Go to the documentation of this file.
26 int main(
int argc,
char **argv)
30 unsigned int length = 0;
33 if (argc < 3 || argc > 4)
36 input = fopen(argv[1],
"rb");
40 output = fopen(argv[2],
"wb");
47 size_t arglen = strlen(argv[1]);
50 for (
int i = 0;
i < arglen;
i++) {
51 if (argv[1][
i] ==
'.')
53 else if (argv[1][
i] ==
'/')
58 fprintf(
output,
"const unsigned char ff_%s_data[] = { ",
name);
65 fprintf(
output,
"0x00 };\n");
66 fprintf(
output,
"const unsigned int ff_%s_len = %u;\n",
name, length);
static void error(const char *err)
AVFrame * ff_get_video_buffer(AVFilterLink *link, int w, int h)
Request a picture buffer with a specific set of permissions.
static av_always_inline void generate(int8_t *out, int out_stride, const uint8_t *in, int in_stride, H274FilmGrainDatabase *database, const AVFilmGrainH274Params *h274, int c, int invert, int deblock, int y_offset, int x_offset)
AVFrame * ff_get_audio_buffer(AVFilterLink *link, int nb_samples)
Request an audio samples buffer with a specific set of permissions.
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it be(in the first position) for now. Options ------- Then comes the options array. This is what will define the user accessible options. For example
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 L1
static void process(NormalizeContext *s, AVFrame *in, AVFrame *out)
static void draw_horiz_band(AVCodecContext *ctx, const AVFrame *fr, int offset[4], int slice_position, int type, int height)
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option name
they must not be accessed directly The fifo field contains the frames that are queued in the input for processing by the filter The status_in and status_out fields contains the queued status(EOF or error) of the link
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 L2
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C program
The reader does not expect b to be semantically here and if the code is changed by maybe adding a cast
static int mix(int c0, int c1)
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated space
status_out is the status that have been taken into it is final when it is not The typical task of an activate callback is to first check the backward status of output and if relevant forward it to the corresponding input Then
int av_frame_get_buffer(AVFrame *frame, int align)
Allocate new buffer(s) for audio or video data.
static const ElemCat * elements[ELEMENT_COUNT]
===============The purpose of these rules is to ensure that frames flow in the filter graph without getting stuck and accumulating somewhere. Simple filters that output one frame for each input frame should not have to worry about it. There are two design for filters:one using the filter_frame() and request_frame() callbacks and the other using the activate() callback. The design using filter_frame() and request_frame() is legacy, but it is suitable for filters that have a single input and process one frame at a time. New filters with several inputs, that treat several frames at a time or that require a special treatment at EOF should probably use the design using activate(). activate -------- This method is called when something must be done in a filter scheduling
static double cb(void *priv, double x, double y)
static void nothing(void *foo)
The official guide to swscale for confused that is
#define u(width, name, range_min, range_max)
The official guide to swscale for confused developers
int ff_filter_frame(AVFilterLink *link, AVFrame *frame)
Send a frame of data to the next filter.
#define AVERROR_EOF
End of file.
static int get_format(AVCodecContext *avctx, const enum AVPixelFormat *pix_fmts)
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is which means that it is highly recommended that you submit your filters to the FFmpeg development mailing list and make sure that they are applied your filters are likely to have a very short lifetime due to more or less regular internal API and a limited and testing changes the pixels in whatever fashion you and outputs the modified frame The most simple way of doing this is to take a similar filter We ll pick edgedetect
static const int8_t mv[256][2]
filter_frame For filters that do not use the this method is called when a frame is pushed to the filter s input It can be called at any time except in a reentrant way If the input frame is enough to produce output
The exact code depends on how similar the blocks are and how related they are to the and needs to apply these operations to the correct inlink or outlink if there are several Macros are available to factor that when no extra processing is inlink
static int16_t basis[64][64]
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
static const OMX_CALLBACKTYPE callbacks
This structure describes decoded (raw) audio or video data.
int av_frame_make_writable(AVFrame *frame)
Ensure that the frame data is writable, avoiding data copy if possible.
trying all byte sequences megabyte in length and selecting the best looking sequence will yield cases to try But a word about which is also called distortion Distortion can be quantified by almost any quality measurement one chooses the sum of squared differences is used but more complex methods that consider psychovisual effects can be used as well It makes no difference in this discussion First step
Note except for filters that can have queued frames and sources
static const uint16_t table[]
static int linear(InterplayACMContext *s, unsigned ind, unsigned col)
static uint8_t half(int a, int b)
int ff_request_frame(AVFilterLink *link)
Request an input frame from the filter at the other end of the link.
the frame and frame reference mechanism is intended to as much as possible
#define allocate(name, size)
filter_frame For filters that do not use the this method is called when a frame is pushed to the filter s input It can be called at any time except in a reentrant way If the input frame is enough to produce then the filter should push the output frames on the output link immediately As an exception to the previous rule if the input frame is enough to produce several output frames then the filter needs output only at least one per link The additional frames can be left buffered in the filter
status_in is a status change that must be taken into account after all frames in fifo have been processed
static int config_props(AVFilterLink *outlink)
static void add_bytes(HYuvDecContext *s, uint8_t *dst, uint8_t *src, int w)
static void put_pixel(uint16_t *dst, ptrdiff_t linesize, const int16_t *in, int bits_per_raw_sample)
Add bias value, clamp and output pixels of a slice.
A link between two filters.
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter Makefile
#define FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink)
Forward the status on an output link to an input link.
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 H0
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before ff_thread_await_progress() has been called on them. reget_buffer() and buffer age optimizations no longer work. *The contents of buffers must not be written to after ff_thread_report_progress() has been called on them. This includes draw_edges(). Porting codecs to frame threading
trying all byte sequences megabyte in length and selecting the best looking sequence will yield cases to try But a word about quality
static void convert(float y, float u, float v, float *b, float *g, float *r)
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before as well as code calling up to before the decode process starts Call have so the codec calls ff_thread_report set FF_CODEC_CAP_ALLOCATE_PROGRESS in FFCodec caps_internal and use ff_thread_get_buffer() to allocate frames. Otherwise decode directly into the user-supplied frames. Call ff_thread_report_progress() after some part of the current picture has decoded. A good place to put this is where draw_horiz_band() is called - add this if it isn 't called anywhere
#define bit(string, value)
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 Filter
int ff_inlink_consume_frame(AVFilterLink *link, AVFrame **rframe)
Take a frame from the link's FIFO and update the link's stats.
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a it should return
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that some(invalid) inputs can trigger overflows(undefined behavior). In these cases
static const chunk_decoder decoder[8]
if it could not for temporary reasons
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like for each output line the vertical scaler pulls lines from a ring buffer When the ring buffer does not contain the wanted then it is pulled from the input slice through the input converter and horizontal scaler The result is also stored in the ring buffer to serve future vertical scaler requests When no more output can be generated because lines from a future slice would be then all remaining lines in the current slice are horizontally scaled and put in the ring buffer[This is done for luma and chroma, each with possibly different numbers of lines per picture.] Input to YUV Converter When the input to the main path is not planar bits per component YUV or bit it is converted to planar bit YUV Two sets of converters exist for this the other leaves the full chroma resolution
if it could not because there are no more frames
status_out is the status that have been taken into it is final when it is not The typical task of an activate callback is to first check the backward status of output and if relevant forward it to the corresponding input if relevant
Writing a table generator This documentation is preliminary Parts of the API are not good and should be changed Basic concepts A table generator consists of two *_tablegen c and *_tablegen h The h file will provide the variable declarations and initialization code for the tables
FFmpeg s bug feature request tracker new issues and changes to existing issues can be done through a web interface Issues can be different kinds of things we want to keep track of but that do not belong into the source tree itself This includes bug feature requests and license violations We might add more items to this list in the future
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf type
static av_always_inline void update(SilenceDetectContext *s, AVFrame *insamples, int is_silence, int current_sample, int64_t nb_samples_notify, AVRational time_base)
static av_always_inline float scale(float x, float s)
status_out is the status that have been taken into account
uint8_t pi<< 24) CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_U8,(uint64_t)((*(const uint8_t *) pi - 0x80U))<< 56) CONV_FUNC(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8,(*(const uint8_t *) pi - 0x80) *(1.0f/(1<< 7))) CONV_FUNC(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8,(*(const uint8_t *) pi - 0x80) *(1.0/(1<< 7))) CONV_FUNC(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16,(*(const int16_t *) pi >>8)+0x80) CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S16, *(const int16_t *) pi *(1<< 16)) CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S16,(uint64_t)(*(const int16_t *) pi)<< 48) CONV_FUNC(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, *(const int16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, *(const int16_t *) pi *(1.0/(1<< 15))) CONV_FUNC(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32,(*(const int32_t *) pi >>24)+0x80) CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S32,(uint64_t)(*(const int32_t *) pi)<< 32) CONV_FUNC(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, *(const int32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, *(const int32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S64,(*(const int64_t *) pi >>56)+0x80) CONV_FUNC(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S64, *(const int64_t *) pi *(1.0f/(UINT64_C(1)<< 63))) CONV_FUNC(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S64, *(const int64_t *) pi *(1.0/(UINT64_C(1)<< 63))) CONV_FUNC(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, av_clip_uint8(lrintf(*(const float *) pi *(1<< 7))+0x80)) CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, av_clip_int16(lrintf(*(const float *) pi *(1<< 15)))) CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, av_clipl_int32(llrintf(*(const float *) pi *(1U<< 31)))) CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_FLT, llrintf(*(const float *) pi *(UINT64_C(1)<< 63))) CONV_FUNC(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, av_clip_uint8(lrint(*(const double *) pi *(1<< 7))+0x80)) CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, av_clip_int16(lrint(*(const double *) pi *(1<< 15)))) CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, av_clipl_int32(llrint(*(const double *) pi *(1U<< 31)))) CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_DBL, llrint(*(const double *) pi *(UINT64_C(1)<< 63))) #define FMT_PAIR_FUNC(out, in) static conv_func_type *const fmt_pair_to_conv_functions[AV_SAMPLE_FMT_NB *AV_SAMPLE_FMT_NB]={ FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8, AV_SAMPLE_FMT_U8), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_U8), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_U8), FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_U8), FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_U8), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_U8), FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8, AV_SAMPLE_FMT_S16), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S16), FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16), FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_S16), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_S16), FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8, AV_SAMPLE_FMT_S32), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S32), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S32), FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S32), FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_S32), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_S32), FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8, AV_SAMPLE_FMT_FLT), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLT), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_FLT), FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_FLT), FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_FLT), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_FLT), FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8, AV_SAMPLE_FMT_DBL), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_DBL), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_DBL), FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_DBL), FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_DBL), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_DBL), FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8, AV_SAMPLE_FMT_S64), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S64), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S64), FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S64), FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_S64), FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_S64), };static void cpy1(uint8_t **dst, const uint8_t **src, int len){ memcpy(*dst, *src, len);} static void cpy2(uint8_t **dst, const uint8_t **src, int len){ memcpy(*dst, *src, 2 *len);} static void cpy4(uint8_t **dst, const uint8_t **src, int len){ memcpy(*dst, *src, 4 *len);} static void cpy8(uint8_t **dst, const uint8_t **src, int len){ memcpy(*dst, *src, 8 *len);} AudioConvert *swri_audio_convert_alloc(enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt, int channels, const int *ch_map, int flags) { AudioConvert *ctx;conv_func_type *f=fmt_pair_to_conv_functions[av_get_packed_sample_fmt(out_fmt)+AV_SAMPLE_FMT_NB *av_get_packed_sample_fmt(in_fmt)];if(!f) return NULL;ctx=av_mallocz(sizeof(*ctx));if(!ctx) return NULL;if(channels==1){ in_fmt=av_get_planar_sample_fmt(in_fmt);out_fmt=av_get_planar_sample_fmt(out_fmt);} ctx->channels=channels;ctx->conv_f=f;ctx->ch_map=ch_map;if(in_fmt==AV_SAMPLE_FMT_U8||in_fmt==AV_SAMPLE_FMT_U8P) memset(ctx->silence, 0x80, sizeof(ctx->silence));if(out_fmt==in_fmt &&!ch_map) { switch(av_get_bytes_per_sample(in_fmt)){ case 1:ctx->simd_f=cpy1;break;case 2:ctx->simd_f=cpy2;break;case 4:ctx->simd_f=cpy4;break;case 8:ctx->simd_f=cpy8;break;} } return ctx;} void swri_audio_convert_free(AudioConvert **ctx) { av_freep(ctx);} int swri_audio_convert(AudioConvert *ctx, AudioData *out, AudioData *in, int len) { int ch;int off=0;const int os=(out->planar ? 1 :out->ch_count) *out->bps;unsigned misaligned=0;av_assert0(ctx->channels==out->ch_count);if(ctx->in_simd_align_mask) { int planes=in->planar ? in->ch_count :1;unsigned m=0;for(ch=0;ch< planes;ch++) m|=(intptr_t) in->ch[ch];misaligned|=m &ctx->in_simd_align_mask;} if(ctx->out_simd_align_mask) { int planes=out->planar ? out->ch_count :1;unsigned m=0;for(ch=0;ch< planes;ch++) m|=(intptr_t) out->ch[ch];misaligned|=m &ctx->out_simd_align_mask;} if(ctx->simd_f &&!ctx->ch_map &&!misaligned){ off=len &~15;av_assert1(off >=0);av_assert1(off<=len);av_assert2(ctx->channels==SWR_CH_MAX||!in->ch[ctx->channels]);if(off >0){ if(out->planar==in->planar){ int planes=out->planar ? out->ch_count :1;for(ch=0;ch< planes;ch++){ ctx->simd_f(out->ch+ch,(const uint8_t **) in->ch+ch, off *(out-> planar
static void idct(int16_t block[64])
A filter pad used for either input or output.
spatial_decomposition_type s header_state qlog s header_state mv_scale s header_state qbias s header_state block_max_depth s header_state qlogs
static int aligned(int val)
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s your new playground is ready Some little details about what s going which in turn will define variables for the build system and the C
trying all byte sequences megabyte in length and selecting the best looking sequence will yield cases to try But first
FFmpeg currently uses a custom build this text attempts to document some of its obscure features and options Makefile variables
static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, int *outFilterSize, int xInc, int srcW, int dstW, int filterAlign, int one, int flags, int cpu_flags, SwsVector *srcFilter, SwsVector *dstFilter, double param[2], int srcPos, int dstPos)
static void set(uint8_t *a[], int ch, int index, int ch_count, enum AVSampleFormat f, double v)
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before as well as code calling up to before the decode process starts Call have so the codec calls ff_thread_report await_progress()
void ff_thread_report_progress(ThreadFrame *f, int n, int field)
Notify later decoding threads when part of their reference picture is ready.
#define check(x, y, S, v)
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like YUV
static void decode(AVCodecContext *dec_ctx, AVPacket *pkt, AVFrame *frame, FILE *outfile)
static void ff_outlink_set_status(AVFilterLink *link, int status, int64_t pts)
Set the status field of a link from the source filter.
void ff_inlink_request_frame(AVFilterLink *link)
Mark that a frame is wanted on the link.
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not matter(as it is from invalid input). In some cases the input can be checked easily in others checking the input is computationally too intensive. In these remaining cases a unsigned type can be used instead of a signed type. unsigned overflows are defined in C. SUINT ----- As we have above established there is a need to use "unsigned" sometimes in computations which work with signed integers(which overflow). Using "unsigned" for signed integers has the very significant potential to cause confusion as in unsigned a
static const int offsets[]
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample format(the sample packing is implied by the sample format) and sample rate. The lists are not just lists
#define SWS_FULL_CHR_H_INP
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is which means that it is highly recommended that you submit your filters to the FFmpeg development mailing list and make sure that they are applied your filters are likely to have a very short lifetime due to more or less regular internal API changes
int(* init)(AVBSFContext *ctx)
#define filters(fmt, type, inverse, clp, inverset, clip, one, clip_fn, packed)
This is the more generic form
Note except for filters that can have queued frames and request_frame does not push and as a reaction
the buffer is automatically deallocated once all corresponding references have been destroyed The characteristics of the data(resolution, sample rate, etc.) are stored in the reference
int flags
A combination of AVFILTER_FLAG_*.
Undefined Behavior In the C some operations are like signed integer dereferencing freed pointers
static const double htaps[HTAPS]
The 2nd half (48 coeffs) of a 96-tap symmetric lowpass filter.
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s your new playground is ready Some little details about what s going on
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this field
write_fileheader() adds some minor things like a "this is a generated file" comment and some standard includes. tablegen.h defines some write functions for one- and two-dimensional arrays for standard types - they print only the "core" parts so they are easier to reuse for multi-dimensional arrays so the outermost
static int pix_norm1(const uint8_t *pix, int line_size, int w)
static int export(AVFilterContext *ctx, StreamContext *sc, int input)
=============================================Slice threading - *The client 's draw_horiz_band() must be thread-safe according to the comment in avcodec.h. Frame threading - *Restrictions with slice threading also apply. *Custom get_buffer2() and get_format() callbacks must be thread-safe. *There is one frame of delay added for every thread beyond the first one. Clients must be able to handle this clients
#define XMM_CLOBBERS(...)
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a link
int ff_inlink_make_frame_writable(AVFilterLink *link, AVFrame **rframe)
Make sure a frame is writable.
static void callback(void *priv_data, int index, uint8_t *buf, int buf_size, int64_t time, enum dshowDeviceType devtype)
must be printed separately If there s no standard function for printing the type you the WRITE_1D_FUNC_ARGV macro is a very quick way to create one See libavcodec dv_tablegen c for an example The h file This file should the initialization functions should not do and instead of the variable declarations the generated *_tables h file should be included Since that will be generated in the build the path must be included
the definition of that something depends on the semantic of the filter The callback must examine the status of the filter s links and proceed accordingly The status of output links is stored in the status_in and status_out fields and tested by the then the processing requires a frame on this link and the filter is expected to make efforts in that direction The status of input links is stored by the fifo and status_out fields
vertical halfpel samples are found by H2[y][x]
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option keep it simple and lowercase description are in without and describe what they for example set the foo of the bar offset is the offset of the field in your context
#define AV_CODEC_CAP_FRAME_THREADS
Codec supports frame-level multithreading.
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format negotiation
must be printed separately If there s no standard function for printing the type you need
static const int16_t quant_table[64]
and forward the result(frame or status change) to the corresponding input. If nothing is possible
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across frames
int av_frame_copy_props(AVFrame *dst, const AVFrame *src)
Copy only "metadata" fields from src to dst.
static int bias(int x, int c)
static void push(HysteresisContext *s, int x, int y, int w)
must be printed separately If there s no standard function for printing the type you the WRITE_1D_FUNC_ARGV macro is a very quick way to create one See libavcodec dv_tablegen c for an example The h file This file should the initialization functions should not do anything
the definition of that something depends on the semantic of the filter The callback must examine the status of the filter s links and proceed accordingly The status of output links is stored in the frame_wanted_out
#define LIBAVFILTER_VERSION_MICRO
must be printed separately If there s no standard function for printing the type you the WRITE_1D_FUNC_ARGV macro is a very quick way to create one See libavcodec dv_tablegen c for an example The h file This file should contain
static const int8_t transform[32][32]
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option keep it simple and lowercase description are in without period
static double coefficients[8 *8]
filter_frame For filters that do not use the activate() callback
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 this can end with a L or a H
FFmpeg currently uses a custom build system
must be printed separately If there s no standard function for printing the type you the WRITE_1D_FUNC_ARGV macro is a very quick way to create one See libavcodec dv_tablegen c for an example The h file This file should the initialization functions should not do and instead of the variable declarations the generated *_tables h file should be included Since that will be generated in the build the path must be i e not Makefile changes To make the automatic table creation work
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining list
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If not
static const struct codec_string codecs[]
static void difference(IPlane *g, IPlane *f, int y0, int y1)
different references for the same buffer can show different characteristics In particular
size_t av_cpu_max_align(void)
Get the maximum data alignment that may be required by FFmpeg.
static int filter_frame(DBEDecodeContext *s, AVFrame *frame)
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several inputs
static void get(const uint8_t *pixels, int stride, int16_t *block)
static int resample(SwrContext *s, AudioData *out_param, int out_count, const AudioData *in_param, int in_count)
int ff_inlink_acknowledge_status(AVFilterLink *link, int *rstatus, int64_t *rpts)
Test and acknowledge the change of status on the link.
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not c
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like for each output line the vertical scaler pulls lines from a ring buffer When the ring buffer does not contain the wanted then it is pulled from the input slice through the input converter and horizontal scaler The result is also stored in the ring buffer to serve future vertical scaler requests When no more output can be generated because lines from a future slice would be then all remaining lines in the current slice are converted
static av_always_inline void mv_scale(Mv *dst, const Mv *src, int td, int tb)
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 H1
FFmpeg multithreading methods
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a source
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i http
static void interleave(uint8_t *dst, uint8_t *src, int w, int h, int dst_linesize, int src_linesize, enum FilterMode mode, int swap)
int main(int argc, char **argv)
FF_FILTER_FORWARD_STATUS_ALL(outlink, filter)
const OptionDef options[]
static void blur(uint8_t *dst, int dst_step, const uint8_t *src, int src_step, int len, int radius, int pixsize)
vertical horizontal halfpel samples are found by H3[y][x]
static AVFrame * get_audio_buffer(AVFilterLink *inlink, int nb_samples)
void ff_inlink_set_status(AVFilterLink *link, int status)
Set the status on an input link.
static void copy(const float *p1, float *p2, const int length)
int av_frame_ref(AVFrame *dst, const AVFrame *src)
Set up a new reference to the data described by the source frame.
static int shift(int a, int b)
OutputFile ** output_files
#define FF_CODEC_CAP_ALLOCATE_PROGRESS
static int pix_sum(const uint8_t *pix, int line_size, int w, int h)
int av_frame_is_writable(AVFrame *frame)
Check if the frame data is writable.
static int64_t prediction(int delta, ChannelContext *c)
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s your new playground is ready Some little details about what s going which in turn will define variables for the build system and the and we are assuming vf_foobar is as well We are also assuming vf_foobar is not an edge detector so you can update the boilerplate with your credits Doxy Next chunk is the Doxygen about the file See does
printf("static const uint8_t my_array[100] = {\n")
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like for each output line the vertical scaler pulls lines from a ring buffer When the ring buffer does not contain the wanted then it is pulled from the input slice through the input converter and horizontal scaler The result is also stored in the ring buffer to serve future vertical scaler requests When no more output can be generated because lines from a future slice would be then all remaining lines in the current slice are horizontally scaled and put in the ring buffer[This is done for luma and chroma, each with possibly different numbers of lines per picture.] Input to YUV Converter When the input to the main path is not planar bits per component YUV or bit gray
const char * description
A description of the filter.
static const uint8_t header[24]
static const uint8_t *BS_FUNC() align(BSCTX *bc)
Skip bits to a byte boundary.
static char * split(char *message, char delim)
#define CONFIG_FOOBAR_FILTER
static void encode(AVCodecContext *ctx, AVFrame *frame, AVPacket *pkt, FILE *output)
The reader does not expect b to be semantically here and if the code is changed by maybe adding a a division or other the signedness will almost certainly be mistaken To avoid this confusion a new type was SUINT is the C unsigned type but it holds a signed int to use the same example SUINT a
FF_FILTER_FORWARD_WANTED(outlink, inlink)
static int forward(AudioFWTDNContext *s, const double *in, int in_length, double **out, int *out_length, int ch, uint64_t sn)
if it could not because there are no more it should return AVERROR_EOF The typical implementation of request_frame for a filter with several inputs will look like that
static double fact(double i)
static void predictor(uint8_t *src, ptrdiff_t size)
and forward the test the status of outputs and forward it to the corresponding return FFERROR_NOT_READY If the filters stores internally one or a few frame for some input
static int interpolation(DeclickChannel *c, const double *src, int ar_order, double *acoefficients, int *index, int nb_errors, double *auxiliary, double *interpolated)
The reader does not expect b to be semantically here and if the code is changed by maybe adding a a division or other the signedness will almost certainly be mistaken To avoid this confusion a new type was introduced
#define XMM_CLOBBERS_ONLY(...)
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec implementations
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is which means that it is highly recommended that you submit your filters to the FFmpeg development mailing list and make sure that they are applied your filters are likely to have a very short lifetime due to more or less regular internal API and a limited distribution
#define AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC
Some filters support a generic "enable" expression option that can be used to enable or disable a fil...
static int64_t filesize(AVIOContext *pb)
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not b
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel layout
#define i(width, name, range_min, range_max)
and forward the test the status of outputs and forward it to the corresponding return FFERROR_NOT_READY If the filters stores internally one or a few frame for some it can consider them to be part of the FIFO and delay acknowledging a status change accordingly Example code
int w
agreed upon image width
#define LIBAVFILTER_VERSION_MINOR
The exact code depends on how similar the blocks are and how related they are to the and needs to apply these operations to the correct inlink or outlink if there are several Macros are available to factor that when no extra processing is needed
enum AVPacketSideDataType packet
any process which generates a stream compliant to the syntactical and semantic requirements and which is decodable by the process described in this spec shall be considered a conformant Snow encoder but not strictly required ilog2(x) is the rounded down logarithm of x with basis 2 ilog2(0)=0Type definitions
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default value
#define CONFIG_HARDCODED_TABLES
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf description
#define MC(PEL, DIR, WIDTH, TAP, DIR1, FILT_DIR)
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s your new playground is ready Some little details about what s going which in turn will define variables for the build system and the and we are assuming vf_foobar is as well We are also assuming vf_foobar is not an edge detector filter
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before as well as code calling up to before the decode process starts Call have update_thread_context() run it in the next thread. Add AV_CODEC_CAP_FRAME_THREADS to the codec capabilities. There will be very little speed gain at this point but it should work. If there are inter-frame dependencies
static int headroom(int *la)
static const uint8_t plain[]
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like for each output line the vertical scaler pulls lines from a ring buffer When the ring buffer does not contain the wanted then it is pulled from the input slice through the input converter and horizontal scaler The result is also stored in the ring buffer to serve future vertical scaler requests When no more output can be generated because lines from a future slice would be needed
FFmpeg currently uses a custom build this text attempts to document some of its obscure features and options Makefile the full command issued by make and its output will be shown on the screen DBG Preprocess x86 external assembler files to a dbg asm file in the object directory
Undefined Behavior In the C language
static int array[MAX_W *MAX_W]
Writing a table generator This documentation is preliminary Parts of the API are not good and should be changed Basic concepts A table generator consists of two files
enum AVPixelFormat pixfmt
__asm__(".macro parse_r var r\n\t" "\\var = -1\n\t" _IFC_REG(0) _IFC_REG(1) _IFC_REG(2) _IFC_REG(3) _IFC_REG(4) _IFC_REG(5) _IFC_REG(6) _IFC_REG(7) _IFC_REG(8) _IFC_REG(9) _IFC_REG(10) _IFC_REG(11) _IFC_REG(12) _IFC_REG(13) _IFC_REG(14) _IFC_REG(15) _IFC_REG(16) _IFC_REG(17) _IFC_REG(18) _IFC_REG(19) _IFC_REG(20) _IFC_REG(21) _IFC_REG(22) _IFC_REG(23) _IFC_REG(24) _IFC_REG(25) _IFC_REG(26) _IFC_REG(27) _IFC_REG(28) _IFC_REG(29) _IFC_REG(30) _IFC_REG(31) ".iflt \\var\n\t" ".error \"Unable to parse register name \\r\"\n\t" ".endif\n\t" ".endm")
static const float pred[4]
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like for each output line the vertical scaler pulls lines from a ring buffer When the ring buffer does not contain the wanted then it is pulled from the input slice through the input converter and horizontal scaler The result is also stored in the ring buffer to serve future vertical scaler requests When no more output can be generated because lines from a future slice would be then all remaining lines in the current slice are horizontally scaled and put in the ring buffer[This is done for luma and chroma, each with possibly different numbers of lines per picture.] Input to YUV Converter When the input to the main path is not planar bits per component YUV or bit it is converted to planar bit YUV Two sets of converters exist for this currently
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output links
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a frame
static int FUNC() comment(CodedBitstreamContext *ctx, RWContext *rw, JPEGRawComment *current)
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before as well as code calling up to before the decode process starts Call ff_thread_finish_setup() afterwards. If some code can 't be moved
static int request_frame(AVFilterLink *outlink)
Tag MUST be and< 10hcoeff half pel interpolation filter coefficients, hcoeff[0] are the 2 middle coefficients[1] are the next outer ones and so on, resulting in a filter like:...eff[2], hcoeff[1], hcoeff[0], hcoeff[0], hcoeff[1], hcoeff[2] ... the sign of the coefficients is not explicitly stored but alternates after each coeff and coeff[0] is positive, so ...,+,-,+,-,+,+,-,+,-,+,... hcoeff[0] is not explicitly stored but found by subtracting the sum of all stored coefficients with signs from 32 hcoeff[0]=32 - hcoeff[1] - hcoeff[2] - ... a good choice for hcoeff and htaps is htaps=6 hcoeff={40,-10, 2} an alternative which requires more computations at both encoder and decoder side and may or may not be better is htaps=8 hcoeff={42,-14, 6,-2}ref_frames minimum of the number of available reference frames and max_ref_frames for example the first frame after a key frame always has ref_frames=1spatial_decomposition_type wavelet type 0 is a 9/7 symmetric compact integer wavelet 1 is a 5/3 symmetric compact integer wavelet others are reserved stored as delta from last, last is reset to 0 if always_reset||keyframeqlog quality(logarithmic quantizer scale) stored as delta from last, last is reset to 0 if always_reset||keyframemv_scale stored as delta from last, last is reset to 0 if always_reset||keyframe FIXME check that everything works fine if this changes between framesqbias dequantization bias stored as delta from last, last is reset to 0 if always_reset||keyframeblock_max_depth maximum depth of the block tree stored as delta from last, last is reset to 0 if always_reset||keyframequant_table quantization tableHighlevel bitstream structure:==============================--------------------------------------------|Header|--------------------------------------------|------------------------------------|||Block0||||split?||||yes no||||......... intra?||||:Block01 :yes no||||:Block02 :....... ..........||||:Block03 ::y DC ::ref index:||||:Block04 ::cb DC ::motion x :||||......... :cr DC ::motion y :||||....... ..........|||------------------------------------||------------------------------------|||Block1|||...|--------------------------------------------|------------ ------------ ------------|||Y subbands||Cb subbands||Cr subbands||||--- ---||--- ---||--- ---|||||LL0||HL0||||LL0||HL0||||LL0||HL0|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||LH0||HH0||||LH0||HH0||||LH0||HH0|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||HL1||LH1||||HL1||LH1||||HL1||LH1|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||HH1||HL2||||HH1||HL2||||HH1||HL2|||||...||...||...|||------------ ------------ ------------|--------------------------------------------Decoding process:=================------------|||Subbands|------------||||------------|Intra DC||||LL0 subband prediction ------------|\ Dequantization ------------------- \||Reference frames|\ IDWT|------- -------|Motion \|||Frame 0||Frame 1||Compensation . OBMC v -------|------- -------|--------------. \------> Frame n output Frame Frame<----------------------------------/|...|------------------- Range Coder:============Binary Range Coder:------------------- The implemented range coder is an adapted version based upon "Range encoding: an algorithm for removing redundancy from a digitised message." by G. N. N. Martin. The symbols encoded by the Snow range coder are bits(0|1). The associated probabilities are not fix but change depending on the symbol mix seen so far. bit seen|new state ---------+----------------------------------------------- 0|256 - state_transition_table[256 - old_state];1|state_transition_table[old_state];state_transition_table={ 0, 0, 0, 0, 0, 0, 0, 0, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 190, 191, 192, 194, 194, 195, 196, 197, 198, 199, 200, 201, 202, 202, 204, 205, 206, 207, 208, 209, 209, 210, 211, 212, 213, 215, 215, 216, 217, 218, 219, 220, 220, 222, 223, 224, 225, 226, 227, 227, 229, 229, 230, 231, 232, 234, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 248, 0, 0, 0, 0, 0, 0, 0};FIXME Range Coding of integers:------------------------- FIXME Neighboring Blocks:===================left and top are set to the respective blocks unless they are outside of the image in which case they are set to the Null block top-left is set to the top left block unless it is outside of the image in which case it is set to the left block if this block has no larger parent block or it is at the left side of its parent block and the top right block is not outside of the image then the top right block is used for top-right else the top-left block is used Null block y, cb, cr are 128 level, ref, mx and my are 0 Motion Vector Prediction:=========================1. the motion vectors of all the neighboring blocks are scaled to compensate for the difference of reference frames scaled_mv=(mv *(256 *(current_reference+1)/(mv.reference+1))+128)> the median of the scaled left
static const int16_t steps[16]
main external API structure.
static const ParseSyntax syntax[]
static float compare(const AVFrame *haystack, const AVFrame *obj, int offx, int offy)
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining again
int h
agreed upon image height
the frame and frame reference mechanism is intended to as much as expensive copies of that data while still allowing the filters to produce correct results The data is stored in buffers represented by AVFrame structures Several references can point to the same frame buffer
static const AVFilterPad outputs[]
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before as well as code calling get_buffer()
FFmpeg currently uses a custom build this text attempts to document some of its obscure features and options Makefile the full command issued by make and its output will be shown on the screen DBG Preprocess x86 external assembler files to a dbg asm file in the object which then gets compiled Helps in developing those assembler files DESTDIR Destination directory for the install useful to prepare packages or install FFmpeg in cross environments GEN Set to ‘1’ to generate the missing or mismatched references Makefile builds all the libraries and the executables fate Run the fate test note that you must have installed it fate list List all fate regression test targets install Install headers
A Quick Description Of Rate Distortion Theory We want to encode a video
static void test(const char *pattern, const char *host)
static int swscale(SwsContext *c, const uint8_t *src[], int srcStride[], int srcSliceY, int srcSliceH, uint8_t *dst[], int dstStride[], int dstSliceY, int dstSliceH)
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return values
static int ref[MAX_W *MAX_W]
Filter the word “frame” indicates either a video frame or a group of audio samples
static float mean(const float *input, int size)
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 note
int ff_outlink_get_status(AVFilterLink *link)
Get the status on an output link.
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s your new playground is ready Some little details about what s going which in turn will define variables for the build system and the and we are assuming vf_foobar is as well We are also assuming vf_foobar is not an edge detector so you can update the boilerplate with your credits Doxy Next chunk is the Doxygen about the file See and add some references if you feel like it Context Skip the headers and scroll down to the definition of FoobarContext This is your state context It is already filled with when you get it so do not worry about uninitialized reads into this context This is where you put all global information that you need
static const int factor[16]
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s your new playground is ready Some little details about what s going which in turn will define variables for the build system and the and we are assuming vf_foobar is as well We are also assuming vf_foobar is not an edge detector so you can update the boilerplate with your credits Doxy Next chunk is the Doxygen about the file See https
static int bad(InterplayACMContext *s, unsigned ind, unsigned col)
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is which means that it is highly recommended that you submit your filters to the FFmpeg development mailing list and make sure that they are applied your filters are likely to have a very short lifetime due to more or less regular internal API and a limited review
void(* flush)(AVBSFContext *ctx)
Undefined Behavior In the C some operations are like signed integer overflow
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 L3
s EdgeDetect Foobar g libavfilter vf_edgedetect c libavfilter vf_foobar c edit libavfilter and add an entry for foobar following the pattern of the other filters edit libavfilter allfilters and add an entry for foobar following the pattern of the other filters configure make j< whatever > ffmpeg ffmpeg i you should get a foobar png with Lena edge detected That s it
static const struct @91 transforms[18]
static float project(float origin_x, float origin_y, float dest_x, float dest_y, float point_x, float point_y, int type)
FF_FILTER_FORWARD_STATUS(inlink, outlink)
static double cr(void *priv, double x, double y)
the definition of that something depends on the semantic of the filter The callback must examine the status of the filter s links and proceed accordingly The status of output links is stored in the status_in and status_out fields and tested by the then the processing requires a frame on this link and the filter is expected to make efforts in that direction The status of input links is stored by the status_in
the frame and frame reference mechanism is intended to avoid
the pkt_dts and pkt_pts fields in AVFrame will work as usual Restrictions on codec whose streams don t reset across will not work because their bitstreams cannot be decoded in parallel *The contents of buffers must not be read before as well as code calling up to before the decode process starts Call have so the codec calls ff_thread_report set FF_CODEC_CAP_ALLOCATE_PROGRESS in FFCodec caps_internal and use as it s useful too and the implementation is trivial when you re doing this Note that draw_edges() needs to be called before reporting progress. Before accessing a reference frame or its MVs
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already ready
static pthread_once_t once
static int query_formats(AVFilterContext *ctx)
int linesize[AV_NUM_DATA_POINTERS]
For video, a positive or negative value, which is typically indicating the size in bytes of each pict...
static const double coeff[2][5]
The exact code depends on how similar the blocks are and how related they are to the block
static int sse(MpegEncContext *s, const uint8_t *src1, const uint8_t *src2, int w, int h, int stride)
static av_cold int uninit(AVCodecContext *avctx)
the definition of that something depends on the semantic of the filter The callback must examine the status of the filter s links and proceed accordingly The status of output links is stored in the status_in and status_out fields and tested by the ff_outlink_frame_wanted() function. If this function returns true
The official guide to swscale for confused that consecutive non overlapping rectangles of dimension(0, slice_top) -(picture_width
static void diff_bytes(HYuvEncContext *s, uint8_t *dst, const uint8_t *src0, const uint8_t *src1, int w)
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is which means that it is highly recommended that you submit your filters to the FFmpeg development mailing list and make sure that they are applied Otherwise
This document is a tutorial initiation for writing simple filters in libavfilter libavfilter is monolithic
int main(int argc, char **argv)
F H1 F F H1 F F F F H1<-F-------F-------F v v v H2 H3 H2 ^ ^ ^ F-------F-------F-> H1<-F-------F-------F|||||||||F H1 F|||||||||F H1 Funavailable fullpel samples(outside the picture for example) shall be equalto the closest available fullpel sampleSmaller pel interpolation:--------------------------if diag_mc is set then points which lie on a line between 2 vertically, horizontally or diagonally adjacent halfpel points shall be interpolatedlinearly with rounding to nearest and halfway values rounded up.points which lie on 2 diagonals at the same time should only use the onediagonal not containing the fullpel point F--> O q O<--h1-> O q O<--F v \/v \/v O O O O O O O|/|\|q q q q q|/|\|O O O O O O O ^/\ ^/\ ^ h2--> O q O<--h3-> O q O<--h2 v \/v \/v O O O O O O O|\|/|q q q q q|\|/|O O O O O O O ^/\ ^/\ ^ F--> O q O<--h1-> O q O<--Fthe remaining points shall be bilinearly interpolated from theup to 4 surrounding halfpel and fullpel points, again rounding should be tonearest and halfway values rounded upcompliant Snow decoders MUST support 1-1/8 pel luma and 1/2-1/16 pel chromainterpolation at leastOverlapped block motion compensation:-------------------------------------FIXMELL band prediction:===================Each sample in the LL0 subband is predicted by the median of the left, top andleft+top-topleft samples, samples outside the subband shall be considered tobe 0. To reverse this prediction in the decoder apply the following.for(y=0;y< height;y++){ for(x=0;x< width;x++){ sample[y][x]+=median(sample[y-1][x], sample[y][x-1], sample[y-1][x]+sample[y][x-1]-sample[y-1][x-1]);}}sample[-1][ *]=sample[ *][-1]=0;width, height here are the width and height of the LL0 subband not of the finalvideoDequantization:===============FIXMEWavelet Transform:==================Snow supports 2 wavelet transforms, the symmetric biorthogonal 5/3 integertransform and an integer approximation of the symmetric biorthogonal 9/7daubechies wavelet.2D IDWT(inverse discrete wavelet transform) --------------------------------------------The 2D IDWT applies a 2D filter recursively, each time combining the4 lowest frequency subbands into a single subband until only 1 subbandremains.The 2D filter is done by first applying a 1D filter in the vertical directionand then applying it in the horizontal one. --------------- --------------- --------------- ---------------|LL0|HL0|||||||||||||---+---|HL1||L0|H0|HL1||LL1|HL1|||||LH0|HH0|||||||||||||-------+-------|-> L1 H1 LH1 HH1 LH1 HH1 LH1 HH1 this can end with a L or a the number of elements shall be w s[-1] shall be considered equivalent to s[1] s[w] shall be considered equivalent to s[w-2] perform the lifting steps in order as described below Integer filter
static void sanitize(uint8_t *line)
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option keep it simple and lowercase description are in without and describe what they do
static int passed(HysteresisContext *s, int x, int y, int w)
static AVFrame * get_video_buffer(AVFilterLink *inlink, int w, int h)
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input Therefore
Undefined Behavior In the C some operations are undefined
static uint32_t BS_FUNC() read(BSCTX *bc, unsigned int n)
Return n bits from the buffer, n has to be in the 0-32 range.
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf foobar
static const uint8_t dither[8][8]