shithub: libvpx

Info  •  Files  •  Log  •  Branches

Support 16K sequence coding

Fixed a couple of variable/function definitions, as well as header
handling to support 16K sequence coding at high bit-rates.

The width and height are each specified by two bytes in the header.
Use an extra byte to explicitly indicate the scaling factors in
both directions, each ranging from 0 to 15.

Tested coding up to 16400x16400 dimension.

Change-Id: Ibc2225c6036620270f2c0cf5172d1760aaec10ec

--- a/vp9/common/vp9_onyx.h

+++ b/vp9/common/vp9_onyx.h

@@ -94,7 +94,7 @@

     int Width;              // width of data passed to the compressor

     int Height;             // height of data passed to the compressor

     double frame_rate;       // set to passed in framerate

-    int target_bandwidth;    // bandwidth to be used in kilobits per second

+    int64_t target_bandwidth;    // bandwidth to be used in kilobits per second

     int noise_sensitivity;   // parameter used for applying pre processing blur: recommendation 0

     int Sharpness;          // parameter used for sharpening output: recommendation 0:

@@ -135,9 +135,9 @@

     int over_shoot_pct;

     // buffering parameters

-    int starting_buffer_level;  // in seconds

-    int optimal_buffer_level;

-    int maximum_buffer_size;

+    int64_t starting_buffer_level;  // in seconds

+    int64_t optimal_buffer_level;

+    int64_t maximum_buffer_size;

     // controlling quality

     int fixed_q;

--- a/vp9/decoder/vp9_decodframe.c

+++ b/vp9/decoder/vp9_decodframe.c

@@ -1267,13 +1267,14 @@

        * if we have enough data. Otherwise we will end up with the wrong

        * size.

        */

-      if (data + 4 < data_end) {

-        pc->Width = (data[0] | (data[1] << 8)) & 0x3fff;

-        pc->horiz_scale = data[1] >> 6;

-        pc->Height = (data[2] | (data[3] << 8)) & 0x3fff;

-        pc->vert_scale = data[3] >> 6;

+      if (data + 5 < data_end) {

+        pc->Width  = (data[0] | (data[1] << 8));

+        pc->Height = (data[2] | (data[3] << 8));

+

+        pc->horiz_scale = data[4] >> 4;

+        pc->vert_scale  = data[4] & 0x0F;

       }

-      data += 4;

+      data += 5;

       if (width != pc->Width || height != pc->Height) {

         if (pc->Width <= 0) {

--- a/vp9/encoder/vp9_bitstream.c

+++ b/vp9/encoder/vp9_bitstream.c

@@ -1500,17 +1500,20 @@

   {

     int v;

-    /* TODO(jkoleszar): support arbitrary resolutions */

-    v = (pc->horiz_scale << 14) | pc->Width;

+    // support arbitrary resolutions

+    v = pc->Width;

     cx_data[0] = v;

     cx_data[1] = v >> 8;

-    v = (pc->vert_scale << 14) | pc->Height;

+    v = pc->Height;

     cx_data[2] = v;

     cx_data[3] = v >> 8;

-    extra_bytes_packed += 4;

-    cx_data += 4;

+    // use a separate byte to store the scale factors, each ranging 0-15

+    cx_data[4] = (pc->horiz_scale << 4) | (pc->vert_scale);

+

+    extra_bytes_packed += 5;

+    cx_data += 5;

   }

   vp9_start_encode(&header_bc, cx_data);

--- a/vp9/encoder/vp9_encodeframe.c

+++ b/vp9/encoder/vp9_encodeframe.c

@@ -1247,8 +1247,8 @@

   MACROBLOCKD *const xd = &x->e_mbd;

   int totalrate;

-  // printf("encode_frame_internal frame %d (%d)\n",

-  //        cpi->common.current_video_frame, cpi->common.show_frame);

+//   fprintf(stderr, "encode_frame_internal frame %d (%d)\n",

+//          cpi->common.current_video_frame, cpi->common.show_frame);

   // Compute a modified set of reference frame probabilities to use when

   // prediction fails. These are based on the current general estimates for

@@ -1329,12 +1329,11 @@

       // Take tiles into account and give start/end MB

       int tile_col;

       TOKENEXTRA *tp = cpi->tok;

-

       for (tile_col = 0; tile_col < cm->tile_columns; tile_col++) {

         TOKENEXTRA *tp_old = tp;

-

         // For each row of SBs in the frame

         vp9_get_tile_col_offsets(cm, tile_col);

+

         for (mb_row = 0; mb_row < cm->mb_rows; mb_row += 4) {

           encode_sb_row(cpi, mb_row, &tp, &totalrate);

         }

--- a/vp9/encoder/vp9_onyx_if.c

+++ b/vp9/encoder/vp9_onyx_if.c

@@ -1086,14 +1086,12 @@

     cpi->max_gf_interval = cpi->twopass.static_scene_max_gf_interval;

 }

-

-static int

-rescale(int val, int num, int denom) {

+static int64_t rescale(int val, int64_t num, int denom) {

   int64_t llnum = num;

   int64_t llden = denom;

   int64_t llval = val;

-  return (int)(llval * llnum / llden);

+  return (llval * llnum / llden);

 }

 static void set_tile_limits(VP9_COMP *cpi) {

--- a/vp9/encoder/vp9_onyx_int.h

+++ b/vp9/encoder/vp9_onyx_int.h

@@ -485,7 +485,7 @@

   int kf_boost;

   int kf_zeromotion_pct;

-  int target_bandwidth;

+  int64_t target_bandwidth;

   struct vpx_codec_pkt_list  *output_pkt_list;

 #if 0

--- a/vp9/vp9_cx_iface.c

+++ b/vp9/vp9_cx_iface.c

@@ -126,8 +126,8 @@

 static vpx_codec_err_t validate_config(vpx_codec_alg_priv_t      *ctx,

                                        const vpx_codec_enc_cfg_t *cfg,

                                        const struct vp8_extracfg *vp8_cfg) {

-  RANGE_CHECK(cfg, g_w,                   1, 16383); /* 14 bits available */

-  RANGE_CHECK(cfg, g_h,                   1, 16383); /* 14 bits available */

+  RANGE_CHECK(cfg, g_w,                   1, 65535); /* 16 bits available */

+  RANGE_CHECK(cfg, g_h,                   1, 65535); /* 16 bits available */

   RANGE_CHECK(cfg, g_timebase.den,        1, 1000000000);

   RANGE_CHECK(cfg, g_timebase.num,        1, cfg->g_timebase.den);

   RANGE_CHECK_HI(cfg, g_profile,          3);

--- a/vp9/vp9_dx_iface.c

+++ b/vp9/vp9_dx_iface.c

@@ -229,8 +229,8 @@

       if (c[0] != 0x9d || c[1] != 0x01 || c[2] != 0x2a)

         res = VPX_CODEC_UNSUP_BITSTREAM;

-      si->w = (c[3] | (c[4] << 8)) & 0x3fff;

-      si->h = (c[5] | (c[6] << 8)) & 0x3fff;

+      si->w = (c[3] | (c[4] << 8));

+      si->h = (c[5] | (c[6] << 8));

       /*printf("w=%d, h=%d\n", si->w, si->h);*/

       if (!(si->h | si->w))