aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorPaul B Mahol <onemda@gmail.com>2017-01-17 15:54:57 +0100
committerPaul B Mahol <onemda@gmail.com>2017-01-19 00:37:26 +0100
commit6c43f33ac2e7606b2013f6261144389589394196 (patch)
treef1e1bba0257696d011e4577cba6a34cb7eafd21d
parent90096e42e1e9c075a1991e592c5965cf3f643cd1 (diff)
downloadffmpeg-6c43f33ac2e7606b2013f6261144389589394196.tar.gz
avcodec/wmaprodec: >2 channel support for XMA
Signed-off-by: Paul B Mahol <onemda@gmail.com>
-rw-r--r--libavcodec/wmaprodec.c274
-rw-r--r--libavformat/wavdec.c3
2 files changed, 204 insertions, 73 deletions
diff --git a/libavcodec/wmaprodec.c b/libavcodec/wmaprodec.c
index 105e27999e..a53c64c1d3 100644
--- a/libavcodec/wmaprodec.c
+++ b/libavcodec/wmaprodec.c
@@ -207,19 +207,19 @@ typedef struct WMAProDecodeCtx {
int subframe_offset; ///< subframe offset in the bit reservoir
uint8_t packet_loss; ///< set in case of bitstream error
uint8_t packet_done; ///< set when a packet is fully decoded
- uint8_t skip_packets;
/* frame decode state */
uint32_t frame_num; ///< current frame number (not used for decoding)
- int num_frames;
GetBitContext gb; ///< bitstream reader context
int buf_bit_size; ///< buffer size in bits
uint8_t drc_gain; ///< gain for the DRC tool
int8_t skip_frame; ///< skip output step
int8_t parsed_all_subframes; ///< all subframes decoded?
+ uint8_t skip_packets;
/* subframe/block decode state */
int16_t subframe_len; ///< current subframe length
+ int8_t nb_channels; ///< number of channels in stream (XMA1/2)
int8_t channels_for_cur_subframe; ///< number of channels that contain the subframe
int8_t channel_indexes_for_cur_subframe[WMAPRO_MAX_CHANNELS];
int8_t num_bands; ///< number of scale factor bands
@@ -234,6 +234,13 @@ typedef struct WMAProDecodeCtx {
WMAProChannelCtx channel[WMAPRO_MAX_CHANNELS]; ///< per channel data
} WMAProDecodeCtx;
+typedef struct XMADecodeCtx {
+ WMAProDecodeCtx xma[4];
+ AVFrame *frames[4];
+ int current_stream;
+ float samples[8][512 * 64];
+ int offset[4];
+} XMADecodeCtx;
/**
*@brief helper function to print the most important members of the context
@@ -250,7 +257,7 @@ static av_cold void dump_context(WMAProDecodeCtx *s)
PRINT("log2 frame size", s->log2_frame_size);
PRINT("max num subframes", s->max_num_subframes);
PRINT("len prefix", s->len_prefix);
- PRINT("num channels", s->avctx->channels);
+ PRINT("num channels", s->nb_channels);
}
/**
@@ -258,9 +265,8 @@ static av_cold void dump_context(WMAProDecodeCtx *s)
*@param avctx codec context
*@return 0 on success, < 0 otherwise
*/
-static av_cold int decode_end(AVCodecContext *avctx)
+static av_cold int decode_end(WMAProDecodeCtx *s)
{
- WMAProDecodeCtx *s = avctx->priv_data;
int i;
av_freep(&s->fdsp);
@@ -271,6 +277,15 @@ static av_cold int decode_end(AVCodecContext *avctx)
return 0;
}
+static av_cold int wmapro_decode_end(AVCodecContext *avctx)
+{
+ WMAProDecodeCtx *s = avctx->priv_data;
+
+ decode_end(s);
+
+ return 0;
+}
+
static av_cold int get_rate(AVCodecContext *avctx)
{
if (avctx->codec_id != AV_CODEC_ID_WMAPRO) { // XXX: is this really only for XMA?
@@ -291,9 +306,8 @@ static av_cold int get_rate(AVCodecContext *avctx)
*@param avctx codec context
*@return 0 on success, -1 otherwise
*/
-static av_cold int decode_init(AVCodecContext *avctx)
+static av_cold int decode_init(WMAProDecodeCtx *s, AVCodecContext *avctx)
{
- WMAProDecodeCtx *s = avctx->priv_data;
uint8_t *edata_ptr = avctx->extradata;
unsigned int channel_mask;
int i, bits;
@@ -326,7 +340,6 @@ static av_cold int decode_init(AVCodecContext *avctx)
s->decode_flags = 0x10d6;
channel_mask = avctx->extradata ? AV_RL32(edata_ptr+2) : 0;
s->bits_per_sample = 16;
-
} else if (avctx->codec_id == AV_CODEC_ID_XMA1) {
s->decode_flags = 0x10d6;
s->bits_per_sample = 16;
@@ -346,8 +359,9 @@ static av_cold int decode_init(AVCodecContext *avctx)
}
if (avctx->codec_id != AV_CODEC_ID_WMAPRO && avctx->channels > 2) {
- avpriv_report_missing_feature(avctx, ">2 channels support");
- return AVERROR_PATCHWELCOME;
+ s->nb_channels = 2;
+ } else {
+ s->nb_channels = avctx->channels;
}
/** generic init */
@@ -406,18 +420,18 @@ static av_cold int decode_init(AVCodecContext *avctx)
return AVERROR_INVALIDDATA;
}
- if (avctx->channels < 0) {
+ if (s->nb_channels <= 0) {
av_log(avctx, AV_LOG_ERROR, "invalid number of channels %d\n",
- avctx->channels);
+ s->nb_channels);
return AVERROR_INVALIDDATA;
- } else if (avctx->channels > WMAPRO_MAX_CHANNELS) {
+ } else if (s->nb_channels > WMAPRO_MAX_CHANNELS) {
avpriv_request_sample(avctx,
"More than %d channels", WMAPRO_MAX_CHANNELS);
return AVERROR_PATCHWELCOME;
}
/** init previous block len */
- for (i = 0; i < avctx->channels; i++)
+ for (i = 0; i < s->nb_channels; i++)
s->channel[i].prev_block_len = s->samples_per_frame;
/** extract lfe channel position */
@@ -543,6 +557,18 @@ static av_cold int decode_init(AVCodecContext *avctx)
}
/**
+ *@brief Initialize the decoder.
+ *@param avctx codec context
+ *@return 0 on success, -1 otherwise
+ */
+static av_cold int wmapro_decode_init(AVCodecContext *avctx)
+{
+ WMAProDecodeCtx *s = avctx->priv_data;
+
+ return decode_init(s, avctx);
+}
+
+/**
*@brief Decode the subframe length.
*@param s context
*@param offset sample offset in the frame
@@ -603,7 +629,7 @@ static int decode_tilehdr(WMAProDecodeCtx *s)
{
uint16_t num_samples[WMAPRO_MAX_CHANNELS] = { 0 };/**< sum of samples for all currently known subframes of a channel */
uint8_t contains_subframe[WMAPRO_MAX_CHANNELS]; /**< flag indicating if a channel contains the current subframe */
- int channels_for_cur_subframe = s->avctx->channels; /**< number of channels that contain the current subframe */
+ int channels_for_cur_subframe = s->nb_channels; /**< number of channels that contain the current subframe */
int fixed_channel_layout = 0; /**< flag indicating that all channels use the same subframe offsets and sizes */
int min_channel_len = 0; /**< smallest sum of samples (channels with this length will be processed first) */
int c;
@@ -615,7 +641,7 @@ static int decode_tilehdr(WMAProDecodeCtx *s)
*/
/** reset tiling information */
- for (c = 0; c < s->avctx->channels; c++)
+ for (c = 0; c < s->nb_channels; c++)
s->channel[c].num_subframes = 0;
if (s->max_num_subframes == 1 || get_bits1(&s->gb))
@@ -626,7 +652,7 @@ static int decode_tilehdr(WMAProDecodeCtx *s)
int subframe_len;
/** check which channels contain the subframe */
- for (c = 0; c < s->avctx->channels; c++) {
+ for (c = 0; c < s->nb_channels; c++) {
if (num_samples[c] == min_channel_len) {
if (fixed_channel_layout || channels_for_cur_subframe == 1 ||
(min_channel_len == s->samples_per_frame - s->min_samples_per_subframe))
@@ -643,7 +669,7 @@ static int decode_tilehdr(WMAProDecodeCtx *s)
/** add subframes to the individual channels and find new min_channel_len */
min_channel_len += subframe_len;
- for (c = 0; c < s->avctx->channels; c++) {
+ for (c = 0; c < s->nb_channels; c++) {
WMAProChannelCtx* chan = &s->channel[c];
if (contains_subframe[c]) {
@@ -670,7 +696,7 @@ static int decode_tilehdr(WMAProDecodeCtx *s)
}
} while (min_channel_len < s->samples_per_frame);
- for (c = 0; c < s->avctx->channels; c++) {
+ for (c = 0; c < s->nb_channels; c++) {
int i;
int offset = 0;
for (i = 0; i < s->channel[c].num_subframes; i++) {
@@ -696,8 +722,8 @@ static void decode_decorrelation_matrix(WMAProDecodeCtx *s,
int i;
int offset = 0;
int8_t rotation_offset[WMAPRO_MAX_CHANNELS * WMAPRO_MAX_CHANNELS];
- memset(chgroup->decorrelation_matrix, 0, s->avctx->channels *
- s->avctx->channels * sizeof(*chgroup->decorrelation_matrix));
+ memset(chgroup->decorrelation_matrix, 0, s->nb_channels *
+ s->nb_channels * sizeof(*chgroup->decorrelation_matrix));
for (i = 0; i < chgroup->num_channels * (chgroup->num_channels - 1) >> 1; i++)
rotation_offset[i] = get_bits(&s->gb, 6);
@@ -750,7 +776,7 @@ static int decode_channel_transform(WMAProDecodeCtx* s)
/** in the one channel case channel transforms are pointless */
s->num_chgroups = 0;
- if (s->avctx->channels > 1) {
+ if (s->nb_channels > 1) {
int remaining_channels = s->channels_for_cur_subframe;
if (get_bits1(&s->gb)) {
@@ -797,7 +823,7 @@ static int decode_channel_transform(WMAProDecodeCtx* s)
}
} else {
chgroup->transform = 1;
- if (s->avctx->channels == 2) {
+ if (s->nb_channels == 2) {
chgroup->decorrelation_matrix[0] = 1.0;
chgroup->decorrelation_matrix[1] = -1.0;
chgroup->decorrelation_matrix[2] = 1.0;
@@ -1087,7 +1113,7 @@ static void inverse_channel_transform(WMAProDecodeCtx *s)
(*ch)[y] = sum;
}
}
- } else if (s->avctx->channels == 2) {
+ } else if (s->nb_channels == 2) {
int len = FFMIN(sfb[1], s->subframe_len) - sfb[0];
s->fdsp->vector_fmul_scalar(ch_data[0] + sfb[0],
ch_data[0] + sfb[0],
@@ -1140,7 +1166,7 @@ static int decode_subframe(WMAProDecodeCtx *s)
int offset = s->samples_per_frame;
int subframe_len = s->samples_per_frame;
int i;
- int total_samples = s->samples_per_frame * s->avctx->channels;
+ int total_samples = s->samples_per_frame * s->nb_channels;
int transmit_coeffs = 0;
int cur_subwoofer_cutoff;
@@ -1150,7 +1176,7 @@ static int decode_subframe(WMAProDecodeCtx *s)
== the next block of the channel with the smallest number of
decoded samples
*/
- for (i = 0; i < s->avctx->channels; i++) {
+ for (i = 0; i < s->nb_channels; i++) {
s->channel[i].grouped = 0;
if (offset > s->channel[i].decoded_samples) {
offset = s->channel[i].decoded_samples;
@@ -1164,7 +1190,7 @@ static int decode_subframe(WMAProDecodeCtx *s)
/** get a list of all channels that contain the estimated block */
s->channels_for_cur_subframe = 0;
- for (i = 0; i < s->avctx->channels; i++) {
+ for (i = 0; i < s->nb_channels; i++) {
const int cur_subframe = s->channel[i].cur_subframe;
/** subtract already processed samples */
total_samples -= s->channel[i].decoded_samples;
@@ -1377,11 +1403,10 @@ static int decode_subframe(WMAProDecodeCtx *s)
*/
static int decode_frame(WMAProDecodeCtx *s, AVFrame *frame, int *got_frame_ptr)
{
- AVCodecContext *avctx = s->avctx;
GetBitContext* gb = &s->gb;
int more_frames = 0;
int len = 0;
- int i, ret;
+ int i;
/** get frame length */
if (s->len_prefix)
@@ -1396,9 +1421,9 @@ static int decode_frame(WMAProDecodeCtx *s, AVFrame *frame, int *got_frame_ptr)
}
/** read postproc transform */
- if (s->avctx->channels > 1 && get_bits1(gb)) {
+ if (s->nb_channels > 1 && get_bits1(gb)) {
if (get_bits1(gb)) {
- for (i = 0; i < avctx->channels * avctx->channels; i++)
+ for (i = 0; i < s->nb_channels * s->nb_channels; i++)
skip_bits(gb, 4);
}
}
@@ -1433,7 +1458,7 @@ static int decode_frame(WMAProDecodeCtx *s, AVFrame *frame, int *got_frame_ptr)
/** reset subframe states */
s->parsed_all_subframes = 0;
- for (i = 0; i < avctx->channels; i++) {
+ for (i = 0; i < s->nb_channels; i++) {
s->channel[i].decoded_samples = 0;
s->channel[i].cur_subframe = 0;
s->channel[i].reuse_sf = 0;
@@ -1447,19 +1472,12 @@ static int decode_frame(WMAProDecodeCtx *s, AVFrame *frame, int *got_frame_ptr)
}
}
- /* get output buffer */
- frame->nb_samples = s->samples_per_frame;
- if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
- s->packet_loss = 1;
- return 0;
- }
-
/** copy samples to the output buffer */
- for (i = 0; i < avctx->channels; i++)
+ for (i = 0; i < s->nb_channels; i++)
memcpy(frame->extended_data[i], s->channel[i].out,
s->samples_per_frame * sizeof(*s->channel[i].out));
- for (i = 0; i < avctx->channels; i++) {
+ for (i = 0; i < s->nb_channels; i++) {
/** reuse second half of the IMDCT output for the next frame */
memcpy(&s->channel[i].out[0],
&s->channel[i].out[s->samples_per_frame],
@@ -1564,17 +1582,9 @@ static void save_bits(WMAProDecodeCtx *s, GetBitContext* gb, int len,
skip_bits(&s->gb, s->frame_offset);
}
-/**
- *@brief Decode a single WMA packet.
- *@param avctx codec context
- *@param data the output buffer
- *@param avpkt input packet
- *@return number of bytes that were read from the input buffer
- */
-static int decode_packet(AVCodecContext *avctx, void *data,
- int *got_frame_ptr, AVPacket* avpkt)
+static int decode_packet(AVCodecContext *avctx, WMAProDecodeCtx *s,
+ void *data, int *got_frame_ptr, AVPacket *avpkt)
{
- WMAProDecodeCtx *s = avctx->priv_data;
GetBitContext* gb = &s->pgb;
const uint8_t* buf = avpkt->data;
int buf_size = avpkt->size;
@@ -1583,11 +1593,6 @@ static int decode_packet(AVCodecContext *avctx, void *data,
*got_frame_ptr = 0;
- if (s->skip_packets > 0) {
- s->skip_packets--;
- return FFMIN(avpkt->size, avctx->block_align);
- }
-
if (s->packet_done || s->packet_loss) {
s->packet_done = 0;
@@ -1613,7 +1618,8 @@ static int decode_packet(AVCodecContext *avctx, void *data,
packet_sequence_number = get_bits(gb, 4);
skip_bits(gb, 2);
} else {
- s->num_frames = get_bits(gb, 6);
+ int num_frames = get_bits(gb, 6);
+ ff_dlog(avctx, "packet[%d]: number of frames %d\n", avctx->frame_number, num_frames);
packet_sequence_number = 0;
}
@@ -1622,6 +1628,7 @@ static int decode_packet(AVCodecContext *avctx, void *data,
if (avctx->codec_id != AV_CODEC_ID_WMAPRO) {
skip_bits(gb, 3);
s->skip_packets = get_bits(gb, 8);
+ ff_dlog(avctx, "packet[%d]: skip packets %d\n", avctx->frame_number, s->skip_packets);
}
ff_dlog(avctx, "packet[%d]: nbpf %x\n", avctx->frame_number,
@@ -1665,7 +1672,6 @@ static int decode_packet(AVCodecContext *avctx, void *data,
s->num_saved_bits = 0;
s->packet_loss = 0;
}
-
} else {
int frame_size;
s->buf_bit_size = (avpkt->size - s->next_packet_start) << 3;
@@ -1687,8 +1693,9 @@ static int decode_packet(AVCodecContext *avctx, void *data,
the "previous frame" data from the next packet so that
we get a buffer that only contains full frames */
s->packet_done = !decode_frame(s, data, got_frame_ptr);
- } else
+ } else {
s->packet_done = 1;
+ }
}
if (remaining_bits(s, gb) < 0) {
@@ -1711,6 +1718,129 @@ static int decode_packet(AVCodecContext *avctx, void *data,
}
/**
+ *@brief Decode a single WMA packet.
+ *@param avctx codec context
+ *@param data the output buffer
+ *@param avpkt input packet
+ *@return number of bytes that were read from the input buffer
+ */
+static int wmapro_decode_packet(AVCodecContext *avctx, void *data,
+ int *got_frame_ptr, AVPacket *avpkt)
+{
+ WMAProDecodeCtx *s = avctx->priv_data;
+ AVFrame *frame = data;
+ int ret;
+
+ /* get output buffer */
+ frame->nb_samples = s->samples_per_frame;
+ if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
+ s->packet_loss = 1;
+ return 0;
+ }
+
+ return decode_packet(avctx, s, data, got_frame_ptr, avpkt);
+}
+
+static int xma_decode_packet(AVCodecContext *avctx, void *data,
+ int *got_frame_ptr, AVPacket *avpkt)
+{
+ XMADecodeCtx *s = avctx->priv_data;
+ int got_stream_frame_ptr = 0;
+ AVFrame *frame = data;
+ int i, ret, offset = INT_MAX;
+
+ ret = decode_packet(avctx, &s->xma[s->current_stream], s->frames[s->current_stream],
+ &got_stream_frame_ptr, avpkt);
+
+ if (got_stream_frame_ptr) {
+ memcpy(&s->samples[s->current_stream * 2 + 0][s->offset[s->current_stream] * 512],
+ s->frames[s->current_stream]->extended_data[0], 512 * 4);
+ memcpy(&s->samples[s->current_stream * 2 + 1][s->offset[s->current_stream] * 512],
+ s->frames[s->current_stream]->extended_data[1], 512 * 4);
+ s->offset[s->current_stream]++;
+ }
+
+ if (s->xma[s->current_stream].packet_done ||
+ s->xma[s->current_stream].packet_loss) {
+ int bret;
+
+ if (s->xma[0].skip_packets == 0) {
+ s->current_stream = 0;
+ } else if (s->xma[1].skip_packets == 0) {
+ s->current_stream = 1;
+ } else if (s->xma[2].skip_packets == 0) {
+ s->current_stream = 2;
+ } else if (s->xma[3].skip_packets == 0) {
+ s->current_stream = 3;
+ } else {
+ s->current_stream++;
+ if (s->current_stream >= avctx->channels / 2)
+ s->current_stream = 0;
+ }
+ for (i = 0; i < avctx->channels / 2; i++) {
+ s->xma[i].skip_packets = FFMAX(0, s->xma[i].skip_packets - 1);
+ }
+
+ for (i = 0; i < avctx->channels / 2; i++) {
+ offset = FFMIN(offset, s->offset[i]);
+ }
+
+ if (offset > 0) {
+ frame->nb_samples = 512 * offset;
+ if ((bret = ff_get_buffer(avctx, frame, 0)) < 0)
+ return bret;
+
+ for (i = 0; i < avctx->channels / 2; i++) {
+ memcpy(frame->extended_data[i * 2 + 0], s->samples[i * 2 + 0], frame->nb_samples * 4);
+ memcpy(frame->extended_data[i * 2 + 1], s->samples[i * 2 + 1], frame->nb_samples * 4);
+ s->offset[i] -= offset;
+ if (s->offset[i]) {
+ memmove(s->samples[i * 2 + 0], s->samples[i * 2 + 0] + frame->nb_samples, s->offset[i] * 4 * 512);
+ memmove(s->samples[i * 2 + 1], s->samples[i * 2 + 1] + frame->nb_samples, s->offset[i] * 4 * 512);
+ }
+ }
+
+ *got_frame_ptr = 1;
+ }
+ }
+
+ return ret;
+}
+
+static av_cold int xma_decode_init(AVCodecContext *avctx)
+{
+ XMADecodeCtx *s = avctx->priv_data;
+ int i, ret;
+
+ for (i = 0; i < avctx->channels / 2; i++) {
+ ret = decode_init(&s->xma[i], avctx);
+ s->frames[i] = av_frame_alloc();
+ if (!s->frames[i])
+ return AVERROR(ENOMEM);
+ s->frames[i]->nb_samples = 512;
+ if ((ret = ff_get_buffer(avctx, s->frames[i], 0)) < 0) {
+ return AVERROR(ENOMEM);
+ }
+
+ }
+
+ return ret;
+}
+
+static av_cold int xma_decode_end(AVCodecContext *avctx)
+{
+ XMADecodeCtx *s = avctx->priv_data;
+ int i;
+
+ for (i = 0; i < avctx->channels / 2; i++) {
+ decode_end(&s->xma[i]);
+ av_frame_free(&s->frames[i]);
+ }
+
+ return 0;
+}
+
+/**
*@brief Clear decoder buffers (for seeking).
*@param avctx codec context
*/
@@ -1720,7 +1850,7 @@ static void flush(AVCodecContext *avctx)
int i;
/** reset output buffer as a part of it is used during the windowing of a
new frame */
- for (i = 0; i < avctx->channels; i++)
+ for (i = 0; i < s->nb_channels; i++)
memset(s->channel[i].out, 0, s->samples_per_frame *
sizeof(*s->channel[i].out));
s->packet_loss = 1;
@@ -1736,9 +1866,9 @@ AVCodec ff_wmapro_decoder = {
.type = AVMEDIA_TYPE_AUDIO,
.id = AV_CODEC_ID_WMAPRO,
.priv_data_size = sizeof(WMAProDecodeCtx),
- .init = decode_init,
- .close = decode_end,
- .decode = decode_packet,
+ .init = wmapro_decode_init,
+ .close = wmapro_decode_end,
+ .decode = wmapro_decode_packet,
.capabilities = AV_CODEC_CAP_SUBFRAMES | AV_CODEC_CAP_DR1,
.flush = flush,
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
@@ -1750,12 +1880,11 @@ AVCodec ff_xma1_decoder = {
.long_name = NULL_IF_CONFIG_SMALL("Xbox Media Audio 1"),
.type = AVMEDIA_TYPE_AUDIO,
.id = AV_CODEC_ID_XMA1,
- .priv_data_size = sizeof(WMAProDecodeCtx),
- .init = decode_init,
- .close = decode_end,
- .decode = decode_packet,
+ .priv_data_size = sizeof(XMADecodeCtx),
+ .init = xma_decode_init,
+ .close = xma_decode_end,
+ .decode = xma_decode_packet,
.capabilities = AV_CODEC_CAP_SUBFRAMES | AV_CODEC_CAP_DR1,
- .flush = flush,
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_NONE },
};
@@ -1765,12 +1894,11 @@ AVCodec ff_xma2_decoder = {
.long_name = NULL_IF_CONFIG_SMALL("Xbox Media Audio 2"),
.type = AVMEDIA_TYPE_AUDIO,
.id = AV_CODEC_ID_XMA2,
- .priv_data_size = sizeof(WMAProDecodeCtx),
- .init = decode_init,
- .close = decode_end,
- .decode = decode_packet,
+ .priv_data_size = sizeof(XMADecodeCtx),
+ .init = xma_decode_init,
+ .close = xma_decode_end,
+ .decode = xma_decode_packet,
.capabilities = AV_CODEC_CAP_SUBFRAMES | AV_CODEC_CAP_DR1,
- .flush = flush,
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_NONE },
};
diff --git a/libavformat/wavdec.c b/libavformat/wavdec.c
index 987155e39e..4046809cf6 100644
--- a/libavformat/wavdec.c
+++ b/libavformat/wavdec.c
@@ -567,6 +567,9 @@ break_loop:
st->codecpar->block_align == st->codecpar->channels * 4 &&
st->codecpar->bits_per_coded_sample == 24) {
st->codecpar->codec_id = AV_CODEC_ID_PCM_F24LE;
+ } else if (st->codecpar->codec_id == AV_CODEC_ID_XMA1 ||
+ st->codecpar->codec_id == AV_CODEC_ID_XMA2) {
+ st->codecpar->block_align = 2048;
}
ff_metadata_conv_ctx(s, NULL, wav_metadata_conv);