docs/filters: add documentation to all existing OpenCL filters
---
Hi! Sorry for delay.
Added general instruction for format conversion in the start of overall section
and separatly for overlay_opencl.
As far as I understand, different format conversion depends on which formats
are supported
r ff_vf_extractplanes;
extern AVFilter ff_vf_fade;
extern AVFilter ff_vf_fftdnoiz;
diff --git a/libavfilter/opencl/neighbor.cl b/libavfilter/opencl/neighbor.cl
new file mode 100644
index 000..2235e3e
--- /dev/null
+++ b/libavfilter/opencl/neighbor.cl
@@ -0,0 +1,87 @@
+/*
+ * Copyright (c) 2018 Danil I
es;
extern AVFilter ff_vf_fade;
extern AVFilter ff_vf_fftdnoiz;
diff --git a/libavfilter/opencl/neighbor.cl b/libavfilter/opencl/neighbor.cl
new file mode 100644
index 000..2235e3e
--- /dev/null
+++ b/libavfilter/opencl/neighbor.cl
@@ -0,0 +1,87 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
f_vf_extractplanes;
extern AVFilter ff_vf_fade;
extern AVFilter ff_vf_fftdnoiz;
diff --git a/libavfilter/opencl/neighbor.cl b/libavfilter/opencl/neighbor.cl
new file mode 100644
index 000..2235e3e
--- /dev/null
+++ b/libavfilter/opencl/neighbor.cl
@@ -0,0 +1,87 @@
+/*
+ * Copyright (c) 2
docs/filters: add documentation to all existing OpenCL filters
---
Thanks, fixed!
Danil.
doc/filters.texi | 415 +++
1 file changed, 415 insertions(+)
diff --git a/doc/filters.texi b/doc/filters.texi
index 705d48e..363fe7d 100644
---
docs/filters: add documentation to all existing OpenCL filters
---
Thanks for your comments! Most of the issues have been fixed.
>The filer source has many more options defined.
>In addition, there are many missing values for the tonemap algo; it appears
>not all are effected. If that's the
Add lumakey_opencl filter. Behaves like existing lumakey filter.
---
On Wed, Jul 25, 2018 at 10:50:43AM -0300, James Almer wrote:
>> On 7/25/2018 9:13 AM, Danil Iashchenko wrote:
>> > Add lumakey_opencl filter. Behaves like existing lumakey filter.
>>
>> Isn
Filter ff_vf_lut2;
extern AVFilter ff_vf_lut3d;
diff --git a/libavfilter/opencl/lumakey.cl b/libavfilter/opencl/lumakey.cl
new file mode 100644
index 000..dbee63e
--- /dev/null
+++ b/libavfilter/opencl/lumakey.cl
@@ -0,0 +1,43 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * Th
doc/filters: update info about avgblur filter to make it more understandable
---
doc/filters.texi | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 705d48e..e3005c3 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -5431,13
moved program_opencl and openclsrc filters to OpenCL section
---
doc/filters.texi | 404 +++
1 file changed, 202 insertions(+), 202 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index d206972..d9458cd 100644
---
docs/filters: add documentation to all existing OpenCL filters
---
doc/filters.texi | 458 +++
1 file changed, 458 insertions(+)
diff --git a/doc/filters.texi b/doc/filters.texi
index 705d48e..d206972 100644
--- a/doc/filters.texi
+++
Switch to use CL_FAIL_ON_ERROR for error handling
---
>Hi Danil,
>The patch looks good, But I think it would be more proper to use a title like
>"switch to use CL_FAIL_ON_ERROR " or "use CL_FAIL_ON_ERROR for error handling".
>Ruiling
Hi!
Changed title to a proper one.
Thanks!
Danil.
---
libavfilter/vf_convolution_opencl.c | 46 +
1 file changed, 11 insertions(+), 35 deletions(-)
diff --git a/libavfilter/vf_convolution_opencl.c
b/libavfilter/vf_convolution_opencl.c
index e2ba81b..00246b2 100644
--- a/libavfilter/vf_convolution_opencl.c
Fixed using uninitialized value "global_work[0]" when calling "av_log".
---
libavfilter/vf_avgblur_opencl.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/libavfilter/vf_avgblur_opencl.c b/libavfilter/vf_avgblur_opencl.c
index d1d3eb1..772b14e 100644
---
p;
+extern AVFilter ff_vf_vflip_opencl;
extern AVFilter ff_vf_vfrdet;
extern AVFilter ff_vf_vidstabdetect;
extern AVFilter ff_vf_vidstabtransform;
diff --git a/libavfilter/opencl/vflip.cl b/libavfilter/opencl/vflip.cl
new file mode 100644
index 000..4ed2f43
--- /dev/null
+++ b/libavfilter/opencl/
add documentation to all existing OpenCL filters
---
>I like this patch, the other thing, is it have any plan to update
>openCL part like VAAPI (https://trac.ffmpeg.org/wiki/Hardware/VAAPI)
>in https://trac.ffmpeg.org/wiki/HWAccelIntro, thanks.
I am going to add it during the next week, thanks!
Add documentation to all existing OpenCL filters, except tonemap filter.
---
Creation of the seperate section for OpenCL filters is motivated by developing
support for Vulkan and CUDA and it is easier to show the capabilities of each
platform by adding new sections.
doc/filters.texi | 193
The very last clFinish() should be ok.
Signed-off-by: Ruiling Song
---
libavfilter/vf_avgblur_opencl.c | 1 -
1 file changed, 1 deletion(-)
Looks good to me. Thanks!
Danil.
diff --git a/libavfilter/vf_avgblur_opencl.c b/libavfilter/vf_avgblur_opencl.c
index bc6bcab..99ed1ca 100644
---
Add opencl version of sobel, prewitt, roberts filters.
---
Fixed, Thanks!
Docs to all opencl-based filters will be added in another patch.
configure | 3 +
libavfilter/Makefile| 8 +-
libavfilter/allfilters.c| 3 +
Add opencl version of sobel, prewitt, roberts filters.
---
configure | 3 +
libavfilter/Makefile| 8 +-
libavfilter/allfilters.c| 3 +
libavfilter/opencl/convolution.cl | 82 ++
libavfilter/vf_convolution_opencl.c | 306
filter/boxblur.c
new file mode 100644
index 000..5ef3a05
--- /dev/null
+++ b/libavfilter/boxblur.c
@@ -0,0 +1,107 @@
+/*
+ * Copyright (c) 2002 Michael Niedermayer
+ * Copyright (c) 2011 Stefano Sabatini
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpe
er/boxblur.c b/libavfilter/boxblur.c
new file mode 100644
index 000..1d4895e
--- /dev/null
+++ b/libavfilter/boxblur.c
@@ -0,0 +1,105 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * mod
---
Hi!
I like your idea with OCL_FAIL_ON_ERR(), but still do not know which one is
better.
My idea relies on fact, that there are only few OpenCL functions which are used
multiple times in filters:
clSetKernelArg, clCreateKernel(in case when there are multiple kernels) and
maybe
---
libavfilter/opencl.h| 15 ++
libavfilter/vf_convolution_opencl.c | 43
libavfilter/vf_overlay_opencl.c | 44 +++-
libavfilter/vf_unsharp_opencl.c | 57 ++---
4 files changed,
/libavfilter/boxblur.c
new file mode 100644
index 000..1d4895e
--- /dev/null
+++ b/libavfilter/boxblur.c
@@ -0,0 +1,105 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the term
;
extern AVFilter ff_vf_ciescope;
diff --git a/libavfilter/boxblur.c b/libavfilter/boxblur.c
new file mode 100644
index 000..1d4895e
--- /dev/null
+++ b/libavfilter/boxblur.c
@@ -0,0 +1,105 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free
a/libavfilter/vf_avgblur_opencl.c b/libavfilter/vf_avgblur_opencl.c
index 48cebb5..d4759de 100644
--- a/libavfilter/vf_avgblur_opencl.c
+++ b/libavfilter/vf_avgblur_opencl.c
@@ -1,5 +1,6 @@
/*
* Copyright (c) 2018 Dylan Fernando
+ * Copyright (c) 2018 Danil Iashchenko
*
* This file is part of FFmpeg
Fixed, thanks!
Also implemented macro for other filters: avgblur_opencl, convolution_opencl,
unsharp_opencl, overlay_opencl.
---
libavfilter/opencl.h| 13 +
libavfilter/vf_avgblur_opencl.c | 48 ---
libavfilter/vf_convolution_opencl.c | 43
---
libavfilter/opencl.h| 13 +++
libavfilter/vf_convolution_opencl.c | 43 ++---
2 files changed, 19 insertions(+), 37 deletions(-)
diff --git a/libavfilter/opencl.h b/libavfilter/opencl.h
index c0a4519..16cdfbe 100644
---
v/null
+++ b/libavfilter/opencl/boxblur.cl
@@ -0,0 +1,53 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by
extern AVFilter ff_vf_bwdif;
extern AVFilter ff_vf_chromakey;
extern AVFilter ff_vf_ciescope;
diff --git a/libavfilter/opencl/boxblur.cl b/libavfilter/opencl/boxblur.cl
new file mode 100644
index 000..a0d9dcd
--- /dev/null
+++ b/libavfilter/opencl/boxblur.cl
@@ -0,0 +1,57 @@
+/*
+ * Copyright (c) 2
ution.cl
b/libavfilter/opencl/convolution.cl
new file mode 100644
index 000..c0748cc
--- /dev/null
+++ b/libavfilter/opencl/convolution.cl
@@ -0,0 +1,42 @@
+/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it a
+++ b/libavfilter/opencl/convolution.cl
@@ -1,4 +1,6 @@
/*
+ * Copyright (c) 2018 Danil Iashchenko
+ *
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
diff --git a/libavfilter/vf_convolution_opencl.c
b/libavfilter/vf_convolution_opencl.c
index
Hi there!
Thank you for advices, I have fixed the problem when per plane matrices
application was incorrect. Now it works as expected and behaves like the
existing vf_convolution filter.
Tested for yuv and nv12 formats.
The following filters from ffmpeg documentation
Behaves like the existing convolution filter, except working on OpenCL
hardware frames.
Takes exactly the same options: 4 convolution matrices, 4 rdiv values, 4 bias
values.
If not specified, default parameters are applied.
Matrices can be different sizes.
NEW IN THIS PATCH:
-fixed bug, if
Behaves like the existing convolution filter, except working on OpenCL
hardware frames.
Takes exactly the same options: 4 convolution matrices, 4 rdiv values, 4 bias
values.
If not specified, default parameters are applied.
Matrices can be different sizes.
filter applies:
matrix0, rdiv0, bias0
---
libavfilter/vf_convolution_opencl.c | 39 ++---
1 file changed, 15 insertions(+), 24 deletions(-)
diff --git a/libavfilter/vf_convolution_opencl.c
b/libavfilter/vf_convolution_opencl.c
index 60e2d1f..a14005b 100644
--- a/libavfilter/vf_convolution_opencl.c
---
configure | 1 +
libavfilter/Makefile| 1 +
libavfilter/allfilters.c| 1 +
libavfilter/opencl/convolution.cl | 42
libavfilter/opencl_source.h | 3 +
libavfilter/vf_convolution_opencl.c | 464
Hi there. Thank you for your advices. I implemented 4 matrix / 4 rdiv / 4 bias
option for each plane of image, fixed mem-leaks, added error messages if the
matrix is invalid.
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
Hi there. Thank you for your advices.
I implemented 4 matrix / 4 rdiv / 4 bias support, fixed mem-leaks,
add error messages if matrix is incorrect, removed local kernel.
Kind regards. Danil
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
---
configure | 1 +
libavfilter/Makefile| 2 +
libavfilter/allfilters.c| 1 +
libavfilter/opencl/convolution.cl | 80
libavfilter/opencl_source.h | 2 +
libavfilter/vf_convolution_opencl.c | 374
*** Hi everybody! I Implemented vf_convolution_opencl filter as qualification
task for GSoC Video filtering with OpenCL project***
Danil Iashchenko (1):
add vf_convolution_opencl
configure | 1 +
libavfilter/Makefile| 2 +
libavfilter
42 matches
Mail list logo