From: Gabriel Machado
Some scaling filters implemented as OpenCL kernels. Can be used as:
scale_opencl=::flags=
where can be `neighbor', `bilinear', `bicubic' or `fast_bicubic'
This is an initial draft, there's still a long way to go in terms of
completeness, config
100644
index 000..3436694
--- /dev/null
+++ b/libavfilter/opencl/scale.cl
@@ -0,0 +1,111 @@
+/*
+ * Copyright (c) 2018 Gabriel Machado
+ *
+ * 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 Publi
ilter/opencl/scale.cl
@@ -0,0 +1,111 @@
+/*
+ * Copyright (c) 2018 Gabriel Machado
+ *
+ * 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 the Free Software Foundatio
, scale2ref, vf);
diff --git a/libavfilter/opencl/scale.cl b/libavfilter/opencl/scale.cl
new file mode 100644
index 000..777344e
--- /dev/null
+++ b/libavfilter/opencl/scale.cl
@@ -0,0 +1,113 @@
+/*
+ * Copyright (c) 2018 Gabriel Machado
+ *
+ * This file is part of FFmpeg.
+ *
+ * FF
From: Gabriel Machado
Implemented the scaling filters using discrete convolution and ported most
of the filters from libswscale.
---
configure | 1 +
libavfilter/Makefile | 1 +
libavfilter/allfilters.c | 1 +
libavfilter/opencl/scale.cl | 67
On 3/31/18 10:08 PM Michael Niedermayer wrote:
> doesnt apply anymore
Fixed.
From 2b78da0f9d2e49d3ae5d5c8c751371666ded85f3 Mon Sep 17 00:00:00 2001
From: Gabriel Machado
Date: Sat, 31 Mar 2018 23:03:18 -0300
Subject: [PATCH] avfilter: add OpenCL scale filter
---
config
you intentionally need double precision, it is ok to leave it as it is.
> There are also some other places like this in the OpenCL file.
Changed to floating point constants, thanks.
- Gabriel
From b5cbdef9abc3727e1456915e50bbf1d9c87751d7 Mon Sep 17 00:00:00 2001
From: Gabriel Machado
Da
