22 14:38:57 +02:00
Paul B Mahol
ca7095a907 avfilter/af_aiir: move response drawing as last step 2020-05-22 14:14:15 +02:00
Paul B Mahol
8c825e43f8 avfilter/af_aiir: fix first denominator calculation 2020-05-22 14:12:06 +02:00
Paul B Mahol
07a9e5ec5e avfilter/af_aiir: add more descriptive options aliases 2020-05-22 12:37:17 +02:00
Paul B Mahol
ffda57b800 avfilter/af_aiir: export normalize option 
And enable it in all modes by default.
 2020-05-22 12:30:59 +02:00
Paul B Mahol
1fc5ddf774 avfilter/af_aiir: fix first delay value 2020-05-22 11:02:45 +02:00
Paul B Mahol
86822cfcd9 avfilter/af_aiir: fix phase and group delay calculation 
Properly unwrap phase.
 2020-05-20 12:08:32 +02:00
Paul B Mahol
b559a5882f avfilter/af_aiir: fix invalid memory access with tf filtering 2020-05-19 20:10:34 +02:00
Paul B Mahol
5646d02cb3 avfilter/vf_chromakey: fix formula for calculation of difference 2020-05-19 20:10:34 +02:00
Paul B Mahol
8b0575d763 avfilter/vf_colorkey: fix formula for calculation of difference 
Also fixes colorhold filtering.
 2020-05-19 20:10:34 +02:00
Paul B Mahol
f63939dedb avfilter: add gradients source video filter 2020-05-19 20:10:34 +02:00
Paul B Mahol
1ead7ed5bf avfilter/vsrc_sierpinski: unbreak configuring rate value 2020-05-16 11:13:46 +02:00
Paul B Mahol
d99ed7367b avfilter/vsrc_mandelbrot: unbreak configuring rate value 2020-05-16 11:10:32 +02:00
Marton Balint
b4bcae4e0e Revert "avfilter/vf_framerate: if metadata lavfi.scd.mafd exists, we'll use it first" 
This reverts commit 339593ca90.

Fixes null pointer dereference.

Signed-off-by: Marton Balint <cus@passwd.hu>
 2020-05-15 18:35:51 +02:00
Marton Balint
7f3a946216 Revert "avfilter/vf_minterpolate: if metadata lavfi.scd.mafd exists, we'll use it first" 
This reverts commit d88e1c9838.

Fixes null pointer dereference.

Signed-off-by: Marton Balint <cus@passwd.hu>
 2020-05-15 18:35:51 +02:00
Limin Wang
815a3b393c avfilter/vf_minterpolate: change the default threshold to get better scene change detect result 
./ffmpeg -loglevel debug -i ../fate-suite/svq3/Vertical400kbit.sorenson3.mov -vf
 minterpolate=fps=60:mi_mode=blend -an -f null -
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 1600
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 4120
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 5780
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 6700
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 8140
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 9740
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 14060
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 15680
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 18480
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 20020
 [Parsed_minterpolate_0 @ 0x7fe7f3e193c0] scene changed, input pts 21740

 The results are consistent with tests/ref/fate/filter-metadata-scenedetect

 For the master, it'll detect more than 20 scene change for the same source.

Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-14 23:03:07 +08:00
Limin Wang
61cc009d53 avfilter/vf_minterpolate: correct the mafd calculation 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-14 23:03:07 +08:00
Limin Wang
d88e1c9838 avfilter/vf_minterpolate: if metadata lavfi.scd.mafd exists, we'll use it first 
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-14 23:03:07 +08:00
Limin Wang
339593ca90 avfilter/vf_framerate: if metadata lavfi.scd.mafd exists, we'll use it first 
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-14 23:03:06 +08:00
Limin Wang
d62d4e05e8 avfilter/vf_scdet: add filter to detect scene change 
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-14 23:03:06 +08:00
lance.lmwang@gmail.com
5ed20a74b7 avfilter/vf_colorbalance:: Fix for framecrc bitexact for 32bit and 64bit system 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-14 09:15:46 +08:00
Martin Storsjö
353aecbb28 pixblockdsp, avdct: Add get_pixels_unaligned 
Use this in vf_spp.c, where the get_pixels operation is done on
unaligned source addresses.

Hook up the x86 (mmx and sse) versions of get_pixels to this
function pointer, as those implementations seem to support unaligned
use.

This fixes fate-filter-spp on armv7.

Signed-off-by: Martin Storsjö <martin@martin.st>
 2020-05-13 13:20:08 +03:00
Anton Khirnov
b84a2b91fd vf_showinfo: support AV_FRAME_DATA_VIDEO_ENC_PARAMS 2020-05-12 09:37:47 +02:00
Michael Niedermayer
52cc86cae1 avfilter/vf_spp: Fix endian-dependance in add_block() 
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-05-12 01:00:28 +02:00
Michael Niedermayer
dde86d0ddf avfilter/vf_spp: Remove unused AVCodecContext 
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2020-05-12 01:00:28 +02:00
Lynne
70d396c8af Revert "hwcontext_vulkan: only use one semaphore per image" 
This reverts commit 97b526c192.
It broke the API, and assumed no other APIs used multiple semaphores.
This also disallowed certain optimizations to happen.

Dealing with APIs that give or expect single semaphores is easier when
we use per-image semaphores.
 2020-05-11 23:48:26 +01:00
Limin Wang
a058d81570 avfilter/vf_signalstats: reindent after last commit 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-10 22:32:34 +08:00
Limin Wang
25bb72af62 avfilter/vf_signalstats: make the alloc array used for any bit depth 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-10 22:32:34 +08:00
Limin Wang
764efda192 avfilter/vf_signalstats: add SignalstatsContext.maxsize variable 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-10 22:32:34 +08:00
Limin Wang
a8c762a17d avfilter/vf_signalstats: rename config_props -> config_output 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-10 22:32:34 +08:00
Guo, Yejun
71e28c5422 dnn/native: add native support for minimum 
it can be tested with model file generated with below python script:
import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
x1 = tf.minimum(0.7, x)
x2 = tf.maximum(x1, 0.4)
y = tf.identity(x2, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
 2020-05-08 15:22:27 +08:00
Sebastian Dröge
ed3da72583 s->target_i and global are in dB but s->target_tp and true_peak are 
linear. Instead of mixing these in the calculations, convert the former
first to have all following calculations in the same unit.

Signed-off-by: Kyle Swanson <k@ylo.ph>
 2020-05-01 21:36:07 -07:00
Limin Wang
f441fadbcf avfilter/af_adelay: Check sscanf() return value 
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-01 06:36:36 +08:00
Limin Wang
ee5d6d2ef8 avfilter/vf_mix: Check sscanf() return value 
Reviewed-by: Nicolas George <george@nsup.org>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-05-01 06:36:36 +08:00
Paul B Mahol
0a181c2cb1 avfilter/af_crossfeed: make options runtime configurable 2020-04-30 12:18:36 +02:00
Paul B Mahol
35d6001815 avfilter/af_crossfeed: make slope configurable 2020-04-30 12:18:36 +02:00
Paul B Mahol
0ec61fcff7 avfilter/af_crossfeed: fix calculation of alpha parameter 
Use A in calculation.
 2020-04-30 12:18:36 +02:00
Paul B Mahol
c7d8082357 avfilter: add asubboost filter 2020-04-30 12:18:36 +02:00
Paul B Mahol
2970846fc1 avfilter/vf_v360: make FOV adjusted for dual fisheye too 
Remove any usage of padding for this format.
 2020-04-29 19:18:25 +02:00
Paul B Mahol
d8147c4e2d avfilter/af_amix: make weights option runtime configured 2020-04-29 19:18:25 +02:00
Steven Liu
d813e43b3d avfilter/vf_v360: adjustment out_pad and in_pad maximum value to 1/10 
Because not every user know about in_pad and out_pad reasonable value range
so maybe try to set 1.0, but setting 1.0 is so hugh to get an fatal error.

Suggested-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
 2020-04-29 13:48:02 +01:00
Paul B Mahol
f53fc935b8 avfilter/vf_pixscope: also show standard deviation of zoomed area 2020-04-26 12:27:36 +02:00
Paul B Mahol
35bcbfd6da avfilter/f_interleave: no need to check for inlink eof if non-empty queue 
Also set state to ready if there is any inlink with queued frame.
 2020-04-25 13:35:22 +02:00
Paul B Mahol
ec3b5deab4 avfilter/f_interleave: make sure that all frames in inlink queue are used 2020-04-25 13:07:13 +02:00
Paul B Mahol
b29b934e4f avfilter/f_interleave: add duration option 2020-04-25 12:35:22 +02:00
Lynne
b136a98303 scale_vulkan: take frame cropping parameters in account when scaling 
Then sample_aspect_ratio line at the bottom was cargo-culted from the vaapi
scaling filter, but its unnecesary.
 2020-04-23 18:23:45 +01:00
Lynne
f66ac83c22 overlay_vulkan: add support for overlaying images with an alpha channel 2020-04-23 18:23:45 +01:00
Guo, Yejun
8ce9d88f93 dnn/native: add native support for divide 
it can be tested with model file generated with below python script:
import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
z1 = 2 / x
z2 = 1 / z1
z3 = z2 / 0.25 + 0.3
z4 = z3 - x * 1.5 - 0.3
y = tf.identity(z4, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
 2020-04-22 13:15:00 +08:00
Guo, Yejun
ef79408e97 dnn/native: add native support for 'mul' 
it can be tested with model file generated from above python script:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
z1 = 0.5 + 0.3 * x
z2 = z1 * 4
z3 = z2 - x - 2.0
y = tf.identity(z3, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
 2020-04-22 13:14:47 +08:00
Guo, Yejun
6aa7e07e7c dnn/native: add native support for 'add' 
It can be tested with the model file generated with below python script:

import tensorflow as tf
import numpy as np
import imageio

in_img = imageio.imread('input.jpg')
in_img = in_img.astype(np.float32)/255.0
in_data = in_img[np.newaxis, :]

x = tf.placeholder(tf.float32, shape=[1, None, None, 3], name='dnn_in')
z1 = 0.039 + x
z2 = x + 0.042
z3 = z1 + z2
z4 = z3 - 0.381
z5 = z4 - x
y = tf.math.maximum(z5, 0.0, name='dnn_out')

sess=tf.Session()
sess.run(tf.global_variables_initializer())

graph_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, ['dnn_out'])
tf.train.write_graph(graph_def, '.', 'image_process.pb', as_text=False)

print("image_process.pb generated, please use \
path_to_ffmpeg/tools/python/convert.py to generate image_process.model\n")

output = sess.run(y, feed_dict={x: in_data})
imageio.imsave("out.jpg", np.squeeze(output))

Signed-off-by: Guo, Yejun <yejun.guo@intel.com>
 2020-04-22 13:14:30 +08:00
Paul B Mahol
4457f75c65 avfilter: add maskedthreshold filter 2020-04-18 12:34:49 +02:00
Paul B Mahol
9987f6d01e avfilter: add tmedian filter 2020-04-18 12:34:49 +02:00
Limin Wang
403bee30a5 avfilter/af_acrossover: Check sscanf() return value 
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
 2020-04-17 16:56:13 +02:00
Paul B Mahol
24db9372eb avfilter/af_astats: add noise floor count stat 2020-04-17 13:28:01 +02:00
Paul B Mahol
715da29501 avfilter/af_astats: add slice threading support 2020-04-17 13:26:44 +02:00