Merge remote-tracking branch 'qatar/master'

* qatar/master:
  wmaenc: use float planar sample format
  (e)ac3enc: use planar sample format
  aacenc: use planar sample format
  adpcmenc: use planar sample format for adpcm_ima_wav and adpcm_ima_qt
  adpcmenc: move 'ch' variable to higher scope
  adpcmenc: fix 3 instances of variable shadowing
  adpcm_ima_wav: simplify encoding
  libvorbis: use planar sample format
  libmp3lame: use planar sample formats
  vorbisenc: use float planar sample format
  ffm: do not write or read the audio sample format
  parseutils: fix parsing of invalid alpha values
  doc/RELEASE_NOTES: update for the 9 release.
  smoothstreamingenc: Add a more verbose error message
  smoothstreamingenc: Ignore the return value from mkdir
  smoothstreamingenc: Try writing a manifest when opening the muxer
  smoothstreamingenc: Move the output_chunk_list and write_manifest functions up
  smoothstreamingenc: Properly return errors from ism_flush to the caller
  smoothstreamingenc: Check the output UrlContext before accessing it

Conflicts:
	doc/RELEASE_NOTES
	libavcodec/aacenc.c
	libavcodec/ac3enc_template.c
	libavcodec/wmaenc.c
	tests/ref/lavf/ffm

Merged-by: Michael Niedermayer <michaelni@gmx.at>

12 files changed, 172 insertions, 221 deletions

diff --git a/libavcodec/aacenc.c b/libavcodec/aacenc.c
index 383cb5a7c1..9178babda6 100644
--- a/libavcodec/aacenc.c
+++ b/libavcodec/aacenc.c
@@ -479,31 +479,28 @@ static void put_bitstream_info(AVCodecContext *avctx, AACEncContext *s,
 }
 
 /*
- * Deinterleave input samples.
+ * Copy input samples.
  * Channels are reordered from libavcodec's default order to AAC order.
  */
-static void deinterleave_input_samples(AACEncContext *s, const AVFrame *frame)
+static void copy_input_samples(AACEncContext *s, const AVFrame *frame)
 {
-    int ch, i;
-    const int sinc = s->channels;
-    const uint8_t *channel_map = aac_chan_maps[sinc - 1];
+    int ch;
+    int end = 2048 + (frame ? frame->nb_samples : 0);
+    const uint8_t *channel_map = aac_chan_maps[s->channels - 1];
 
-    /* deinterleave and remap input samples */
-    for (ch = 0; ch < sinc; ch++) {
+    /* copy and remap input samples */
+    for (ch = 0; ch < s->channels; ch++) {
         /* copy last 1024 samples of previous frame to the start of the current frame */
         memcpy(&s->planar_samples[ch][1024], &s->planar_samples[ch][2048], 1024 * sizeof(s->planar_samples[0][0]));
 
-        /* deinterleave */
-        i = 2048;
+        /* copy new samples and zero any remaining samples */
         if (frame) {
-            const float *sptr = ((const float *)frame->data[0]) + channel_map[ch];
-            for (; i < 2048 + frame->nb_samples; i++) {
-                s->planar_samples[ch][i] = *sptr;
-                sptr += sinc;
-            }
+            memcpy(&s->planar_samples[ch][2048],
+                   frame->extended_data[channel_map[ch]],
+                   frame->nb_samples * sizeof(s->planar_samples[0][0]));
         }
-        memset(&s->planar_samples[ch][i], 0,
-               (3072 - i) * sizeof(s->planar_samples[0][0]));
+        memset(&s->planar_samples[ch][end], 0,
+               (3072 - end) * sizeof(s->planar_samples[0][0]));
     }
 }
 
@@ -526,7 +523,7 @@ static int aac_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
             return ret;
     }
 
-    deinterleave_input_samples(s, frame);
+    copy_input_samples(s, frame);
     if (s->psypp)
         ff_psy_preprocess(s->psypp, s->planar_samples, s->channels);
 
@@ -827,7 +824,7 @@ AVCodec ff_aac_encoder = {
     .supported_samplerates = avpriv_mpeg4audio_sample_rates,
     .capabilities   = CODEC_CAP_SMALL_LAST_FRAME | CODEC_CAP_DELAY |
                       CODEC_CAP_EXPERIMENTAL,
-    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLT,
+    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLTP,
                                                      AV_SAMPLE_FMT_NONE },
     .long_name      = NULL_IF_CONFIG_SMALL("AAC (Advanced Audio Coding)"),
     .priv_class     = &aacenc_class,
diff --git a/libavcodec/ac3enc_fixed.c b/libavcodec/ac3enc_fixed.c
index fa109f9791..5d8dd5c3f1 100644
--- a/libavcodec/ac3enc_fixed.c
+++ b/libavcodec/ac3enc_fixed.c
@@ -160,7 +160,7 @@ AVCodec ff_ac3_fixed_encoder = {
     .init            = ac3_fixed_encode_init,
     .encode2         = ff_ac3_fixed_encode_frame,
     .close           = ff_ac3_encode_close,
-    .sample_fmts     = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
+    .sample_fmts     = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16P,
                                                       AV_SAMPLE_FMT_NONE },
     .long_name       = NULL_IF_CONFIG_SMALL("ATSC A/52A (AC-3)"),
     .priv_class      = &ac3enc_class,
diff --git a/libavcodec/ac3enc_float.c b/libavcodec/ac3enc_float.c
index 5afe0fe2ab..7864f41019 100644
--- a/libavcodec/ac3enc_float.c
+++ b/libavcodec/ac3enc_float.c
@@ -158,7 +158,7 @@ AVCodec ff_ac3_encoder = {
     .init            = ff_ac3_encode_init,
     .encode2         = ff_ac3_float_encode_frame,
     .close           = ff_ac3_encode_close,
-    .sample_fmts     = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLT,
+    .sample_fmts     = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLTP,
                                                       AV_SAMPLE_FMT_NONE },
     .long_name       = NULL_IF_CONFIG_SMALL("ATSC A/52A (AC-3)"),
     .priv_class      = &ac3enc_class,
diff --git a/libavcodec/ac3enc_template.c b/libavcodec/ac3enc_template.c
index e81bfce5a8..904e0bb9ef 100644
--- a/libavcodec/ac3enc_template.c
+++ b/libavcodec/ac3enc_template.c
@@ -4,20 +4,20 @@
  * Copyright (c) 2006-2011 Justin Ruggles <justin.ruggles@gmail.com>
  * Copyright (c) 2006-2010 Prakash Punnoor <prakash@punnoor.de>
  *
- * This file is part of Libav.
+ * This file is part of FFmpeg.
  *
- * Libav is free software; you can redistribute it and/or
+ * 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.
  *
- * Libav is distributed in the hope that it will be useful,
+ * 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 Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
@@ -68,30 +68,23 @@ alloc_fail:
 
 
 /*
- * Deinterleave input samples.
+ * Copy input samples.
  * Channels are reordered from FFmpeg's default order to AC-3 order.
  */
-static void deinterleave_input_samples(AC3EncodeContext *s,
-                                       const SampleType *samples)
+static void copy_input_samples(AC3EncodeContext *s, SampleType **samples)
 {
-    int ch, i;
+    int ch;
 
-    /* deinterleave and remap input samples */
+    /* copy and remap input samples */
     for (ch = 0; ch < s->channels; ch++) {
-        const SampleType *sptr;
-        int sinc;
-
         /* copy last 256 samples of previous frame to the start of the current frame */
         memcpy(&s->planar_samples[ch][0], &s->planar_samples[ch][AC3_BLOCK_SIZE * s->num_blocks],
                AC3_BLOCK_SIZE * sizeof(s->planar_samples[0][0]));
 
-        /* deinterleave */
-        sinc = s->channels;
-        sptr = samples + s->channel_map[ch];
-        for (i = AC3_BLOCK_SIZE; i < AC3_BLOCK_SIZE * (s->num_blocks + 1); i++) {
-            s->planar_samples[ch][i] = *sptr;
-            sptr += sinc;
-        }
+        /* copy new samples for current frame */
+        memcpy(&s->planar_samples[ch][AC3_BLOCK_SIZE],
+               samples[s->channel_map[ch]],
+               AC3_BLOCK_SIZE * s->num_blocks * sizeof(s->planar_samples[0][0]));
     }
 }
 
@@ -395,7 +388,6 @@ int AC3_NAME(encode_frame)(AVCodecContext *avctx, AVPacket *avpkt,
                            const AVFrame *frame, int *got_packet_ptr)
 {
     AC3EncodeContext *s = avctx->priv_data;
-    const SampleType *samples = (const SampleType *)frame->data[0];
     int ret;
 
     if (s->options.allow_per_frame_metadata) {
@@ -407,7 +399,7 @@ int AC3_NAME(encode_frame)(AVCodecContext *avctx, AVPacket *avpkt,
     if (s->bit_alloc.sr_code == 1 || s->eac3)
         ff_ac3_adjust_frame_size(s);
 
-    deinterleave_input_samples(s, samples);
+    copy_input_samples(s, (SampleType **)frame->extended_data);
 
     apply_mdct(s);
 
diff --git a/libavcodec/adpcmenc.c b/libavcodec/adpcmenc.c
index a3194abb3c..f016ebd910 100644
--- a/libavcodec/adpcmenc.c
+++ b/libavcodec/adpcmenc.c
@@ -274,12 +274,11 @@ static inline uint8_t adpcm_yamaha_compress_sample(ADPCMChannelStatus *c,
 
 static void adpcm_compress_trellis(AVCodecContext *avctx,
                                    const int16_t *samples, uint8_t *dst,
-                                   ADPCMChannelStatus *c, int n)
+                                   ADPCMChannelStatus *c, int n, int stride)
 {
     //FIXME 6% faster if frontier is a compile-time constant
     ADPCMEncodeContext *s = avctx->priv_data;
     const int frontier = 1 << avctx->trellis;
-    const int stride   = avctx->channels;
     const int version  = avctx->codec->id;
     TrellisPath *paths       = s->paths, *p;
     TrellisNode *node_buf    = s->node_buf;
@@ -481,13 +480,15 @@ static void adpcm_compress_trellis(AVCodecContext *avctx,
 static int adpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
                               const AVFrame *frame, int *got_packet_ptr)
 {
-    int n, i, st, pkt_size, ret;
+    int n, i, ch, st, pkt_size, ret;
     const int16_t *samples;
+    int16_t **samples_p;
     uint8_t *dst;
     ADPCMEncodeContext *c = avctx->priv_data;
     uint8_t *buf;
 
     samples = (const int16_t *)frame->data[0];
+    samples_p = (int16_t **)frame->extended_data;
     st = avctx->channels == 2;
 
     if (avctx->codec_id == AV_CODEC_ID_ADPCM_SWF)
@@ -500,91 +501,71 @@ static int adpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
 
     switch(avctx->codec->id) {
     case AV_CODEC_ID_ADPCM_IMA_WAV:
-        n = frame->nb_samples / 8;
-        c->status[0].prev_sample = samples[0];
-        /* c->status[0].step_index = 0;
-        XXX: not sure how to init the state machine */
-        bytestream_put_le16(&dst, c->status[0].prev_sample);
-        *dst++ = c->status[0].step_index;
-        *dst++ = 0; /* unknown */
-        samples++;
-        if (avctx->channels == 2) {
-            c->status[1].prev_sample = samples[0];
-            /* c->status[1].step_index = 0; */
-            bytestream_put_le16(&dst, c->status[1].prev_sample);
-            *dst++ = c->status[1].step_index;
-            *dst++ = 0;
-            samples++;
+    {
+        int blocks, j;
+
+        blocks = (frame->nb_samples - 1) / 8;
+
+        for (ch = 0; ch < avctx->channels; ch++) {
+            ADPCMChannelStatus *status = &c->status[ch];
+            status->prev_sample = samples_p[ch][0];
+            /* status->step_index = 0;
+               XXX: not sure how to init the state machine */
+            bytestream_put_le16(&dst, status->prev_sample);
+            *dst++ = status->step_index;
+            *dst++ = 0; /* unknown */
         }
 
-        /* stereo: 4 bytes (8 samples) for left,
-            4 bytes for right, 4 bytes left, ... */
+        /* stereo: 4 bytes (8 samples) for left, 4 bytes for right */
         if (avctx->trellis > 0) {
-            FF_ALLOC_OR_GOTO(avctx, buf, 2 * n * 8, error);
-            adpcm_compress_trellis(avctx, samples, buf, &c->status[0], n * 8);
-            if (avctx->channels == 2)
-                adpcm_compress_trellis(avctx, samples + 1, buf + n * 8,
-                                       &c->status[1], n * 8);
-            for (i = 0; i < n; i++) {
-                *dst++ = buf[8 * i + 0] | (buf[8 * i + 1] << 4);
-                *dst++ = buf[8 * i + 2] | (buf[8 * i + 3] << 4);
-                *dst++ = buf[8 * i + 4] | (buf[8 * i + 5] << 4);
-                *dst++ = buf[8 * i + 6] | (buf[8 * i + 7] << 4);
-                if (avctx->channels == 2) {
-                    uint8_t *buf1 = buf + n * 8;
-                    *dst++ = buf1[8 * i + 0] | (buf1[8 * i + 1] << 4);
-                    *dst++ = buf1[8 * i + 2] | (buf1[8 * i + 3] << 4);
-                    *dst++ = buf1[8 * i + 4] | (buf1[8 * i + 5] << 4);
-                    *dst++ = buf1[8 * i + 6] | (buf1[8 * i + 7] << 4);
+            FF_ALLOC_OR_GOTO(avctx, buf, avctx->channels * blocks * 8, error);
+            for (ch = 0; ch < avctx->channels; ch++) {
+                adpcm_compress_trellis(avctx, &samples_p[ch][1],
+                                       buf + ch * blocks * 8, &c->status[ch],
+                                       blocks * 8, 1);
+            }
+            for (i = 0; i < blocks; i++) {
+                for (ch = 0; ch < avctx->channels; ch++) {
+                    uint8_t *buf1 = buf + ch * blocks * 8 + i * 8;
+                    for (j = 0; j < 8; j += 2)
+                        *dst++ = buf1[j] | (buf1[j + 1] << 4);
                 }
             }
             av_free(buf);
         } else {
-            for (; n > 0; n--) {
-                *dst    = adpcm_ima_compress_sample(&c->status[0], samples[0]);
-                *dst++ |= adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels    ]) << 4;
-                *dst    = adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels * 2]);
-                *dst++ |= adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels * 3]) << 4;
-                *dst    = adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels * 4]);
-                *dst++ |= adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels * 5]) << 4;
-                *dst    = adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels * 6]);
-                *dst++ |= adpcm_ima_compress_sample(&c->status[0], samples[avctx->channels * 7]) << 4;
-                /* right channel */
-                if (avctx->channels == 2) {
-                    *dst    = adpcm_ima_compress_sample(&c->status[1], samples[1 ]);
-                    *dst++ |= adpcm_ima_compress_sample(&c->status[1], samples[3 ]) << 4;
-                    *dst    = adpcm_ima_compress_sample(&c->status[1], samples[5 ]);
-                    *dst++ |= adpcm_ima_compress_sample(&c->status[1], samples[7 ]) << 4;
-                    *dst    = adpcm_ima_compress_sample(&c->status[1], samples[9 ]);
-                    *dst++ |= adpcm_ima_compress_sample(&c->status[1], samples[11]) << 4;
-                    *dst    = adpcm_ima_compress_sample(&c->status[1], samples[13]);
-                    *dst++ |= adpcm_ima_compress_sample(&c->status[1], samples[15]) << 4;
+            for (i = 0; i < blocks; i++) {
+                for (ch = 0; ch < avctx->channels; ch++) {
+                    ADPCMChannelStatus *status = &c->status[ch];
+                    const int16_t *smp = &samples_p[ch][1 + i * 8];
+                    for (j = 0; j < 8; j += 2) {
+                        *dst++ = adpcm_ima_compress_sample(status, smp[j    ]) |
+                                (adpcm_ima_compress_sample(status, smp[j + 1]) << 4);
+                    }
                 }
-                samples += 8 * avctx->channels;
             }
         }
         break;
+    }
     case AV_CODEC_ID_ADPCM_IMA_QT:
     {
-        int ch, i;
         PutBitContext pb;
         init_put_bits(&pb, dst, pkt_size * 8);
 
         for (ch = 0; ch < avctx->channels; ch++) {
-            put_bits(&pb, 9, (c->status[ch].prev_sample & 0xFFFF) >> 7);
-            put_bits(&pb, 7,  c->status[ch].step_index);
+            ADPCMChannelStatus *status = &c->status[ch];
+            put_bits(&pb, 9, (status->prev_sample & 0xFFFF) >> 7);
+            put_bits(&pb, 7,  status->step_index);
             if (avctx->trellis > 0) {
                 uint8_t buf[64];
-                adpcm_compress_trellis(avctx, samples+ch, buf, &c->status[ch], 64);
+                adpcm_compress_trellis(avctx, &samples_p[ch][1], buf, status,
+                                       64, 1);
                 for (i = 0; i < 64; i++)
                     put_bits(&pb, 4, buf[i ^ 1]);
             } else {
                 for (i = 0; i < 64; i += 2) {
                     int t1, t2;
-                    t1 = adpcm_ima_qt_compress_sample(&c->status[ch],
-                                                      samples[avctx->channels * (i + 0) + ch]);
-                    t2 = adpcm_ima_qt_compress_sample(&c->status[ch],
-                                                      samples[avctx->channels * (i + 1) + ch]);
+                    t1 = adpcm_ima_qt_compress_sample(status, samples_p[ch][i    ]);
+                    t2 = adpcm_ima_qt_compress_sample(status, samples_p[ch][i + 1]);
                     put_bits(&pb, 4, t2);
                     put_bits(&pb, 4, t1);
                 }
@@ -596,7 +577,6 @@ static int adpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
     }
     case AV_CODEC_ID_ADPCM_SWF:
     {
-        int i;
         PutBitContext pb;
         init_put_bits(&pb, dst, pkt_size * 8);
 
@@ -617,10 +597,11 @@ static int adpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
         if (avctx->trellis > 0) {
             FF_ALLOC_OR_GOTO(avctx, buf, 2 * n, error);
             adpcm_compress_trellis(avctx, samples + avctx->channels, buf,
-                                   &c->status[0], n);
+                                   &c->status[0], n, avctx->channels);
             if (avctx->channels == 2)
                 adpcm_compress_trellis(avctx, samples + avctx->channels + 1,
-                                       buf + n, &c->status[1], n);
+                                       buf + n, &c->status[1], n,
+                                       avctx->channels);
             for (i = 0; i < n; i++) {
                 put_bits(&pb, 4, buf[i]);
                 if (avctx->channels == 2)
@@ -661,15 +642,18 @@ static int adpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
             bytestream_put_le16(&dst, c->status[i].sample2);
 
         if (avctx->trellis > 0) {
-            int n = avctx->block_align - 7 * avctx->channels;
+            n = avctx->block_align - 7 * avctx->channels;
             FF_ALLOC_OR_GOTO(avctx, buf, 2 * n, error);
             if (avctx->channels == 1) {
-                adpcm_compress_trellis(avctx, samples, buf, &c->status[0], n);
+                adpcm_compress_trellis(avctx, samples, buf, &c->status[0], n,
+                                       avctx->channels);
                 for (i = 0; i < n; i += 2)
                     *dst++ = (buf[i] << 4) | buf[i + 1];
             } else {
-                adpcm_compress_trellis(avctx, samples,     buf,     &c->status[0], n);
-                adpcm_compress_trellis(avctx, samples + 1, buf + n, &c->status[1], n);
+                adpcm_compress_trellis(avctx, samples,     buf,
+                                       &c->status[0], n, avctx->channels);
+                adpcm_compress_trellis(avctx, samples + 1, buf + n,
+                                       &c->status[1], n, avctx->channels);
                 for (i = 0; i < n; i++)
                     *dst++ = (buf[i] << 4) | buf[n + i];
             }
@@ -689,12 +673,15 @@ static int adpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
             FF_ALLOC_OR_GOTO(avctx, buf, 2 * n * 2, error);
             n *= 2;
             if (avctx->channels == 1) {
-                adpcm_compress_trellis(avctx, samples, buf, &c->status[0], n);
+                adpcm_compress_trellis(avctx, samples, buf, &c->status[0], n,
+                                       avctx->channels);
                 for (i = 0; i < n; i += 2)
                     *dst++ = buf[i] | (buf[i + 1] << 4);
             } else {
-                adpcm_compress_trellis(avctx, samples,     buf,     &c->status[0], n);
-                adpcm_compress_trellis(avctx, samples + 1, buf + n, &c->status[1], n);
+                adpcm_compress_trellis(avctx, samples,     buf,
+                                       &c->status[0], n, avctx->channels);
+                adpcm_compress_trellis(avctx, samples + 1, buf + n,
+                                       &c->status[1], n, avctx->channels);
                 for (i = 0; i < n; i++)
                     *dst++ = buf[i] | (buf[n + i] << 4);
             }
@@ -722,7 +709,11 @@ static const enum AVSampleFormat sample_fmts[] = {
     AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE
 };
 
-#define ADPCM_ENCODER(id_, name_, long_name_)               \
+static const enum AVSampleFormat sample_fmts_p[] = {
+    AV_SAMPLE_FMT_S16P, AV_SAMPLE_FMT_NONE
+};
+
+#define ADPCM_ENCODER(id_, name_, sample_fmts_, long_name_) \
 AVCodec ff_ ## name_ ## _encoder = {                        \
     .name           = #name_,                               \
     .type           = AVMEDIA_TYPE_AUDIO,                   \
@@ -731,12 +722,12 @@ AVCodec ff_ ## name_ ## _encoder = {                        \
     .init           = adpcm_encode_init,                    \
     .encode2        = adpcm_encode_frame,                   \
     .close          = adpcm_encode_close,                   \
-    .sample_fmts    = sample_fmts,                          \
+    .sample_fmts    = sample_fmts_,                         \
     .long_name      = NULL_IF_CONFIG_SMALL(long_name_),     \
 }
 
-ADPCM_ENCODER(AV_CODEC_ID_ADPCM_IMA_QT, adpcm_ima_qt,   "ADPCM IMA QuickTime");
-ADPCM_ENCODER(AV_CODEC_ID_ADPCM_IMA_WAV, adpcm_ima_wav, "ADPCM IMA WAV");
-ADPCM_ENCODER(AV_CODEC_ID_ADPCM_MS, adpcm_ms,           "ADPCM Microsoft");
-ADPCM_ENCODER(AV_CODEC_ID_ADPCM_SWF, adpcm_swf,         "ADPCM Shockwave Flash");
-ADPCM_ENCODER(AV_CODEC_ID_ADPCM_YAMAHA, adpcm_yamaha,   "ADPCM Yamaha");
+ADPCM_ENCODER(AV_CODEC_ID_ADPCM_IMA_QT,  adpcm_ima_qt,  sample_fmts_p, "ADPCM IMA QuickTime");
+ADPCM_ENCODER(AV_CODEC_ID_ADPCM_IMA_WAV, adpcm_ima_wav, sample_fmts_p, "ADPCM IMA WAV");
+ADPCM_ENCODER(AV_CODEC_ID_ADPCM_MS,      adpcm_ms,      sample_fmts,   "ADPCM Microsoft");
+ADPCM_ENCODER(AV_CODEC_ID_ADPCM_SWF,     adpcm_swf,     sample_fmts,   "ADPCM Shockwave Flash");
+ADPCM_ENCODER(AV_CODEC_ID_ADPCM_YAMAHA,  adpcm_yamaha,  sample_fmts,   "ADPCM Yamaha");
diff --git a/libavcodec/eac3enc.c b/libavcodec/eac3enc.c
index c5ac6d1516..bb9ef4fd3e 100644
--- a/libavcodec/eac3enc.c
+++ b/libavcodec/eac3enc.c
@@ -258,7 +258,7 @@ AVCodec ff_eac3_encoder = {
     .init            = ff_ac3_encode_init,
     .encode2         = ff_ac3_float_encode_frame,
     .close           = ff_ac3_encode_close,
-    .sample_fmts     = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLT,
+    .sample_fmts     = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLTP,
                                                       AV_SAMPLE_FMT_NONE },
     .long_name       = NULL_IF_CONFIG_SMALL("ATSC A/52 E-AC-3"),
     .priv_class      = &eac3enc_class,
diff --git a/libavcodec/libmp3lame.c b/libavcodec/libmp3lame.c
index 427cacdf8e..ce17e06ef8 100644
--- a/libavcodec/libmp3lame.c
+++ b/libavcodec/libmp3lame.c
@@ -33,6 +33,7 @@
 #include "libavutil/opt.h"
 #include "avcodec.h"
 #include "audio_frame_queue.h"
+#include "dsputil.h"
 #include "internal.h"
 #include "mpegaudio.h"
 #include "mpegaudiodecheader.h"
@@ -46,8 +47,9 @@ typedef struct LAMEContext {
     uint8_t buffer[BUFFER_SIZE];
     int buffer_index;
     int reservoir;
-    void *planar_samples[2];
+    float *samples_flt[2];
     AudioFrameQueue afq;
+    DSPContext dsp;
 } LAMEContext;
 
 
@@ -58,8 +60,8 @@ static av_cold int mp3lame_encode_close(AVCodecContext *avctx)
 #if FF_API_OLD_ENCODE_AUDIO
     av_freep(&avctx->coded_frame);
 #endif
-    av_freep(&s->planar_samples[0]);
-    av_freep(&s->planar_samples[1]);
+    av_freep(&s->samples_flt[0]);
+    av_freep(&s->samples_flt[1]);
 
     ff_af_queue_close(&s->afq);
 
@@ -127,93 +129,63 @@ static av_cold int mp3lame_encode_init(AVCodecContext *avctx)
     }
 #endif
 
-    /* sample format */
-    if (avctx->sample_fmt == AV_SAMPLE_FMT_S32 ||
-        avctx->sample_fmt == AV_SAMPLE_FMT_FLT) {
+    /* allocate float sample buffers */
+    if (avctx->sample_fmt == AV_SAMPLE_FMT_FLTP) {
         int ch;
         for (ch = 0; ch < avctx->channels; ch++) {
-            s->planar_samples[ch] = av_malloc(avctx->frame_size *
-                                              av_get_bytes_per_sample(avctx->sample_fmt));
-            if (!s->planar_samples[ch]) {
+            s->samples_flt[ch] = av_malloc(avctx->frame_size *
+                                           sizeof(*s->samples_flt[ch]));
+            if (!s->samples_flt[ch]) {
                 ret = AVERROR(ENOMEM);
                 goto error;
             }
         }
     }
 
+    ff_dsputil_init(&s->dsp, avctx);
+
     return 0;
 error:
     mp3lame_encode_close(avctx);
     return ret;
 }
 
-#define DEINTERLEAVE(type, scale) do {                  \
-    int ch, i;                                          \
-    for (ch = 0; ch < s->avctx->channels; ch++) {       \
-        const type *input = samples;                    \
-        type      *output = s->planar_samples[ch];      \
-        input += ch;                                    \
-        for (i = 0; i < nb_samples; i++) {              \
-            output[i] = *input * scale;                 \
-            input += s->avctx->channels;                \
-        }                                               \
-    }                                                   \
+#define ENCODE_BUFFER(func, buf_type, buf_name) do {                        \
+    lame_result = func(s->gfp,                                              \
+                       (const buf_type *)buf_name[0],                       \
+                       (const buf_type *)buf_name[1], frame->nb_samples,    \
+                       s->buffer + s->buffer_index,                         \
+                       BUFFER_SIZE - s->buffer_index);                      \
 } while (0)
 
-static int encode_frame_int16(LAMEContext *s, void *samples, int nb_samples)
-{
-    if (s->avctx->channels > 1) {
-        return lame_encode_buffer_interleaved(s->gfp, samples,
-                                              nb_samples,
-                                              s->buffer + s->buffer_index,
-                                              BUFFER_SIZE - s->buffer_index);
-    } else {
-        return lame_encode_buffer(s->gfp, samples, NULL, nb_samples,
-                                  s->buffer + s->buffer_index,
-                                  BUFFER_SIZE - s->buffer_index);
-    }
-}
-
-static int encode_frame_int32(LAMEContext *s, void *samples, int nb_samples)
-{
-    DEINTERLEAVE(int32_t, 1);
-
-    return lame_encode_buffer_int(s->gfp,
-                                  s->planar_samples[0], s->planar_samples[1],
-                                  nb_samples,
-                                  s->buffer + s->buffer_index,
-                                  BUFFER_SIZE - s->buffer_index);
-}
-
-static int encode_frame_float(LAMEContext *s, void *samples, int nb_samples)
-{
-    DEINTERLEAVE(float, 32768.0f);
-
-    return lame_encode_buffer_float(s->gfp,
-                                    s->planar_samples[0], s->planar_samples[1],
-                                    nb_samples,
-                                    s->buffer + s->buffer_index,
-                                    BUFFER_SIZE - s->buffer_index);
-}
-
 static int mp3lame_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
                                 const AVFrame *frame, int *got_packet_ptr)
 {
     LAMEContext *s = avctx->priv_data;
     MPADecodeHeader hdr;
-    int len, ret;
+    int len, ret, ch;
     int lame_result;
 
     if (frame) {
         switch (avctx->sample_fmt) {
-        case AV_SAMPLE_FMT_S16:
-            lame_result = encode_frame_int16(s, frame->data[0], frame->nb_samples);
+        case AV_SAMPLE_FMT_S16P:
+            ENCODE_BUFFER(lame_encode_buffer, int16_t, frame->data);
             break;
-        case AV_SAMPLE_FMT_S32:
-            lame_result = encode_frame_int32(s, frame->data[0], frame->nb_samples);
+        case AV_SAMPLE_FMT_S32P:
+            ENCODE_BUFFER(lame_encode_buffer_int, int32_t, frame->data);
             break;
-        case AV_SAMPLE_FMT_FLT:
-            lame_result = encode_frame_float(s, frame->data[0], frame->nb_samples);
+        case AV_SAMPLE_FMT_FLTP:
+            if (frame->linesize[0] < 4 * FFALIGN(frame->nb_samples, 8)) {
+                av_log(avctx, AV_LOG_ERROR, "inadequate AVFrame plane padding\n");
+                return AVERROR(EINVAL);
+            }
+            for (ch = 0; ch < avctx->channels; ch++) {
+                s->dsp.vector_fmul_scalar(s->samples_flt[ch],
+                                          (const float *)frame->data[ch],
+                                          32768.0f,
+                                          FFALIGN(frame->nb_samples, 8));
+            }
+            ENCODE_BUFFER(lame_encode_buffer_float, float, s->samples_flt);
             break;
         default:
             return AVERROR_BUG;
@@ -299,9 +271,9 @@ AVCodec ff_libmp3lame_encoder = {
     .encode2               = mp3lame_encode_frame,
     .close                 = mp3lame_encode_close,
     .capabilities          = CODEC_CAP_DELAY | CODEC_CAP_SMALL_LAST_FRAME,
-    .sample_fmts           = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S32,
-                                                             AV_SAMPLE_FMT_FLT,
-                                                             AV_SAMPLE_FMT_S16,
+    .sample_fmts           = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S32P,
+                                                             AV_SAMPLE_FMT_FLTP,
+                                                             AV_SAMPLE_FMT_S16P,
                                                              AV_SAMPLE_FMT_NONE },
     .supported_samplerates = libmp3lame_sample_rates,
     .channel_layouts       = (const uint64_t[]) { AV_CH_LAYOUT_MONO,
diff --git a/libavcodec/libvorbisenc.c b/libavcodec/libvorbisenc.c
index b042e0cb4f..6064dd5649 100644
--- a/libavcodec/libvorbisenc.c
+++ b/libavcodec/libvorbisenc.c
@@ -290,18 +290,16 @@ static int oggvorbis_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
 
     /* send samples to libvorbis */
     if (frame) {
-        const float *audio = (const float *)frame->data[0];
         const int samples = frame->nb_samples;
         float **buffer;
         int c, channels = s->vi.channels;
 
         buffer = vorbis_analysis_buffer(&s->vd, samples);
         for (c = 0; c < channels; c++) {
-            int i;
             int co = (channels > 8) ? c :
                      ff_vorbis_encoding_channel_layout_offsets[channels - 1][c];
-            for (i = 0; i < samples; i++)
-                buffer[c][i] = audio[i * channels + co];
+            memcpy(buffer[c], frame->extended_data[co],
+                   samples * sizeof(*buffer[c]));
         }
         if ((ret = vorbis_analysis_wrote(&s->vd, samples)) < 0) {
             av_log(avctx, AV_LOG_ERROR, "error in vorbis_analysis_wrote()\n");
@@ -383,7 +381,7 @@ AVCodec ff_libvorbis_encoder = {
     .encode2        = oggvorbis_encode_frame,
     .close          = oggvorbis_encode_close,
     .capabilities   = CODEC_CAP_DELAY,
-    .sample_fmts    = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLT,
+    .sample_fmts    = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
                                                       AV_SAMPLE_FMT_NONE },
     .long_name      = NULL_IF_CONFIG_SMALL("libvorbis"),
     .priv_class     = &class,
diff --git a/libavcodec/vorbisenc.c b/libavcodec/vorbisenc.c
index b88791f274..74d6331ee3 100644
--- a/libavcodec/vorbisenc.c
+++ b/libavcodec/vorbisenc.c
@@ -963,10 +963,10 @@ static int residue_encode(vorbis_enc_context *venc, vorbis_enc_residue *rc,
     return 0;
 }
 
-static int apply_window_and_mdct(vorbis_enc_context *venc, const signed short *audio,
-                                 int samples)
+static int apply_window_and_mdct(vorbis_enc_context *venc,
+                                 float **audio, int samples)
 {
-    int i, j, channel;
+    int i, channel;
     const float * win = venc->win[0];
     int window_len = 1 << (venc->log2_blocksize[0] - 1);
     float n = (float)(1 << venc->log2_blocksize[0]) / 4.;
@@ -988,9 +988,8 @@ static int apply_window_and_mdct(vorbis_enc_context *venc, const signed short *a
     if (samples) {
         for (channel = 0; channel < venc->channels; channel++) {
             float * offset = venc->samples + channel*window_len*2 + window_len;
-            j = channel;
-            for (i = 0; i < samples; i++, j += venc->channels)
-                offset[i] = audio[j] / 32768. / n * win[window_len - i - 1];
+            for (i = 0; i < samples; i++)
+                offset[i] = audio[channel][i] / n * win[window_len - i - 1];
         }
     } else {
         for (channel = 0; channel < venc->channels; channel++)
@@ -1005,9 +1004,8 @@ static int apply_window_and_mdct(vorbis_enc_context *venc, const signed short *a
     if (samples) {
         for (channel = 0; channel < venc->channels; channel++) {
             float *offset = venc->saved + channel * window_len;
-            j = channel;
-            for (i = 0; i < samples; i++, j += venc->channels)
-                offset[i] = audio[j] / 32768. / n * win[i];
+            for (i = 0; i < samples; i++)
+                offset[i] = audio[channel][i] / n * win[i];
         }
         venc->have_saved = 1;
     } else {
@@ -1020,7 +1018,7 @@ static int vorbis_encode_frame(AVCodecContext *avccontext, AVPacket *avpkt,
                                const AVFrame *frame, int *got_packet_ptr)
 {
     vorbis_enc_context *venc = avccontext->priv_data;
-    const int16_t *audio = frame ? (const int16_t *)frame->data[0] : NULL;
+    float **audio = frame ? (float **)frame->extended_data : NULL;
     int samples = frame ? frame->nb_samples : 0;
     vorbis_enc_mode *mode;
     vorbis_enc_mapping *mapping;
@@ -1213,7 +1211,7 @@ AVCodec ff_vorbis_encoder = {
     .encode2        = vorbis_encode_frame,
     .close          = vorbis_encode_close,
     .capabilities   = CODEC_CAP_DELAY | CODEC_CAP_EXPERIMENTAL,
-    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
+    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLTP,
                                                      AV_SAMPLE_FMT_NONE },
     .long_name      = NULL_IF_CONFIG_SMALL("Vorbis"),
 };
diff --git a/libavcodec/wma.c b/libavcodec/wma.c
index d5e66e25b3..c65281711a 100644
--- a/libavcodec/wma.c
+++ b/libavcodec/wma.c
@@ -89,6 +89,7 @@ int ff_wma_init(AVCodecContext *avctx, int flags2)
 
     ff_dsputil_init(&s->dsp, avctx);
     ff_fmt_convert_init(&s->fmt_conv, avctx);
+    avpriv_float_dsp_init(&s->fdsp, avctx->flags & CODEC_FLAG_BITEXACT);
 
     if (avctx->codec->id == AV_CODEC_ID_WMAV1) {
         s->version = 1;
diff --git a/libavcodec/wma.h b/libavcodec/wma.h
index 73734ebc7c..1e7f0e5aba 100644
--- a/libavcodec/wma.h
+++ b/libavcodec/wma.h
@@ -22,6 +22,7 @@
 #ifndef AVCODEC_WMA_H
 #define AVCODEC_WMA_H
 
+#include "libavutil/float_dsp.h"
 #include "get_bits.h"
 #include "put_bits.h"
 #include "dsputil.h"
@@ -137,6 +138,7 @@ typedef struct WMACodecContext {
     float lsp_pow_m_table2[(1 << LSP_POW_BITS)];
     DSPContext dsp;
     FmtConvertContext fmt_conv;
+    AVFloatDSPContext fdsp;
 
 #ifdef TRACE
     int frame_count;
diff --git a/libavcodec/wmaenc.c b/libavcodec/wmaenc.c
index 835bc39ce8..63f815c78a 100644
--- a/libavcodec/wmaenc.c
+++ b/libavcodec/wmaenc.c
@@ -93,23 +93,24 @@ static int encode_init(AVCodecContext * avctx){
 }
 
 
-static void apply_window_and_mdct(AVCodecContext * avctx, const signed short * audio, int len) {
+static void apply_window_and_mdct(AVCodecContext * avctx, const AVFrame *frame)
+{
     WMACodecContext *s = avctx->priv_data;
+    float **audio      = (float **)frame->extended_data;
+    int len            = frame->nb_samples;
     int window_index= s->frame_len_bits - s->block_len_bits;
     FFTContext *mdct = &s->mdct_ctx[window_index];
-    int i, j, channel;
+    int ch;
     const float * win = s->windows[window_index];
     int window_len = 1 << s->block_len_bits;
-    float n = window_len/2;
-
-    for (channel = 0; channel < avctx->channels; channel++) {
-        memcpy(s->output, s->frame_out[channel], sizeof(float)*window_len);
-        j = channel;
-        for (i = 0; i < len; i++, j += avctx->channels){
-            s->output[i+window_len]  = audio[j] / n * win[window_len - i - 1];
-            s->frame_out[channel][i] = audio[j] / n * win[i];
-        }
-        mdct->mdct_calc(mdct, s->coefs[channel], s->output);
+    float n = 2.0 * 32768.0 / window_len;
+
+    for (ch = 0; ch < avctx->channels; ch++) {
+        memcpy(s->output, s->frame_out[ch], window_len * sizeof(*s->output));
+        s->dsp.vector_fmul_scalar(s->frame_out[ch], audio[ch], n, len);
+        s->dsp.vector_fmul_reverse(&s->output[window_len], s->frame_out[ch], win, len);
+        s->fdsp.vector_fmul(s->frame_out[ch], s->frame_out[ch], win, len);
+        mdct->mdct_calc(mdct, s->coefs[ch], s->output);
     }
 }
 
@@ -345,13 +346,12 @@ static int encode_superframe(AVCodecContext *avctx, AVPacket *avpkt,
                              const AVFrame *frame, int *got_packet_ptr)
 {
     WMACodecContext *s = avctx->priv_data;
-    const int16_t *samples = (const int16_t *)frame->data[0];
     int i, total_gain, ret, error;
 
     s->block_len_bits= s->frame_len_bits; //required by non variable block len
     s->block_len = 1 << s->block_len_bits;
 
-    apply_window_and_mdct(avctx, samples, frame->nb_samples);
+    apply_window_and_mdct(avctx, frame);
 
     if (s->ms_stereo) {
         float a, b;
@@ -404,7 +404,7 @@ AVCodec ff_wmav1_encoder = {
     .init           = encode_init,
     .encode2        = encode_superframe,
     .close          = ff_wma_end,
-    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
+    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLTP,
                                                      AV_SAMPLE_FMT_NONE },
     .long_name      = NULL_IF_CONFIG_SMALL("Windows Media Audio 1"),
 };
@@ -418,7 +418,7 @@ AVCodec ff_wmav2_encoder = {
     .init           = encode_init,
     .encode2        = encode_superframe,
     .close          = ff_wma_end,
-    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
+    .sample_fmts    = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_FLTP,
                                                      AV_SAMPLE_FMT_NONE },
     .long_name      = NULL_IF_CONFIG_SMALL("Windows Media Audio 2"),
 };