lavc: Add spherical packet side data API

[ffmpeg] / libavcodec / wmavoice.c

diff --git a/libavcodec/wmavoice.c b/libavcodec/wmavoice.c
index f3d0034ea883fad39144ad47c3a75787d0d441e2..19c06f44726f6a0dc388d1ab70bf09ff046d754f 100644 (file)
--- a/libavcodec/wmavoice.c
+++ b/libavcodec/wmavoice.c
@@ -131,7 +131,7 @@ static const struct frame_type_desc {
 /**
  * WMA Voice decoding context.
  */
-typedef struct {
+typedef struct WMAVoiceContext {
     /**
      * @name Global values specified in the stream header / extradata or used all over.
      * @{
@@ -205,7 +205,7 @@ typedef struct {
                                   ///< to #wmavoice_decode_packet() (since
                                   ///< they're part of the previous superframe)
 
-    uint8_t sframe_cache[SFRAME_CACHE_MAXSIZE + FF_INPUT_BUFFER_PADDING_SIZE];
+    uint8_t sframe_cache[SFRAME_CACHE_MAXSIZE + AV_INPUT_BUFFER_PADDING_SIZE];
                                   ///< cache for superframe data split over
                                   ///< multiple packets
     int sframe_cache_size;        ///< set to >0 if we have data from an
@@ -620,7 +620,7 @@ static void calc_input_response(WMAVoiceContext *s, float *lpcs,
     }
 
     /* calculate the Hilbert transform of the gains, which we do (since this
-     * is a sinus input) by doing a phase shift (in theory, H(sin())=cos()).
+     * is a sine input) by doing a phase shift (in theory, H(sin())=cos()).
      * Hilbert_Transform(RDFT(x)) = Laplace_Transform(x), which calculates the
      * "moment" of the LPCs in this filter. */
     s->dct.dct_calc(&s->dct, lpcs);
@@ -1052,9 +1052,10 @@ static void aw_parse_coords(WMAVoiceContext *s, GetBitContext *gb,
  * @param gb bit I/O context
  * @param block_idx block index in frame [0, 1]
  * @param fcb structure containing fixed codebook vector info
+ * @return -1 on error, 0 otherwise
  */
-static void aw_pulse_set2(WMAVoiceContext *s, GetBitContext *gb,
-                          int block_idx, AMRFixed *fcb)
+static int aw_pulse_set2(WMAVoiceContext *s, GetBitContext *gb,
+                         int block_idx, AMRFixed *fcb)
 {
     uint16_t use_mask_mem[9]; // only 5 are used, rest is padding
     uint16_t *use_mask = use_mask_mem + 2;
@@ -1116,7 +1117,7 @@ static void aw_pulse_set2(WMAVoiceContext *s, GetBitContext *gb,
             else if (use_mask[2]) idx = 0x2F;
             else if (use_mask[3]) idx = 0x3F;
             else if (use_mask[4]) idx = 0x4F;
-            else                  return;
+            else return -1;
             idx -= av_log2_16bit(use_mask[idx >> 4]);
         }
         if (use_mask[idx >> 4] & (0x8000 >> (idx & 15))) {
@@ -1133,6 +1134,7 @@ static void aw_pulse_set2(WMAVoiceContext *s, GetBitContext *gb,
     /* set offset for next block, relative to start of that block */
     n = (MAX_FRAMESIZE / 2 - start_off) % fcb->pitch_lag;
     s->aw_next_pulse_off_cache = n ? fcb->pitch_lag - n : 0;
+    return 0;
 }
 
 /**
@@ -1193,7 +1195,7 @@ static void aw_pulse_set1(WMAVoiceContext *s, GetBitContext *gb,
 /**
  * @}
  *
- * Generate a random number from frame_cntr and block_idx, which will lief
+ * Generate a random number from frame_cntr and block_idx, which will live
  * in the range [0, 1000 - block_size] (so it can be used as an index in a
  * table of size 1000 of which you want to read block_size entries).
  *
@@ -1295,7 +1297,18 @@ static void synth_block_fcb_acb(WMAVoiceContext *s, GetBitContext *gb,
      * (fixed) codebook pulses of the speech signal. */
     if (frame_desc->fcb_type == FCB_TYPE_AW_PULSES) {
         aw_pulse_set1(s, gb, block_idx, &fcb);
-        aw_pulse_set2(s, gb, block_idx, &fcb);
+        if (aw_pulse_set2(s, gb, block_idx, &fcb)) {
+            /* Conceal the block with silence and return.
+             * Skip the correct amount of bits to read the next
+             * block from the correct offset. */
+            int r_idx = pRNG(s->frame_cntr, block_idx, size);
+
+            for (n = 0; n < size; n++)
+                excitation[n] =
+                    wmavoice_std_codebook[r_idx + n] * s->silence_gain;
+            skip_bits(gb, 7 + 1);
+            return;
+        }
     } else /* FCB_TYPE_EXC_PULSES */ {
         int offset_nbits = 5 - frame_desc->log_n_blocks;
 
@@ -1505,7 +1518,7 @@ static int synth_frame(AVCodecContext *ctx, GetBitContext *gb, int frame_idx,
             /* Pitch is given per block. Per-block pitches are encoded as an
              * absolute value for the first block, and then delta values
              * relative to this value) for all subsequent blocks. The scale of
-             * this pitch value is semi-logaritmic compared to its use in the
+             * this pitch value is semi-logarithmic compared to its use in the
              * decoder, so we convert it to normal scale also. */
             int block_pitch,
                 t1 = (s->block_conv_table[1] - s->block_conv_table[0]) << 2,
@@ -1659,9 +1672,7 @@ static int check_bits_for_superframe(GetBitContext *orig_gb,
     const struct frame_type_desc *frame_desc;
 
     /* initialize a copy */
-    init_get_bits(gb, orig_gb->buffer, orig_gb->size_in_bits);
-    skip_bits_long(gb, get_bits_count(orig_gb));
-    assert(get_bits_left(gb) == get_bits_left(orig_gb));
+    *gb = *orig_gb;
 
     /* superframe header */
     if (get_bits_left(gb) < 14)
@@ -2053,6 +2064,7 @@ static av_cold void wmavoice_flush(AVCodecContext *ctx)
 
 AVCodec ff_wmavoice_decoder = {
     .name             = "wmavoice",
+    .long_name        = NULL_IF_CONFIG_SMALL("Windows Media Audio Voice"),
     .type             = AVMEDIA_TYPE_AUDIO,
     .id               = AV_CODEC_ID_WMAVOICE,
     .priv_data_size   = sizeof(WMAVoiceContext),
@@ -2060,7 +2072,6 @@ AVCodec ff_wmavoice_decoder = {
     .init_static_data = wmavoice_init_static_data,
     .close            = wmavoice_decode_end,
     .decode           = wmavoice_decode_packet,
-    .capabilities     = CODEC_CAP_SUBFRAMES | CODEC_CAP_DR1,
+    .capabilities     = AV_CODEC_CAP_SUBFRAMES | AV_CODEC_CAP_DR1,
     .flush            = wmavoice_flush,
-    .long_name        = NULL_IF_CONFIG_SMALL("Windows Media Audio Voice"),
 };