[FFmpeg-cvslog] binkaudio: use float sample format

Justin Ruggles git at videolan.org
Wed Sep 19 15:29:32 CEST 2012


ffmpeg | branch: master | Justin Ruggles <justin.ruggles at gmail.com> | Sun Aug 26 20:41:45 2012 -0400| [7bfd1766d1c18f07b0a2dd042418a874d49ea60d] | committer: Justin Ruggles

binkaudio: use float sample format

Use planar for DCT codec, interleaved for RDFT codec.

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7bfd1766d1c18f07b0a2dd042418a874d49ea60d
---

 libavcodec/binkaudio.c |   56 ++++++++++++++++++------------------------------
 1 file changed, 21 insertions(+), 35 deletions(-)

diff --git a/libavcodec/binkaudio.c b/libavcodec/binkaudio.c
index 915e7aa..957af79 100644
--- a/libavcodec/binkaudio.c
+++ b/libavcodec/binkaudio.c
@@ -47,7 +47,6 @@ static float quant_table[96];
 typedef struct {
     AVFrame frame;
     GetBitContext gb;
-    FmtConvertContext fmt_conv;
     int version_b;          ///< Bink version 'b'
     int first;
     int channels;
@@ -58,10 +57,7 @@ typedef struct {
     unsigned int *bands;
     float root;
     DECLARE_ALIGNED(32, FFTSample, coeffs)[BINK_BLOCK_MAX_SIZE];
-    DECLARE_ALIGNED(16, int16_t, previous)[BINK_BLOCK_MAX_SIZE / 16];  ///< coeffs from previous audio block
-    DECLARE_ALIGNED(16, int16_t, current)[BINK_BLOCK_MAX_SIZE / 16];
-    float *coeffs_ptr[MAX_CHANNELS]; ///< pointers to the coeffs arrays for float_to_int16_interleave
-    float *prev_ptr[MAX_CHANNELS];   ///< pointers to the overlap points in the coeffs array
+    float previous[MAX_CHANNELS][BINK_BLOCK_MAX_SIZE / 16];  ///< coeffs from previous audio block
     uint8_t *packet_buffer;
     union {
         RDFTContext rdft;
@@ -78,8 +74,6 @@ static av_cold int decode_init(AVCodecContext *avctx)
     int i;
     int frame_len_bits;
 
-    ff_fmt_convert_init(&s->fmt_conv, avctx);
-
     /* determine frame length */
     if (avctx->sample_rate < 22050) {
         frame_len_bits = 9;
@@ -98,12 +92,14 @@ static av_cold int decode_init(AVCodecContext *avctx)
 
     if (avctx->codec->id == AV_CODEC_ID_BINKAUDIO_RDFT) {
         // audio is already interleaved for the RDFT format variant
+        avctx->sample_fmt = AV_SAMPLE_FMT_FLT;
         sample_rate  *= avctx->channels;
         s->channels = 1;
         if (!s->version_b)
             frame_len_bits += av_log2(avctx->channels);
     } else {
         s->channels = avctx->channels;
+        avctx->sample_fmt = AV_SAMPLE_FMT_FLTP;
     }
 
     s->frame_len     = 1 << frame_len_bits;
@@ -111,9 +107,9 @@ static av_cold int decode_init(AVCodecContext *avctx)
     s->block_size    = (s->frame_len - s->overlap_len) * s->channels;
     sample_rate_half = (sample_rate + 1) / 2;
     if (avctx->codec->id == AV_CODEC_ID_BINKAUDIO_RDFT)
-        s->root = 2.0 / sqrt(s->frame_len);
+        s->root = 2.0 / (sqrt(s->frame_len) * 32768.0);
     else
-        s->root = s->frame_len / sqrt(s->frame_len);
+        s->root = s->frame_len / (sqrt(s->frame_len) * 32768.0);
     for (i = 0; i < 96; i++) {
         /* constant is result of 0.066399999/log10(M_E) */
         quant_table[i] = expf(i * 0.15289164787221953823f) * s->root;
@@ -135,12 +131,6 @@ static av_cold int decode_init(AVCodecContext *avctx)
     s->bands[s->num_bands] = s->frame_len;
 
     s->first = 1;
-    avctx->sample_fmt = AV_SAMPLE_FMT_S16;
-
-    for (i = 0; i < s->channels; i++) {
-        s->coeffs_ptr[i] = s->coeffs + i * s->frame_len;
-        s->prev_ptr[i]   = s->coeffs_ptr[i] + s->frame_len - s->overlap_len;
-    }
 
     if (CONFIG_BINKAUDIO_RDFT_DECODER && avctx->codec->id == AV_CODEC_ID_BINKAUDIO_RDFT)
         ff_rdft_init(&s->trans.rdft, frame_len_bits, DFT_C2R);
@@ -179,7 +169,7 @@ static const uint8_t rle_length_tab[16] = {
  * @param[out] out Output buffer (must contain s->block_size elements)
  * @return 0 on success, negative error code on failure
  */
-static int decode_block(BinkAudioContext *s, int16_t *out, int use_dct)
+static int decode_block(BinkAudioContext *s, float **out, int use_dct)
 {
     int ch, i, j, k;
     float q, quant[25];
@@ -190,7 +180,8 @@ static int decode_block(BinkAudioContext *s, int16_t *out, int use_dct)
         skip_bits(gb, 2);
 
     for (ch = 0; ch < s->channels; ch++) {
-        FFTSample *coeffs = s->coeffs_ptr[ch];
+        FFTSample *coeffs = out[ch];
+
         if (s->version_b) {
             if (get_bits_left(gb) < 64)
                 return AVERROR_INVALIDDATA;
@@ -265,24 +256,19 @@ static int decode_block(BinkAudioContext *s, int16_t *out, int use_dct)
             s->trans.rdft.rdft_calc(&s->trans.rdft, coeffs);
     }
 
-    s->fmt_conv.float_to_int16_interleave(s->current,
-                                          (const float **)s->prev_ptr,
-                                          s->overlap_len, s->channels);
-    s->fmt_conv.float_to_int16_interleave(out, (const float **)s->coeffs_ptr,
-                                          s->frame_len - s->overlap_len,
-                                          s->channels);
-
-    if (!s->first) {
+    for (ch = 0; ch < s->channels; ch++) {
+        int j;
         int count = s->overlap_len * s->channels;
-        int shift = av_log2(count);
-        for (i = 0; i < count; i++) {
-            out[i] = (s->previous[i] * (count - i) + out[i] * i) >> shift;
+        if (!s->first) {
+            j = ch;
+            for (i = 0; i < s->overlap_len; i++, j += s->channels)
+                out[ch][i] = (s->previous[ch][i] * (count - j) +
+                                      out[ch][i] *          j) / count;
         }
+        memcpy(s->previous[ch], &out[ch][s->frame_len - s->overlap_len],
+               s->overlap_len * sizeof(*s->previous[ch]));
     }
 
-    memcpy(s->previous, s->current,
-           s->overlap_len * s->channels * sizeof(*s->previous));
-
     s->first = 0;
 
     return 0;
@@ -311,7 +297,6 @@ static int decode_frame(AVCodecContext *avctx, void *data,
                         int *got_frame_ptr, AVPacket *avpkt)
 {
     BinkAudioContext *s = avctx->priv_data;
-    int16_t *samples;
     GetBitContext *gb = &s->gb;
     int ret, consumed = 0;
 
@@ -339,19 +324,20 @@ static int decode_frame(AVCodecContext *avctx, void *data,
     }
 
     /* get output buffer */
-    s->frame.nb_samples = s->block_size / avctx->channels;
+    s->frame.nb_samples = s->frame_len;
     if ((ret = avctx->get_buffer(avctx, &s->frame)) < 0) {
         av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
         return ret;
     }
-    samples = (int16_t *)s->frame.data[0];
 
-    if (decode_block(s, samples, avctx->codec->id == AV_CODEC_ID_BINKAUDIO_DCT)) {
+    if (decode_block(s, (float **)s->frame.extended_data,
+                     avctx->codec->id == AV_CODEC_ID_BINKAUDIO_DCT)) {
         av_log(avctx, AV_LOG_ERROR, "Incomplete packet\n");
         return AVERROR_INVALIDDATA;
     }
     get_bits_align32(gb);
 
+    s->frame.nb_samples = s->block_size / avctx->channels;
     *got_frame_ptr   = 1;
     *(AVFrame *)data = s->frame;
 



More information about the ffmpeg-cvslog mailing list