Hi Sergey,
How should i test this filter?
I tested it some days ago, the picture get worse from 2nd frame.
input resolution 640x480 to 1280x720;
ffmpeg -i input -vf srcnn output
> On May 7, 2018, at 09:33, Pedro Arthur <bygran...@gmail.com> wrote:
>
> 2018-05-06 17:27 GMT-03:00 James Almer <jamr...@gmail.com>:
>> On 5/5/2018 5:38 PM, James Almer wrote:
>>> On 4/10/2018 2:16 PM, Sergey Lavrushkin wrote:
>>>> diff --git a/libavfilter/vf_srcnn.c b/libavfilter/vf_srcnn.c
>>>> new file mode 100644
>>>> index 0000000000..d9b4891f7f
>>>> --- /dev/null
>>>> +++ b/libavfilter/vf_srcnn.c
>>>> @@ -0,0 +1,420 @@
>>>> +/*
>>>> + * Copyright (c) 2018 Sergey Lavrushkin
>>>> + *
>>>> + * 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 Foundation; either
>>>> + * version 2.1 of the License, or (at your option) any later version.
>>>> + *
>>>> + * FFmpeg is distributed in the hope that it will be useful,
>>>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>>>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
>>>> + * Lesser General Public License for more details.
>>>> + *
>>>> + * You should have received a copy of the GNU Lesser General Public
>>>> + * License along with FFmpeg; if not, write to the Free Software
>>>> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
>>>> 02110-1301 USA
>>>> + */
>>>> +
>>>> +/**
>>>> + * @file
>>>> + * Filter implementing image super-resolution using deep convolutional
>>>> networks.
>>>> + * https://arxiv.org/abs/1501.00092
>>>> + */
>>>> +
>>>> +#include "avfilter.h"
>>>> +#include "formats.h"
>>>> +#include "internal.h"
>>>> +#include "libavutil/opt.h"
>>>> +#include "unistd.h"
>>>
>>> This broke msvc. It should be (if anything) <unistd.h>, and wrapped
>>> inside a HAVE_UNISTD_H preprocessor check.
>>
>> I (accidentally) applied this change as part of an unrelated commit, so
>> that's dealt with at least.
>>
>>>
>>>> +static av_cold int init(AVFilterContext* context)
>>>> +{
>>>> + SRCNNContext *srcnn_context = context->priv;
>>>> + AVIOContext* config_file_context;
>>>> + int64_t file_size, srcnn_size;
>>>> +
>>>> + /// Check specified confguration file name and read network weights
>>>> from it
>>>> + if (!srcnn_context->config_file_path){
>>>> + av_log(context, AV_LOG_INFO, "configuration file for network was
>>>> not specified, using default weights for x2 upsampling\n");
>>>> +
>>>> + /// Create convolution kernels and copy default weights
>>>> + srcnn_context->conv1.input_channels = 1;
>>>> + srcnn_context->conv1.output_channels = 64;
>>>> + srcnn_context->conv1.size = 9;
>>>> + CHECK_ALLOCATION(allocate_copy_conv_data(&srcnn_context->conv1,
>>>> conv1_kernel, conv1_biases), )
>>>> +
>>>> + srcnn_context->conv2.input_channels = 64;
>>>> + srcnn_context->conv2.output_channels = 32;
>>>> + srcnn_context->conv2.size = 1;
>>>> + CHECK_ALLOCATION(allocate_copy_conv_data(&srcnn_context->conv2,
>>>> conv2_kernel, conv2_biases), )
>>>> +
>>>> + srcnn_context->conv3.input_channels = 32;
>>>> + srcnn_context->conv3.output_channels = 1;
>>>> + srcnn_context->conv3.size = 5;
>>>> + CHECK_ALLOCATION(allocate_copy_conv_data(&srcnn_context->conv3,
>>>> conv3_kernel, conv3_biases), )
>>>> + }
>>>> + else if (access(srcnn_context->config_file_path, R_OK) != -1){
>>>
>>> This seems to be the function needed from unistd.h
>>>
>>> Looking at configure and libavformat/file.c it's apparently not
>>> available everywhere (we have a HAVE_ACCESS define, and also check for
>>> mode constants like R_OK).
>>
>> MSVC does not have access(), so i have temporarily made this filter
>> depend on its presence until a proper fix is applied. That target has
>> been broken for a few days now and that shouldn't be the case.
>> It should be a matter or doing the same thing as in libavformat/file.c,
>> i guess, or perhaps use _access() instead, which is available in MSVC.
> Thanks.
>
>
> @Sergey can you provide a definitive fix?
>
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel@ffmpeg.org
>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Thanks
Steven
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
