[FFmpeg-devel] [PATCH v2 14/26] avfilter/textmod: Add textmod, censor and show_speaker filters
ffmpegagent
ffmpegagent at gmail.com
Thu Jan 20 04:48:24 EET 2022
From: softworkz <softworkz at hotmail.com>
- textmod {S -> S)
Modify subtitle text in a number of ways
- censor {S -> S)
Censor subtitles using a word list
- show_speaker {S -> S)
Prepend speaker names from ASS subtitles to the visible text lines
Signed-off-by: softworkz <softworkz at hotmail.com>
---
doc/filters.texi | 206 ++++++++++++
libavfilter/Makefile | 5 +
libavfilter/allfilters.c | 3 +
libavfilter/sf_textmod.c | 710 +++++++++++++++++++++++++++++++++++++++
4 files changed, 924 insertions(+)
create mode 100644 libavfilter/sf_textmod.c
diff --git a/doc/filters.texi b/doc/filters.texi
index 51771f88e4..494ee6f062 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -25721,6 +25721,145 @@ existing filters using @code{--disable-filters}.
Below is a description of the currently available subtitle filters.
+
+ at section censor
+
+Censor selected words in text subtitles.
+
+Inputs:
+ at itemize
+ at item 0: Subtitles[TEXT]
+ at end itemize
+
+Outputs:
+ at itemize
+ at item 0: Subtitles[TEXT]
+ at end itemize
+
+It accepts the following parameters:
+
+ at table @option
+ at item mode
+The censoring mode to apply.
+
+Supported censoring modes are:
+
+ at table @var
+ at item 0, keep_first_last
+Replace all characters with the 'censor_char' except the first and the last character of a word.
+For words with less than 4 characters, the last character will be replaced as well.
+For words with less than 3 characters, the first character will be replaced as well.
+ at item 1, keep_first
+Replace all characters with the 'censor_char' except the first character of a word.
+For words with less than 3 characters, the first character will be replaced as well.
+ at item 2, all
+Replace all characters with the 'censor_char'.
+ at end table
+
+ at item words
+A list of words to censor, separated by 'separator'.
+
+ at item words_file
+Specify a file from which to load the contents for the 'words' parameter.
+
+ at item censor_char
+Single character used as replacement for censoring.
+
+ at item separator
+Delimiter character for words. Used with replace_words and remove_words- Must be a single character.
+The default is '.'.
+
+ at end table
+
+ at subsection Examples
+
+ at itemize
+ at item
+Censor a few given words with a pound character.
+ at example
+ffmpeg -i "http://streams.videolan.org/samples/sub/SSA/subtitle_testing_complex.mkv" -filter_complex "[0:1]censor=words='diss,louder,hope,beam,word':censor_char='#'" -map 0 -y output.mkv
+ at end example
+ at end itemize
+
+
+ at section textmod
+
+Modify subtitle text in a number of ways.
+
+Inputs:
+ at itemize
+ at item 0: Subtitles[TEXT]
+ at end itemize
+
+Outputs:
+ at itemize
+ at item 0: Subtitles[TEXT]
+ at end itemize
+
+It accepts the following parameters:
+
+ at table @option
+ at item mode
+The kind of text modification to apply
+
+Supported operation modes are:
+
+ at table @var
+ at item 0, leet
+Convert subtitle text to 'leet speak'. It's primarily useful for testing as the modification will be visible with almost all text lines.
+ at item 1, to_upper
+Change all text to upper case. Might improve readability.
+ at item 2, to_lower
+Change all text to lower case.
+ at item 3, replace_chars
+Replace one or more characters. Requires the find and replace parameters to be specified.
+Both need to be equal in length.
+The first char in find is replaced by the first char in replace, same for all subsequent chars.
+ at item 4, remove_chars
+Remove certain characters. Requires the find parameter to be specified.
+All chars in the find parameter string will be removed from all subtitle text.
+ at item 5, replace_words
+Replace one or more words. Requires the find and replace parameters to be specified. Multiple words must be separated by the delimiter char specified vie the separator parameter (default: ',').
+The number of words in the find and replace parameters needs to be equal.
+The first word in find is replaced by the first word in replace, same for all subsequent words
+ at item 6, remove_words
+Remove certain words. Requires the find parameter to be specified. Multiple words must be separated by the delimiter char specified vie the separator parameter (default: ',').
+All words in the find parameter string will be removed from all subtitle text.
+ at end table
+
+ at item find
+Required for replace_chars, remove_chars, replace_words and remove_words.
+
+ at item find_file
+Specify a file from which to load the contents for the 'find' parameter.
+
+ at item replace
+Required for replace_chars and replace_words.
+
+ at item replace_file
+Specify a file from which to load the contents for the 'replace' parameter.
+
+ at item separator
+Delimiter character for words. Used with replace_words and remove_words- Must be a single character.
+The default is '.'.
+
+ at end table
+
+ at subsection Examples
+
+ at itemize
+ at item
+Change all characters to upper case while keeping all styles and animations:
+ at example
+ffmpeg -i "https://streams.videolan.org/ffmpeg/mkv_subtitles.mkv" -filter_complex "[0:s]textmod=mode=to_upper" -map 0 -y out.mkv
+ at end example
+ at item
+Remove a set of symbol characters for am improved and smoother visual apperance:
+ at example
+ffmpeg -i "https://streams.videolan.org/ffmpeg/mkv_subtitles.mkv" -filter_complex "[0:s]textmod=mode=remove_chars:find='$&#@*§'" -map 0 -y out.mkv
+ at end example
+ at end itemize
+
@section graphicsub2video
Renders graphic subtitles as video frames.
@@ -25888,6 +26027,73 @@ ffmpeg -i "http://streams.videolan.org/samples/sub/SSA/subtitle_testing_complex.
@end example
@end itemize
+ at section showspeaker
+
+Prepend speaker names to subtitle lines (when available).
+
+Subtitles in ASS/SSA format are often including the names of the persons
+or character for each subtitle line. The showspeaker filter adds those names
+to the actual subtitle text to make it visible on playback.
+
+Inputs:
+ at itemize
+ at item 0: Subtitles[TEXT]
+ at end itemize
+
+Outputs:
+ at itemize
+ at item 0: Subtitles[TEXT]
+ at end itemize
+
+It accepts the following parameters:
+
+ at table @option
+ at item format
+The format for prepending speaker names. Default is 'square_brackets'.
+
+Supported operation modes are:
+
+ at table @var
+ at item 0, square_brackets
+Enclose the speaker name in square brackets, followed by space ('[speaker] text').
+ at item 1, round_brackets
+Enclose the speaker name in round brackets, followed by space ('(speaker) text').
+ at item 2, colon
+Separate the speaker name with a colon and space ('speaker: text').
+ at item 3, plain
+Separate the speaker name with a space only ('speaker text').
+ at end table
+
+ at item line_break
+Set thís parameter to insert a line break between speaker name and text instead of the space character.
+
+ at item style
+Allows to set a specific style for the speaker name text.
+
+This can be either a named style that exists in the ass subtitle header script (e.g. 'Default') or an ass style override code.
+Example: @{\\c&HDD0000&\\be1\\i1\\bord10@}
+This sets the color to blue, enables edge blurring, italic font and a border of size 10.
+
+The behavior is as follows:
+
+- When the style parameter is not provided, the filter will find the first position in the event string that is actual text.
+ The speaker name will be inserted at this position. This allows to have the speaker name shown in the same style like the
+ regular text, in case the string would start with a sequence of style codes.
+- When the style parameter is provided, everything will be prepended to the original text:
+ Style Code or Style name >> Speaker Name >> Style Reset Code >> Original Text
+
+ at end table
+
+ at subsection Examples
+
+ at itemize
+ at item
+Prepend speaker names with blue text, smooth edges and blend/overlay that onto the video.
+ at example
+ffmpeg -i INPUT -filter_complex "showspeaker=format=colon:style='@{\\c&HDD0000&\\be1@}',[0:v]overlay_textsubs"
+ at end example
+ at end itemize
+
@section textsub2video
Converts text subtitles to video frames.
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 48c9182c55..1af4f4b9bc 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -561,6 +561,11 @@ OBJS-$(CONFIG_YUVTESTSRC_FILTER) += vsrc_testsrc.o
OBJS-$(CONFIG_NULLSINK_FILTER) += vsink_nullsink.o
+# subtitle filters
+OBJS-$(CONFIG_CENSOR_FILTER) += sf_textmod.o
+OBJS-$(CONFIG_SHOW_SPEAKER_FILTER) += sf_textmod.o
+OBJS-$(CONFIG_TEXTMOD_FILTER) += sf_textmod.o
+
# multimedia filters
OBJS-$(CONFIG_ABITSCOPE_FILTER) += avf_abitscope.o
OBJS-$(CONFIG_ADRAWGRAPH_FILTER) += f_drawgraph.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 9ba89db128..ac7d71547b 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -547,6 +547,9 @@ extern const AVFilter ff_avf_showvolume;
extern const AVFilter ff_avf_showwaves;
extern const AVFilter ff_avf_showwavespic;
extern const AVFilter ff_vaf_spectrumsynth;
+extern const AVFilter ff_sf_censor;
+extern const AVFilter ff_sf_showspeaker;
+extern const AVFilter ff_sf_textmod;
extern const AVFilter ff_svf_graphicsub2video;
extern const AVFilter ff_svf_textsub2video;
diff --git a/libavfilter/sf_textmod.c b/libavfilter/sf_textmod.c
new file mode 100644
index 0000000000..c75244cfb2
--- /dev/null
+++ b/libavfilter/sf_textmod.c
@@ -0,0 +1,710 @@
+/*
+ * Copyright (c) 2021 softworkz
+ *
+ * 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
+ * text subtitle filter which allows to modify subtitle text in several ways
+ */
+
+#include <libavutil/ass_internal.h>
+
+#include "libavutil/opt.h"
+#include "internal.h"
+#include "libavutil/ass_split_internal.h"
+#include "libavutil/bprint.h"
+#include "libavutil/file.h"
+
+static const char* leet_src = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ";
+static const char* leet_dst = "abcd3f6#1jklmn0pq257uvwxyzAB(D3F6#1JKLMN0PQ257UVWXYZ";
+
+enum TextModFilterType {
+ TM_TEXTMOD,
+ TM_CENSOR,
+ TM_SHOW_SPEAKER,
+};
+
+enum TextModOperation {
+ OP_LEET,
+ OP_TO_UPPER,
+ OP_TO_LOWER,
+ OP_REPLACE_CHARS,
+ OP_REMOVE_CHARS,
+ OP_REPLACE_WORDS,
+ OP_REMOVE_WORDS,
+ NB_OPS,
+};
+
+enum CensorMode {
+ CM_KEEP_FIRST_LAST,
+ CM_KEEP_FIRST,
+ CM_ALL,
+};
+
+enum ShowSpeakerMode {
+ SM_SQUARE_BRACKETS,
+ SM_ROUND_BRACKETS,
+ SM_COLON,
+ SM_PLAIN,
+};
+
+typedef struct TextModContext {
+ const AVClass *class;
+ enum AVSubtitleType format;
+ enum TextModFilterType filter_type;
+ enum TextModOperation operation;
+ enum CensorMode censor_mode;
+ enum ShowSpeakerMode speaker_mode;
+ char *find;
+ char *find_file;
+ char *style;
+ char *replace;
+ char *replace_file;
+ char *separator;
+ char *censor_char;
+ char **find_list;
+ int line_break;
+ int nb_find_list;
+ char **replace_list;
+ int nb_replace_list;
+} TextModContext;
+
+static char **split_string(char *source, int *nb_elems, const char *delim)
+{
+ char **list = NULL;
+ char *temp = NULL;
+ char *ptr = av_strtok(source, delim, &temp);
+
+ while (ptr) {
+ if (strlen(ptr)) {
+ av_dynarray_add(&list, nb_elems, ptr);
+ if (!list)
+ return NULL;
+ }
+
+ ptr = av_strtok(NULL, delim, &temp);
+ }
+
+ if (!list)
+ return NULL;
+
+ for (int i = 0; i < *nb_elems; i++) {
+ list[i] = av_strdup(list[i]);
+ if (!list[i]) {
+ for (int n = 0; n < i; n++)
+ av_free(list[n]);
+ av_free(list);
+ return NULL;
+ }
+ }
+
+ return list;
+}
+
+static int load_text_from_file(AVFilterContext *ctx, const char *file_name, char **text, char separator)
+{
+ int err;
+ uint8_t *textbuf;
+ char *tmp;
+ size_t textbuf_size;
+ int offset = 0;
+
+ if ((err = av_file_map(file_name, &textbuf, &textbuf_size, 0, ctx)) < 0) {
+ av_log(ctx, AV_LOG_ERROR, "The text file '%s' could not be read or is empty\n", file_name);
+ return err;
+ }
+
+ if (textbuf_size > 1 &&
+ (textbuf[0] == 0xFF && textbuf[1] == 0xFE
+ || textbuf[0] == 0xFE && textbuf[1] == 0xFF)) {
+ av_log(ctx, AV_LOG_ERROR, "UTF text files are not supported. File: %s\n", file_name);
+ return AVERROR(EINVAL);
+ }
+
+ if (textbuf_size > 2 && textbuf[0] == 0xEF && textbuf[1] == 0xBB && textbuf[2] == 0xBF)
+ offset = 3; // UTF-8
+
+ if (textbuf_size > SIZE_MAX - 1 || !((tmp = av_strndup((char *)textbuf + offset, textbuf_size - offset)))) {
+ av_file_unmap(textbuf, textbuf_size);
+ return AVERROR(ENOMEM);
+ }
+
+ av_file_unmap(textbuf, textbuf_size);
+
+ for (size_t i = 0; i < strlen(tmp); i++) {
+ switch (tmp[i]) {
+ case '\n':
+ case '\r':
+ case '\f':
+ case '\v':
+ tmp[i] = separator;
+ }
+ }
+
+ *text = tmp;
+
+ return 0;
+}
+
+static int load_files(AVFilterContext *ctx)
+{
+ TextModContext *s = ctx->priv;
+ int ret;
+
+ if (!s->separator || strlen(s->separator) != 1) {
+ av_log(ctx, AV_LOG_ERROR, "A single character needs to be specified for the separator parameter.\n");
+ return AVERROR(EINVAL);
+ }
+
+ if (s->find_file && strlen(s->find_file)) {
+ ret = load_text_from_file(ctx, s->find_file, &s->find, s->separator[0]);
+ if (ret < 0 )
+ return ret;
+ }
+
+ if (s->replace_file && strlen(s->replace_file)) {
+ ret = load_text_from_file(ctx, s->replace_file, &s->replace, s->separator[0]);
+ if (ret < 0 )
+ return ret;
+ }
+
+ return 0;
+}
+
+static int init_censor(AVFilterContext *ctx)
+{
+ TextModContext *s = ctx->priv;
+ int ret;
+
+ s->filter_type = TM_CENSOR;
+ s->operation = OP_REPLACE_WORDS;
+
+ ret = load_files(ctx);
+ if (ret < 0 )
+ return ret;
+
+ if (!s->find || !strlen(s->find)) {
+ av_log(ctx, AV_LOG_ERROR, "Either the 'words' or the 'words_file' parameter needs to be specified\n");
+ return AVERROR(EINVAL);
+ }
+
+ if (!s->censor_char || strlen(s->censor_char) != 1) {
+ av_log(ctx, AV_LOG_ERROR, "A single character needs to be specified for the censor_char parameter\n");
+ return AVERROR(EINVAL);
+ }
+
+ s->find_list = split_string(s->find, &s->nb_find_list, s->separator);
+ if (!s->find_list)
+ return AVERROR(ENOMEM);
+
+ s->replace_list = av_calloc(s->nb_find_list, sizeof(char *));
+ if (!s->replace_list)
+ return AVERROR(ENOMEM);
+
+ for (int i = 0; i < s->nb_find_list; i++) {
+ size_t len, start = 0, end;
+ char *item = av_strdup(s->find_list[i]);
+ if (!item)
+ return AVERROR(ENOMEM);
+
+ len = end = strlen(item);
+
+ switch (s->censor_mode) {
+ case CM_KEEP_FIRST_LAST:
+
+ if (len > 2)
+ start = 1;
+ if (len > 3)
+ end--;
+
+ break;
+ case CM_KEEP_FIRST:
+
+ if (len > 2)
+ start = 1;
+
+ break;
+ }
+
+ for (size_t n = start; n < end; n++)
+ item[n] = s->censor_char[0];
+
+ s->replace_list[i] = item;
+ }
+
+ return 0;
+}
+
+static int init_showspeaker(AVFilterContext *ctx)
+{
+ TextModContext *s = ctx->priv;
+ s->filter_type = TM_SHOW_SPEAKER;
+
+ return 0;
+}
+
+static int init(AVFilterContext *ctx)
+{
+ TextModContext *s = ctx->priv;
+ int ret;
+
+ ret = load_files(ctx);
+ if (ret < 0 )
+ return ret;
+
+ switch (s->operation) {
+ case OP_REPLACE_CHARS:
+ case OP_REMOVE_CHARS:
+ case OP_REPLACE_WORDS:
+ case OP_REMOVE_WORDS:
+ if (!s->find || !strlen(s->find)) {
+ av_log(ctx, AV_LOG_ERROR, "Selected mode requires the 'find' parameter to be specified\n");
+ return AVERROR(EINVAL);
+ }
+ break;
+ }
+
+ switch (s->operation) {
+ case OP_REPLACE_CHARS:
+ case OP_REPLACE_WORDS:
+ if (!s->replace || !strlen(s->replace)) {
+ av_log(ctx, AV_LOG_ERROR, "Selected mode requires the 'replace' parameter to be specified\n");
+ return AVERROR(EINVAL);
+ }
+ break;
+ }
+
+ if (s->operation == OP_REPLACE_CHARS && strlen(s->find) != strlen(s->replace)) {
+ av_log(ctx, AV_LOG_ERROR, "Selected mode requires the 'find' and 'replace' parameters to have the same length\n");
+ return AVERROR(EINVAL);
+ }
+
+ if (s->operation == OP_REPLACE_WORDS || s->operation == OP_REMOVE_WORDS) {
+ if (!s->separator || strlen(s->separator) != 1) {
+ av_log(ctx, AV_LOG_ERROR, "Selected mode requires a single separator char to be specified\n");
+ return AVERROR(EINVAL);
+ }
+
+ s->find_list = split_string(s->find, &s->nb_find_list, s->separator);
+ if (!s->find_list)
+ return AVERROR(ENOMEM);
+
+ if (s->operation == OP_REPLACE_WORDS) {
+
+ s->replace_list = split_string(s->replace, &s->nb_replace_list, s->separator);
+ if (!s->replace_list)
+ return AVERROR(ENOMEM);
+
+ if (s->nb_find_list != s->nb_replace_list) {
+ av_log(ctx, AV_LOG_ERROR, "The number of words in 'find' and 'replace' needs to be equal\n");
+ return AVERROR(EINVAL);
+ }
+ }
+ }
+
+ return 0;
+}
+
+static void uninit(AVFilterContext *ctx)
+{
+ TextModContext *s = ctx->priv;
+
+ for (int i = 0; i < s->nb_find_list; i++)
+ av_freep(&s->find_list[i]);
+
+ s->nb_find_list = 0;
+ av_freep(&s->find_list);
+
+ for (int i = 0; i < s->nb_replace_list; i++)
+ av_freep(&s->replace_list[i]);
+
+ s->nb_replace_list = 0;
+ av_freep(&s->replace_list);
+}
+
+static char *process_text(TextModContext *s, char *text)
+{
+ const char *char_src = s->find;
+ const char *char_dst = s->replace;
+ char *result = NULL;
+ int escape_level = 0, k = 0;
+
+ switch (s->operation) {
+ case OP_LEET:
+ case OP_REPLACE_CHARS:
+
+ if (s->operation == OP_LEET) {
+ char_src = leet_src;
+ char_dst = leet_dst;
+ }
+
+ result = av_strdup(text);
+ if (!result)
+ return NULL;
+
+ for (size_t n = 0; n < strlen(result); n++) {
+ if (result[n] == '{')
+ escape_level++;
+
+ if (!escape_level) {
+ size_t len = strlen(char_src);
+ for (size_t t = 0; t < len; t++) {
+ if (result[n] == char_src[t]) {
+ result[n] = char_dst[t];
+ break;
+ }
+ }
+ }
+
+ if (result[n] == '}')
+ escape_level--;
+ }
+
+ break;
+ case OP_TO_UPPER:
+ case OP_TO_LOWER:
+
+ result = av_strdup(text);
+ if (!result)
+ return NULL;
+
+ for (size_t n = 0; n < strlen(result); n++) {
+ if (result[n] == '{')
+ escape_level++;
+ if (!escape_level)
+ result[n] = s->operation == OP_TO_LOWER ? av_tolower(result[n]) : av_toupper(result[n]);
+ if (result[n] == '}')
+ escape_level--;
+ }
+
+ break;
+ case OP_REMOVE_CHARS:
+
+ result = av_strdup(text);
+ if (!result)
+ return NULL;
+
+ for (size_t n = 0; n < strlen(result); n++) {
+ int skip_char = 0;
+
+ if (result[n] == '{')
+ escape_level++;
+
+ if (!escape_level) {
+ size_t len = strlen(char_src);
+ for (size_t t = 0; t < len; t++) {
+ if (result[n] == char_src[t]) {
+ skip_char = 1;
+ break;
+ }
+ }
+ }
+
+ if (!skip_char)
+ result[k++] = result[n];
+
+ if (result[n] == '}')
+ escape_level--;
+ }
+
+ result[k] = 0;
+
+ break;
+ case OP_REPLACE_WORDS:
+ case OP_REMOVE_WORDS:
+
+ result = av_strdup(text);
+ if (!result)
+ return NULL;
+
+ for (int n = 0; n < s->nb_find_list; n++) {
+ char *tmp = result;
+ const char *replace = (s->operation == OP_REPLACE_WORDS) ? s->replace_list[n] : "";
+
+ result = av_strireplace(result, s->find_list[n], replace);
+ if (!result)
+ return NULL;
+
+ av_free(tmp);
+ }
+
+ break;
+ }
+
+ return result;
+}
+
+static char *process_dialog_show_speaker(TextModContext *s, char *ass_line)
+{
+ ASSDialog *dialog = avpriv_ass_split_dialog(NULL, ass_line);
+ int escape_level = 0;
+ unsigned pos = 0, len;
+ char *result, *text;
+ AVBPrint pbuf;
+
+ if (!dialog)
+ return NULL;
+
+ text = process_text(s, dialog->text);
+ if (!text)
+ return NULL;
+
+ if (!dialog->name || !strlen(dialog->name) || !dialog->text || !strlen(dialog->text))
+ return av_strdup(ass_line);
+
+ // Find insertion point in case the line starts with style codes
+ len = (unsigned)strlen(dialog->text);
+ for (unsigned i = 0; i < len; i++) {
+
+ if (dialog->text[i] == '{')
+ escape_level++;
+
+ if (dialog->text[i] == '}')
+ escape_level--;
+
+ if (escape_level == 0) {
+ pos = i;
+ break;
+ }
+ }
+
+ if (s->style && strlen(s->style))
+ // When a style is specified reset the insertion point
+ // (always add speaker plus style at the start in that case)
+ pos = 0;
+
+ if (pos >= len - 1)
+ return av_strdup(ass_line);
+
+ av_bprint_init(&pbuf, 1, AV_BPRINT_SIZE_UNLIMITED);
+
+ if (pos > 0) {
+ av_bprint_append_data(&pbuf, dialog->text, pos);
+ }
+
+ if (s->style && strlen(s->style)) {
+ if (s->style[0] == '{')
+ // Assume complete and valid style code, e.g. {\c&HFF0000&}
+ av_bprintf(&pbuf, "%s", s->style);
+ else
+ // Otherwise it must be a style name
+ av_bprintf(&pbuf, "{\\r%s}", s->style);
+ }
+
+ switch (s->speaker_mode) {
+ case SM_SQUARE_BRACKETS:
+ av_bprintf(&pbuf, "[%s]", dialog->name);
+ break;
+ case SM_ROUND_BRACKETS:
+ av_bprintf(&pbuf, "(%s)", dialog->name);
+ break;
+ case SM_COLON:
+ av_bprintf(&pbuf, "%s:", dialog->name);
+ break;
+ case SM_PLAIN:
+ av_bprintf(&pbuf, "%s", dialog->name);
+ break;
+ }
+
+ if (s->style && strlen(s->style)) {
+ // Reset line style
+ if (dialog->style && strlen(dialog->style) && !av_strcasecmp(dialog->style, "default"))
+ av_bprintf(&pbuf, "{\\r%s}", dialog->style);
+ else
+ av_bprintf(&pbuf, "{\\r}");
+ }
+
+ if (s->line_break)
+ av_bprintf(&pbuf, "\\N");
+ else
+ av_bprintf(&pbuf, " ");
+
+ av_bprint_append_data(&pbuf, dialog->text + pos, len - pos);
+
+ av_bprint_finalize(&pbuf, &text);
+
+ result = avpriv_ass_get_dialog_ex(dialog->readorder, dialog->layer, dialog->style, dialog->name, dialog->margin_l, dialog->margin_r, dialog->margin_v, text);
+
+ av_free(text);
+ avpriv_ass_free_dialog(&dialog);
+ return result;
+}
+
+static char *process_dialog(TextModContext *s, char *ass_line)
+{
+ ASSDialog *dialog;
+ char *result, *text;
+
+ if (s->filter_type == TM_SHOW_SPEAKER)
+ return process_dialog_show_speaker(s, ass_line);
+
+ dialog = avpriv_ass_split_dialog(NULL, ass_line);
+ if (!dialog)
+ return NULL;
+
+ text = process_text(s, dialog->text);
+ if (!text)
+ return NULL;
+
+ result = avpriv_ass_get_dialog_ex(dialog->readorder, dialog->layer, dialog->style, dialog->name, dialog->margin_l, dialog->margin_r, dialog->margin_v, text);
+
+ av_free(text);
+ avpriv_ass_free_dialog(&dialog);
+ return result;
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+ AVFilterLink *inlink = outlink->src->inputs[0];
+
+ outlink->w = inlink->w;
+ outlink->h = inlink->h;
+ outlink->time_base = inlink->time_base;
+ outlink->frame_rate = inlink->frame_rate;
+
+ return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *frame)
+{
+ TextModContext *s = inlink->dst->priv;
+ AVFilterLink *outlink = inlink->dst->outputs[0];
+ int ret;
+
+ outlink->format = inlink->format;
+
+ ret = av_frame_make_writable(frame);
+ if (ret < 0)
+ return ret;
+
+ for (unsigned i = 0; i < frame->num_subtitle_areas; i++) {
+
+ AVSubtitleArea *area = frame->subtitle_areas[i];
+
+ if (area->ass) {
+ char *tmp = area->ass;
+ area->ass = process_dialog(s, area->ass);
+ av_free(tmp);
+ if (!area->ass)
+ return AVERROR(ENOMEM);
+ }
+ }
+
+ return ff_filter_frame(outlink, frame);
+}
+
+#define OFFSET(x) offsetof(TextModContext, x)
+#define FLAGS (AV_OPT_FLAG_SUBTITLE_PARAM | AV_OPT_FLAG_FILTERING_PARAM)
+
+static const AVOption textmod_options[] = {
+ { "mode", "set operation mode", OFFSET(operation), AV_OPT_TYPE_INT, {.i64=OP_LEET}, OP_LEET, NB_OPS-1, FLAGS, "mode" },
+ { "leet", "convert text to 'leet speak'", 0, AV_OPT_TYPE_CONST, {.i64=OP_LEET}, 0, 0, FLAGS, "mode" },
+ { "to_upper", "change to upper case", 0, AV_OPT_TYPE_CONST, {.i64=OP_TO_UPPER}, 0, 0, FLAGS, "mode" },
+ { "to_lower", "change to lower case", 0, AV_OPT_TYPE_CONST, {.i64=OP_TO_LOWER}, 0, 0, FLAGS, "mode" },
+ { "replace_chars", "replace characters", 0, AV_OPT_TYPE_CONST, {.i64=OP_REPLACE_CHARS}, 0, 0, FLAGS, "mode" },
+ { "remove_chars", "remove characters", 0, AV_OPT_TYPE_CONST, {.i64=OP_REMOVE_CHARS}, 0, 0, FLAGS, "mode" },
+ { "replace_words", "replace words", 0, AV_OPT_TYPE_CONST, {.i64=OP_REPLACE_WORDS}, 0, 0, FLAGS, "mode" },
+ { "remove_words", "remove words", 0, AV_OPT_TYPE_CONST, {.i64=OP_REMOVE_WORDS}, 0, 0, FLAGS, "mode" },
+ { "find", "chars/words to find or remove", OFFSET(find), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { "find_file", "load find param from file", OFFSET(find_file), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { "replace", "chars/words to replace", OFFSET(replace), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { "replace_file", "load replace param from file", OFFSET(replace_file), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { "separator", "word separator", OFFSET(separator), AV_OPT_TYPE_STRING, {.str = ","}, 0, 0, FLAGS, NULL },
+ { NULL },
+};
+
+
+static const AVOption censor_options[] = {
+ { "mode", "set censoring mode", OFFSET(censor_mode), AV_OPT_TYPE_INT, {.i64=CM_KEEP_FIRST_LAST}, 0, 2, FLAGS, "mode" },
+ { "keep_first_last", "censor inner chars", 0, AV_OPT_TYPE_CONST, {.i64=CM_KEEP_FIRST_LAST}, 0, 0, FLAGS, "mode" },
+ { "keep_first", "censor all but first char", 0, AV_OPT_TYPE_CONST, {.i64=CM_KEEP_FIRST}, 0, 0, FLAGS, "mode" },
+ { "all", "censor all chars", 0, AV_OPT_TYPE_CONST, {.i64=CM_ALL}, 0, 0, FLAGS, "mode" },
+ { "words", "list of words to censor", OFFSET(find), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { "words_file", "path to word list file", OFFSET(find_file), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { "separator", "word separator", OFFSET(separator), AV_OPT_TYPE_STRING, {.str = ","}, 0, 0, FLAGS, NULL },
+ { "censor_char", "replacement character", OFFSET(censor_char), AV_OPT_TYPE_STRING, {.str = "*"}, 0, 0, FLAGS, NULL },
+ { NULL },
+};
+
+static const AVOption showspeaker_options[] = {
+ { "format", "speaker name formatting", OFFSET(speaker_mode), AV_OPT_TYPE_INT, {.i64=SM_SQUARE_BRACKETS}, 0, 2, FLAGS, "format" },
+ { "square_brackets", "[speaker] text", 0, AV_OPT_TYPE_CONST, {.i64=SM_SQUARE_BRACKETS}, 0, 0, FLAGS, "format" },
+ { "round_brackets", "(speaker) text", 0, AV_OPT_TYPE_CONST, {.i64=SM_ROUND_BRACKETS}, 0, 0, FLAGS, "format" },
+ { "colon", "speaker: text", 0, AV_OPT_TYPE_CONST, {.i64=SM_COLON}, 0, 0, FLAGS, "format" },
+ { "plain", "speaker text", 0, AV_OPT_TYPE_CONST, {.i64=SM_PLAIN}, 0, 0, FLAGS, "format" },
+ { "line_break", "insert line break", OFFSET(line_break), AV_OPT_TYPE_BOOL, {.i64=0}, 0, 1, FLAGS, NULL },
+ { "style", "ass type name or style code", OFFSET(style), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, FLAGS, NULL },
+ { NULL },
+};
+
+AVFILTER_DEFINE_CLASS(textmod);
+AVFILTER_DEFINE_CLASS(censor);
+AVFILTER_DEFINE_CLASS(showspeaker);
+
+static const AVFilterPad inputs[] = {
+ {
+ .name = "default",
+ .type = AVMEDIA_TYPE_SUBTITLE,
+ .filter_frame = filter_frame,
+ },
+};
+
+static const AVFilterPad outputs[] = {
+ {
+ .name = "default",
+ .type = AVMEDIA_TYPE_SUBTITLE,
+ .config_props = config_output,
+ },
+};
+
+const AVFilter ff_sf_textmod = {
+ .name = "textmod",
+ .description = NULL_IF_CONFIG_SMALL("Modify subtitle text in several ways"),
+ .init = init,
+ .uninit = uninit,
+ .priv_size = sizeof(TextModContext),
+ .priv_class = &textmod_class,
+ FILTER_INPUTS(inputs),
+ FILTER_OUTPUTS(outputs),
+ FILTER_SINGLE_SUBFMT(AV_SUBTITLE_FMT_ASS),
+};
+
+const AVFilter ff_sf_censor = {
+ .name = "censor",
+ .description = NULL_IF_CONFIG_SMALL("Censor words in subtitle text"),
+ .init = init_censor,
+ .uninit = uninit,
+ .priv_size = sizeof(TextModContext),
+ .priv_class = &censor_class,
+ FILTER_INPUTS(inputs),
+ FILTER_OUTPUTS(outputs),
+ FILTER_SINGLE_SUBFMT(AV_SUBTITLE_FMT_ASS),
+};
+
+const AVFilter ff_sf_showspeaker = {
+ .name = "showspeaker",
+ .description = NULL_IF_CONFIG_SMALL("Prepend speaker names to text subtitles (when available)"),
+ .init = init_showspeaker,
+ .uninit = uninit,
+ .priv_size = sizeof(TextModContext),
+ .priv_class = &showspeaker_class,
+ FILTER_INPUTS(inputs),
+ FILTER_OUTPUTS(outputs),
+ FILTER_SINGLE_SUBFMT(AV_SUBTITLE_FMT_ASS),
+};
--
ffmpeg-codebot
More information about the ffmpeg-devel
mailing list