Merge remote-tracking branch 'remotes/origin/unify-roi-features'

README.md

@@ -156,11 +156,20 @@ Video structure:
                                    - frametile: Constrain within the tile.
                                    - frametilemargin: Constrain even more.
       --roi <filename>       : Use a delta QP map for region of interest.
-                               Reads an array of delta QP values from a text
-                               file. The file format is: width and height of
-                               the QP delta map followed by width*height delta
-                               QP values in raster order. The map can be of any
-                               size and will be scaled to the video size.
+                               Reads an array of delta QP values from a file.
+                               Text and binary files are supported and detected
+                               from the file extension (.txt/.bin). If a known
+                               extension is not found, the file is treated as
+                               a text file. The file can include one or many
+                               ROI frames each in the following format:
+                               width and height of the QP delta map followed
+                               by width * height delta QP values in raster
+                               order. In binary format, width and height are
+                               32-bit integers whereas the delta QP values are
+                               signed 8-bit values. The map can be of any size
+                               and will be scaled to the video size. The file
+                               reading will loop if end of the file is reached.
+                               See roi.txt in the examples folder.
       --set-qp-in-cu         : Set QP at CU level keeping pic_init_qp_minus26.
                                in PPS and slice_qp_delta in slize header zero.
       --(no-)erp-aqp         : Use adaptive QP for 360 degree video with

configure.ac

@@ -22,8 +22,8 @@ AC_CONFIG_SRCDIR([src/encmain.c])
 #   - Increment when making new releases and major or minor was not changed since last release.
 #
 # Here is a somewhat sane guide to lib versioning: http://apr.apache.org/versioning.html
-ver_major=6
-ver_minor=6
+ver_major=7
+ver_minor=0
 ver_release=0
 
 # Prevents configure from adding a lot of defines to the CFLAGS

doc/kvazaar.1

@@ -1,4 +1,4 @@
-.TH KVAZAAR "1" "October 2021" "kvazaar v2.1.0" "User Commands"
+.TH KVAZAAR "1" "February 2022" "kvazaar v2.1.0" "User Commands"
 .SH NAME
 kvazaar \- open source HEVC encoder
 .SH SYNOPSIS
@@ -180,11 +180,20 @@ Constrain movement vectors. [none]
 .TP
 \fB\-\-roi <filename>      
 Use a delta QP map for region of interest.
-Reads an array of delta QP values from a text
-file. The file format is: width and height of
-the QP delta map followed by width*height delta
-QP values in raster order. The map can be of any
-size and will be scaled to the video size.
+Reads an array of delta QP values from a file.
+Text and binary files are supported and detected
+from the file extension (.txt/.bin). If a known
+extension is not found, the file is treated as
+a text file. The file can include one or many
+ROI frames each in the following format:
+width and height of the QP delta map followed
+by width * height delta QP values in raster
+order. In binary format, width and height are
+32\-bit integers whereas the delta QP values are
+signed 8\-bit values. The map can be of any size
+and will be scaled to the video size. The file
+reading will loop if end of the file is reached.
+See roi.txt in the examples folder.
 .TP
 \fB\-\-set\-qp\-in\-cu        
 Set QP at CU level keeping pic_init_qp_minus26.

src/cfg.c

@@ -139,9 +139,9 @@ int kvz_config_init(kvz_config *cfg)
   cfg->gop_lp_definition.t = 1;
   cfg->open_gop = true;
 
-  cfg->roi.width = 0;
-  cfg->roi.height = 0;
-  cfg->roi.dqps = NULL;
+  cfg->roi.file_path = NULL;
+  cfg->roi.format = KVZ_ROI_TXT;
+
   cfg->set_qp_in_cu = false;
 
   cfg->erp_aqp = false;
@@ -190,11 +190,11 @@ int kvz_config_destroy(kvz_config *cfg)
 {
   if (cfg) {
     FREE_POINTER(cfg->cqmfile);
+    FREE_POINTER(cfg->roi.file_path);
     FREE_POINTER(cfg->fast_coeff_table_fn);
     FREE_POINTER(cfg->tiles_width_split);
     FREE_POINTER(cfg->tiles_height_split);
     FREE_POINTER(cfg->slice_addresses_in_ts);
-    FREE_POINTER(cfg->roi.dqps);
     FREE_POINTER(cfg->optional_key);
     FREE_POINTER(cfg->fastrd_learning_outdir_fn);
   }
@@ -1241,60 +1241,29 @@ int kvz_config_parse(kvz_config *cfg, const char *name, const char *value)
   }
   else if OPT("implicit-rdpcm")
     cfg->implicit_rdpcm = (bool)atobool(value);
-  else if OPT("roi") {
-    // The ROI description is as follows:
-    // First number is width, second number is height,
-    // then follows width * height number of dqp values.
-    FILE* f = fopen(value, "rb");
-    if (!f) {
-      fprintf(stderr, "Could not open ROI file.\n");
-      return 0;
-    }
-
-    int width = 0;
-    int height = 0;
-    if (!fscanf(f, "%d", &width) || !fscanf(f, "%d", &height)) {
-      fprintf(stderr, "Failed to read ROI size.\n");
-      fclose(f);
-      return 0;
-    }
-
-    if (width <= 0 || height <= 0) {
-      fprintf(stderr, "Invalid ROI size: %dx%d.\n", width, height);
-      fclose(f);
-      return 0;
-    }
 
-    if (width > 10000 || height > 10000) {
-      fprintf(stderr, "ROI dimensions exceed arbitrary value of 10000.\n");
-      fclose(f);
-      return 0;
-    }
+  else if OPT("roi") {
+    static enum kvz_roi_format const formats[] = { KVZ_ROI_TXT, KVZ_ROI_BIN };
+    static const char * const format_names[] = { "txt", "bin", NULL };
 
-    const unsigned size = width * height;
-    int8_t *dqp_array  = calloc((size_t)size, sizeof(cfg->roi.dqps[0]));
-    if (!dqp_array) {
-      fprintf(stderr, "Failed to allocate memory for ROI table.\n");
-      fclose(f);
+    char *roi_file = strdup(value);
+    if (!roi_file) {
+      fprintf(stderr, "Failed to allocate memory for ROI file name.\n");
       return 0;
     }
+    FREE_POINTER(cfg->roi.file_path);
+    cfg->roi.file_path = roi_file;
 
-    FREE_POINTER(cfg->roi.dqps);
-    cfg->roi.dqps   = dqp_array;
-    cfg->roi.width  = width;
-    cfg->roi.height = height;
-
-    for (int i = 0; i < size; ++i) {
-      int number; // Need a pointer to int for fscanf
-      if (fscanf(f, "%d", &number) != 1) {
-        fprintf(stderr, "Reading ROI file failed.\n");
-        fclose(f);
-        return 0;
-      }
-      dqp_array[i] = CLIP(-51, 51, number);
+    // Get file extension or the substring after the last dot
+    char *maybe_extension = strrchr(cfg->roi.file_path, '.');
+    if (!maybe_extension) {
+      cfg->roi.format = KVZ_ROI_TXT;
+    } else {
+      maybe_extension++;
+      int8_t format;
+      bool unknown_format = !parse_enum(maybe_extension, format_names, &format);
+      cfg->roi.format = unknown_format ? KVZ_ROI_TXT : formats[format];
     }
-
-    fclose(f);
   }
   else if OPT("set-qp-in-cu") {
     cfg->set_qp_in_cu = (bool)atobool(value);

src/cli.c

@@ -141,6 +141,7 @@ static const struct option long_options[] = {
   { "force-level",        required_argument, NULL, 0 },
   { "high-tier",                no_argument, NULL, 0 },
   { "me-steps",           required_argument, NULL, 0 },
+  { "roi-file",           required_argument, NULL, 0 },
   { "fast-residual-cost", required_argument, NULL, 0 },
   { "set-qp-in-cu",             no_argument, NULL, 0 },
   { "open-gop",                 no_argument, NULL, 0 },
@@ -500,11 +501,20 @@ void print_help(void)
     "                                   - frametile: Constrain within the tile.\n"
     "                                   - frametilemargin: Constrain even more.\n"
     "      --roi <filename>       : Use a delta QP map for region of interest.\n"
-    "                               Reads an array of delta QP values from a text\n"
-    "                               file. The file format is: width and height of\n"
-    "                               the QP delta map followed by width*height delta\n"
-    "                               QP values in raster order. The map can be of any\n"
-    "                               size and will be scaled to the video size.\n"
+    "                               Reads an array of delta QP values from a file.\n"
+    "                               Text and binary files are supported and detected\n"
+    "                               from the file extension (.txt/.bin). If a known\n"
+    "                               extension is not found, the file is treated as\n"
+    "                               a text file. The file can include one or many\n"
+    "                               ROI frames each in the following format:\n"
+    "                               width and height of the QP delta map followed\n"
+    "                               by width * height delta QP values in raster\n"
+    "                               order. In binary format, width and height are\n"
+    "                               32-bit integers whereas the delta QP values are\n"
+    "                               signed 8-bit values. The map can be of any size\n"
+    "                               and will be scaled to the video size. The file\n"
+    "                               reading will loop if end of the file is reached.\n"
+    "                               See roi.txt in the examples folder.\n"
     "      --set-qp-in-cu         : Set QP at CU level keeping pic_init_qp_minus26.\n"
     "                               in PPS and slice_qp_delta in slize header zero.\n"
     "      --(no-)erp-aqp         : Use adaptive QP for 360 degree video with\n"

src/encmain.c

@@ -427,6 +427,7 @@ int main(int argc, char *argv[])
   FILE *input  = NULL; //!< input file (YUV)
   FILE *output = NULL; //!< output file (HEVC NAL stream)
   FILE *recout = NULL; //!< reconstructed YUV output, --debug
+  FILE *roifile = NULL;
   clock_t start_time = clock();
   clock_t encoding_start_cpu_time;
   KVZ_CLOCK_T encoding_start_real_time;
@@ -566,7 +567,7 @@ int main(int argc, char *argv[])
     // Give arguments via struct to the input thread
     input_handler_args in_args = {
       .available_input_slots = available_input_slots,
-      .filled_input_slots    = filled_input_slots,
+      .filled_input_slots = filled_input_slots,
 
       .input = input,
       .api = api,
@@ -805,6 +806,7 @@ int main(int argc, char *argv[])
   if (input)  fclose(input);
   if (output) fclose(output);
   if (recout) fclose(recout);
+  if (roifile) fclose(roifile);
 
   CHECKPOINTS_FINALIZE();
 

src/encode_coding_tree.c

@@ -758,7 +758,7 @@ void kvz_encode_coding_tree(encoder_state_t * const state,
   bool border_split_y = ctrl->in.height >= abs_y + (LCU_WIDTH >> MAX_DEPTH) + half_cu;
   bool border = border_x || border_y; /*!< are we in any border CU */
 
-  if (depth <= ctrl->max_qp_delta_depth) {
+  if (depth <= state->frame->max_qp_delta_depth) {
     state->must_code_qp_delta = true;
   }