mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-25 11:11:08 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/gst/audioresample/gstaudioresample.c
Jan Schmidt 02a7b31f0e audioresample: Fix buffer size transformations 
When calculating the input/output buffer sizes in the transform_size function,
take the number of channels into account, so we don't end up calculating
a buffer size that only contains a partial number of audio frames.

Also, when going from output size to input size, round down rather than
up, so as to calculate the minimum number of samples that *might* yield
a buffer of the intended destination size.

Fixes: #580470 and #580952
2009-05-01 16:47:53 +01:00

1443 lines
42 KiB
C
Raw Blame History

/* GStreamer
* Copyright (C) 1999 Erik Walthinsen <omega@cse.ogi.edu>
* Copyright (C) 2003,2004 David A. Schleef <ds@schleef.org>
* Copyright (C) 2007-2008 Sebastian Dröge <sebastian.droege@collabora.co.uk>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library 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
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 02111-1307, USA.
*/
/**
* SECTION:element-audioresample
*
* audioresample resamples raw audio buffers to different sample rates using
* a configurable windowing function to enhance quality.
*
* <refsect2>
* <title>Example launch line</title>
* |[
* gst-launch -v filesrc location=sine.ogg ! oggdemux ! vorbisdec ! audioconvert ! audioresample ! audio/x-raw-int, rate=8000 ! alsasink
* ]| Decode an Ogg/Vorbis downsample to 8Khz and play sound through alsa.
* To create the Ogg/Vorbis file refer to the documentation of vorbisenc.
* </refsect2>
*/
/* TODO:
* - Enable SSE/ARM optimizations and select at runtime
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <string.h>
#include <math.h>
#include "gstaudioresample.h"
#include <gst/audio/audio.h>
#include <gst/base/gstbasetransform.h>
#if defined AUDIORESAMPLE_FORMAT_AUTO
#define OIL_ENABLE_UNSTABLE_API
#include <liboil/liboilprofile.h>
#include <liboil/liboil.h>
#endif
GST_DEBUG_CATEGORY (audio_resample_debug);
#define GST_CAT_DEFAULT audio_resample_debug
enum
{
PROP_0,
PROP_QUALITY,
PROP_FILTER_LENGTH
};
#define SUPPORTED_CAPS \
GST_STATIC_CAPS ( \
"audio/x-raw-float, " \
"rate = (int) [ 1, MAX ], " \
"channels = (int) [ 1, MAX ], " \
"endianness = (int) BYTE_ORDER, " \
"width = (int) { 32, 64 }; " \
"audio/x-raw-int, " \
"rate = (int) [ 1, MAX ], " \
"channels = (int) [ 1, MAX ], " \
"endianness = (int) BYTE_ORDER, " \
"width = (int) 32, " \
"depth = (int) 32, " \
"signed = (boolean) true; " \
"audio/x-raw-int, " \
"rate = (int) [ 1, MAX ], " \
"channels = (int) [ 1, MAX ], " \
"endianness = (int) BYTE_ORDER, " \
"width = (int) 24, " \
"depth = (int) 24, " \
"signed = (boolean) true; " \
"audio/x-raw-int, " \
"rate = (int) [ 1, MAX ], " \
"channels = (int) [ 1, MAX ], " \
"endianness = (int) BYTE_ORDER, " \
"width = (int) 16, " \
"depth = (int) 16, " \
"signed = (boolean) true; " \
"audio/x-raw-int, " \
"rate = (int) [ 1, MAX ], " \
"channels = (int) [ 1, MAX ], " \
"endianness = (int) BYTE_ORDER, " \
"width = (int) 8, " \
"depth = (int) 8, " \
"signed = (boolean) true" \
)
/* If TRUE integer arithmetic resampling is faster and will be used if appropiate */
#if defined AUDIORESAMPLE_FORMAT_INT
static gboolean gst_audio_resample_use_int = TRUE;
#elif defined AUDIORESAMPLE_FORMAT_FLOAT
static gboolean gst_audio_resample_use_int = FALSE;
#else
static gboolean gst_audio_resample_use_int = FALSE;
#endif
static GstStaticPadTemplate gst_audio_resample_sink_template =
GST_STATIC_PAD_TEMPLATE ("sink",
GST_PAD_SINK, GST_PAD_ALWAYS, SUPPORTED_CAPS);
static GstStaticPadTemplate gst_audio_resample_src_template =
GST_STATIC_PAD_TEMPLATE ("src",
GST_PAD_SRC, GST_PAD_ALWAYS, SUPPORTED_CAPS);
static void gst_audio_resample_set_property (GObject * object,
guint prop_id, const GValue * value, GParamSpec * pspec);
static void gst_audio_resample_get_property (GObject * object,
guint prop_id, GValue * value, GParamSpec * pspec);
/* vmethods */
static gboolean gst_audio_resample_get_unit_size (GstBaseTransform * base,
GstCaps * caps, guint * size);
static GstCaps *gst_audio_resample_transform_caps (GstBaseTransform * base,
GstPadDirection direction, GstCaps * caps);
static void gst_audio_resample_fixate_caps (GstBaseTransform * base,
GstPadDirection direction, GstCaps * caps, GstCaps * othercaps);
static gboolean gst_audio_resample_transform_size (GstBaseTransform * trans,
GstPadDirection direction, GstCaps * incaps, guint insize,
GstCaps * outcaps, guint * outsize);
static gboolean gst_audio_resample_set_caps (GstBaseTransform * base,
GstCaps * incaps, GstCaps * outcaps);
static GstFlowReturn gst_audio_resample_transform (GstBaseTransform * base,
GstBuffer * inbuf, GstBuffer * outbuf);
static gboolean gst_audio_resample_event (GstBaseTransform * base,
GstEvent * event);
static gboolean gst_audio_resample_start (GstBaseTransform * base);
static gboolean gst_audio_resample_stop (GstBaseTransform * base);
static gboolean gst_audio_resample_query (GstPad * pad, GstQuery * query);
static const GstQueryType *gst_audio_resample_query_type (GstPad * pad);
GST_BOILERPLATE (GstAudioResample, gst_audio_resample, GstBaseTransform,
GST_TYPE_BASE_TRANSFORM);
static void
gst_audio_resample_base_init (gpointer g_class)
{
GstElementClass *gstelement_class = GST_ELEMENT_CLASS (g_class);
gst_element_class_add_pad_template (gstelement_class,
gst_static_pad_template_get (&gst_audio_resample_src_template));
gst_element_class_add_pad_template (gstelement_class,
gst_static_pad_template_get (&gst_audio_resample_sink_template));
gst_element_class_set_details_simple (gstelement_class, "Audio resampler",
"Filter/Converter/Audio", "Resamples audio",
"Sebastian Dröge <sebastian.droege@collabora.co.uk>");
}
static void
gst_audio_resample_class_init (GstAudioResampleClass * klass)
{
GObjectClass *gobject_class = (GObjectClass *) klass;
gobject_class->set_property = gst_audio_resample_set_property;
gobject_class->get_property = gst_audio_resample_get_property;
g_object_class_install_property (gobject_class, PROP_QUALITY,
g_param_spec_int ("quality", "Quality", "Resample quality with 0 being "
"the lowest and 10 being the best",
SPEEX_RESAMPLER_QUALITY_MIN, SPEEX_RESAMPLER_QUALITY_MAX,
SPEEX_RESAMPLER_QUALITY_DEFAULT,
G_PARAM_READWRITE | G_PARAM_CONSTRUCT));
/* FIXME 0.11: Remove this property, it's just for compatibility
* with old audioresample
*/
/**
* GstAudioResample:filter-length:
*
* Length of the resample filter
*
* Deprectated: Use #GstAudioResample:quality property instead
*/
g_object_class_install_property (gobject_class, PROP_FILTER_LENGTH,
g_param_spec_int ("filter-length", "Filter length",
"Length of the resample filter", 0, G_MAXINT, 64, G_PARAM_READWRITE));
GST_BASE_TRANSFORM_CLASS (klass)->start =
GST_DEBUG_FUNCPTR (gst_audio_resample_start);
GST_BASE_TRANSFORM_CLASS (klass)->stop =
GST_DEBUG_FUNCPTR (gst_audio_resample_stop);
GST_BASE_TRANSFORM_CLASS (klass)->transform_size =
GST_DEBUG_FUNCPTR (gst_audio_resample_transform_size);
GST_BASE_TRANSFORM_CLASS (klass)->get_unit_size =
GST_DEBUG_FUNCPTR (gst_audio_resample_get_unit_size);
GST_BASE_TRANSFORM_CLASS (klass)->transform_caps =
GST_DEBUG_FUNCPTR (gst_audio_resample_transform_caps);
GST_BASE_TRANSFORM_CLASS (klass)->fixate_caps =
GST_DEBUG_FUNCPTR (gst_audio_resample_fixate_caps);
GST_BASE_TRANSFORM_CLASS (klass)->set_caps =
GST_DEBUG_FUNCPTR (gst_audio_resample_set_caps);
GST_BASE_TRANSFORM_CLASS (klass)->transform =
GST_DEBUG_FUNCPTR (gst_audio_resample_transform);
GST_BASE_TRANSFORM_CLASS (klass)->event =
GST_DEBUG_FUNCPTR (gst_audio_resample_event);
GST_BASE_TRANSFORM_CLASS (klass)->passthrough_on_same_caps = TRUE;
}
static void
gst_audio_resample_init (GstAudioResample * resample,
GstAudioResampleClass * klass)
{
GstBaseTransform *trans = GST_BASE_TRANSFORM (resample);
resample->quality = SPEEX_RESAMPLER_QUALITY_DEFAULT;
resample->need_discont = FALSE;
gst_pad_set_query_function (trans->srcpad, gst_audio_resample_query);
gst_pad_set_query_type_function (trans->srcpad,
gst_audio_resample_query_type);
}
/* vmethods */
static gboolean
gst_audio_resample_start (GstBaseTransform * base)
{
GstAudioResample *resample = GST_AUDIO_RESAMPLE (base);
resample->next_offset = -1;
resample->next_ts = -1;
resample->next_upstream_ts = -1;
return TRUE;
}
static gboolean
gst_audio_resample_stop (GstBaseTransform * base)
{
GstAudioResample *resample = GST_AUDIO_RESAMPLE (base);
if (resample->state) {
resample->funcs->destroy (resample->state);
resample->state = NULL;
}
resample->funcs = NULL;
g_free (resample->tmp_in);
resample->tmp_in = NULL;
resample->tmp_in_size = 0;
g_free (resample->tmp_out);
resample->tmp_out = NULL;
resample->tmp_out_size = 0;
gst_caps_replace (&resample->sinkcaps, NULL);
gst_caps_replace (&resample->srccaps, NULL);
return TRUE;
}
static gboolean
gst_audio_resample_get_unit_size (GstBaseTransform * base, GstCaps * caps,
guint * size)
{
gint width, channels;
GstStructure *structure;
gboolean ret;
g_return_val_if_fail (size != NULL, FALSE);
/* this works for both float and int */
structure = gst_caps_get_structure (caps, 0);
ret = gst_structure_get_int (structure, "width", &width);
ret &= gst_structure_get_int (structure, "channels", &channels);
if (G_UNLIKELY (!ret))
return FALSE;
*size = (width / 8) * channels;
return TRUE;
}
static GstCaps *
gst_audio_resample_transform_caps (GstBaseTransform * base,
GstPadDirection direction, GstCaps * caps)
{
const GValue *val;
GstStructure *s;
GstCaps *res;
/* transform single caps into input_caps + input_caps with the rate
* field set to our supported range. This ensures that upstream knows
* about downstream's prefered rate(s) and can negotiate accordingly. */
res = gst_caps_copy (caps);
/* first, however, check if the caps contain a range for the rate field, in
* which case that side isn't going to care much about the exact sample rate
* chosen and we should just assume things will get fixated to something sane
* and we may just as well offer our full range instead of the range in the
* caps. If the rate is not an int range value, it's likely to express a
* real preference or limitation and we should maintain that structure as
* preference by putting it first into the transformed caps, and only add
* our full rate range as second option */
s = gst_caps_get_structure (res, 0);
val = gst_structure_get_value (s, "rate");
if (val == NULL || GST_VALUE_HOLDS_INT_RANGE (val)) {
/* overwrite existing range, or add field if it doesn't exist yet */
gst_structure_set (s, "rate", GST_TYPE_INT_RANGE, 1, G_MAXINT, NULL);
} else {
/* append caps with full range to existing caps with non-range rate field */
s = gst_structure_copy (s);
gst_structure_set (s, "rate", GST_TYPE_INT_RANGE, 1, G_MAXINT, NULL);
gst_caps_append_structure (res, s);
}
return res;
}
/* Fixate rate to the allowed rate that has the smallest difference */
static void
gst_audio_resample_fixate_caps (GstBaseTransform * base,
GstPadDirection direction, GstCaps * caps, GstCaps * othercaps)
{
GstStructure *s;
gint rate;
s = gst_caps_get_structure (caps, 0);
if (G_UNLIKELY (!gst_structure_get_int (s, "rate", &rate)))
return;
s = gst_caps_get_structure (othercaps, 0);
gst_structure_fixate_field_nearest_int (s, "rate", rate);
}
static const SpeexResampleFuncs *
gst_audio_resample_get_funcs (gint width, gboolean fp)
{
const SpeexResampleFuncs *funcs = NULL;
if (gst_audio_resample_use_int && (width == 8 || width == 16) && !fp)
funcs = &int_funcs;
else if ((!gst_audio_resample_use_int && (width == 8 || width == 16) && !fp)
|| (width == 32 && fp))
funcs = &float_funcs;
else if ((width == 64 && fp) || ((width == 32 || width == 24) && !fp))
funcs = &double_funcs;
else
g_assert_not_reached ();
return funcs;
}
static SpeexResamplerState *
gst_audio_resample_init_state (GstAudioResample * resample, gint width,
gint channels, gint inrate, gint outrate, gint quality, gboolean fp)
{
SpeexResamplerState *ret = NULL;
gint err = RESAMPLER_ERR_SUCCESS;
const SpeexResampleFuncs *funcs = gst_audio_resample_get_funcs (width, fp);
ret = funcs->init (channels, inrate, outrate, quality, &err);
if (G_UNLIKELY (err != RESAMPLER_ERR_SUCCESS)) {
GST_ERROR_OBJECT (resample, "Failed to create resampler state: %s",
funcs->strerror (err));
return NULL;
}
funcs->skip_zeros (ret);
return ret;
}
static gboolean
gst_audio_resample_update_state (GstAudioResample * resample, gint width,
gint channels, gint inrate, gint outrate, gint quality, gboolean fp)
{
gboolean ret = TRUE;
gboolean updated_latency = FALSE;
updated_latency = (resample->inrate != inrate
|| quality != resample->quality) && resample->state != NULL;
if (resample->state == NULL) {
ret = TRUE;
} else if (resample->channels != channels || fp != resample->fp
|| width != resample->width) {
resample->funcs->destroy (resample->state);
resample->state =
gst_audio_resample_init_state (resample, width, channels, inrate,
outrate, quality, fp);
resample->funcs = gst_audio_resample_get_funcs (width, fp);
ret = (resample->state != NULL);
} else if (resample->inrate != inrate || resample->outrate != outrate) {
gint err = RESAMPLER_ERR_SUCCESS;
err = resample->funcs->set_rate (resample->state, inrate, outrate);
if (G_UNLIKELY (err != RESAMPLER_ERR_SUCCESS))
GST_ERROR_OBJECT (resample, "Failed to update rate: %s",
resample->funcs->strerror (err));
ret = (err == RESAMPLER_ERR_SUCCESS);
} else if (quality != resample->quality) {
gint err = RESAMPLER_ERR_SUCCESS;
err = resample->funcs->set_quality (resample->state, quality);
if (G_UNLIKELY (err != RESAMPLER_ERR_SUCCESS))
GST_ERROR_OBJECT (resample, "Failed to update quality: %s",
resample->funcs->strerror (err));
ret = (err == RESAMPLER_ERR_SUCCESS);
}
resample->width = width;
resample->channels = channels;
resample->fp = fp;
resample->quality = quality;
resample->inrate = inrate;
resample->outrate = outrate;
if (updated_latency)
gst_element_post_message (GST_ELEMENT (resample),
gst_message_new_latency (GST_OBJECT (resample)));
return ret;
}
static void
gst_audio_resample_reset_state (GstAudioResample * resample)
{
if (resample->state)
resample->funcs->reset_mem (resample->state);
}
static gboolean
gst_audio_resample_parse_caps (GstCaps * incaps,
GstCaps * outcaps, gint * width, gint * channels, gint * inrate,
gint * outrate, gboolean * fp)
{
GstStructure *structure;
gboolean ret;
gint mywidth, myinrate, myoutrate, mychannels;
gboolean myfp;
GST_DEBUG ("incaps %" GST_PTR_FORMAT ", outcaps %"
GST_PTR_FORMAT, incaps, outcaps);
structure = gst_caps_get_structure (incaps, 0);
if (g_str_equal (gst_structure_get_name (structure), "audio/x-raw-float"))
myfp = TRUE;
else
myfp = FALSE;
ret = gst_structure_get_int (structure, "rate", &myinrate);
ret &= gst_structure_get_int (structure, "channels", &mychannels);
ret &= gst_structure_get_int (structure, "width", &mywidth);
if (G_UNLIKELY (!ret))
goto no_in_rate_channels;
structure = gst_caps_get_structure (outcaps, 0);
ret = gst_structure_get_int (structure, "rate", &myoutrate);
if (G_UNLIKELY (!ret))
goto no_out_rate;
if (channels)
*channels = mychannels;
if (inrate)
*inrate = myinrate;
if (outrate)
*outrate = myoutrate;
if (width)
*width = mywidth;
if (fp)
*fp = myfp;
return TRUE;
/* ERRORS */
no_in_rate_channels:
{
GST_DEBUG ("could not get input rate and channels");
return FALSE;
}
no_out_rate:
{
GST_DEBUG ("could not get output rate");
return FALSE;
}
}
static gint
_gcd (gint a, gint b)
{
while (b != 0) {
int temp = a;
a = b;
b = temp % b;
}
return ABS (a);
}
static gboolean
gst_audio_resample_transform_size (GstBaseTransform * base,
GstPadDirection direction, GstCaps * caps, guint size, GstCaps * othercaps,
guint * othersize)
{
GstCaps *srccaps, *sinkcaps;
gboolean ret = TRUE;
guint32 ratio_den, ratio_num;
gint inrate, outrate, gcd;
gint bytes_per_samp, channels;
GST_LOG_OBJECT (base, "asked to transform size %d in direction %s",
size, direction == GST_PAD_SINK ? "SINK" : "SRC");
if (direction == GST_PAD_SINK) {
sinkcaps = caps;
srccaps = othercaps;
} else {
sinkcaps = othercaps;
srccaps = caps;
}
/* Get sample width -> bytes_per_samp, channels, inrate, outrate */
ret =
gst_audio_resample_parse_caps (caps, othercaps, &bytes_per_samp,
&channels, &inrate, &outrate, NULL);
if (G_UNLIKELY (!ret)) {
GST_ERROR_OBJECT (base, "Wrong caps");
return FALSE;
}
/* Number of samples in either buffer is size / (width*channels) ->
* calculate the factor */
bytes_per_samp = bytes_per_samp * channels / 8;
/* Convert source buffer size to samples */
size /= bytes_per_samp;
/* Simplify the conversion ratio factors */
gcd = _gcd (inrate, outrate);
ratio_num = inrate / gcd;
ratio_den = outrate / gcd;
if (direction == GST_PAD_SINK) {
/* asked to convert size of an incoming buffer. Round up the output size */
*othersize = (size * ratio_den + ratio_num - 1) / ratio_num;
*othersize *= bytes_per_samp;
} else {
/* asked to convert size of an outgoing buffer. Round down the input size */
*othersize = (size * ratio_num) / ratio_den;
*othersize *= bytes_per_samp;
}
GST_LOG_OBJECT (base, "transformed size %d to %d", size * bytes_per_samp,
*othersize);
return ret;
}
static gboolean
gst_audio_resample_set_caps (GstBaseTransform * base, GstCaps * incaps,
GstCaps * outcaps)
{
gboolean ret;
gint width = 0, inrate = 0, outrate = 0, channels = 0;
gboolean fp;
GstAudioResample *resample = GST_AUDIO_RESAMPLE (base);
GST_LOG ("incaps %" GST_PTR_FORMAT ", outcaps %"
GST_PTR_FORMAT, incaps, outcaps);
ret = gst_audio_resample_parse_caps (incaps, outcaps,
&width, &channels, &inrate, &outrate, &fp);
if (G_UNLIKELY (!ret))
return FALSE;
ret =
gst_audio_resample_update_state (resample, width, channels, inrate,
outrate, resample->quality, fp);
if (G_UNLIKELY (!ret))
return FALSE;
/* save caps so we can short-circuit in the size_transform if the caps
* are the same */
gst_caps_replace (&resample->sinkcaps, incaps);
gst_caps_replace (&resample->srccaps, outcaps);
return TRUE;
}
#define GST_MAXINT24 (8388607)
#define GST_MININT24 (-8388608)
#if (G_BYTE_ORDER == G_LITTLE_ENDIAN)
#define GST_READ_UINT24 GST_READ_UINT24_LE
#define GST_WRITE_UINT24 GST_WRITE_UINT24_LE
#else
#define GST_READ_UINT24 GST_READ_UINT24_BE
#define GST_WRITE_UINT24 GST_WRITE_UINT24_BE
#endif
static void
gst_audio_resample_convert_buffer (GstAudioResample * resample,
const guint8 * in, guint8 * out, guint len, gboolean inverse)
{
len *= resample->channels;
if (inverse) {
if (gst_audio_resample_use_int && resample->width == 8 && !resample->fp) {
gint8 *o = (gint8 *) out;
gint16 *i = (gint16 *) in;
gint32 tmp;
while (len) {
tmp = *i + (G_MAXINT8 >> 1);
*o = CLAMP (tmp >> 8, G_MININT8, G_MAXINT8);
o++;
i++;
len--;
}
} else if (!gst_audio_resample_use_int && resample->width == 8
&& !resample->fp) {
gint8 *o = (gint8 *) out;
gfloat *i = (gfloat *) in;
gfloat tmp;
while (len) {
tmp = *i;
*o = (gint8) CLAMP (tmp * G_MAXINT8 + 0.5, G_MININT8, G_MAXINT8);
o++;
i++;
len--;
}
} else if (!gst_audio_resample_use_int && resample->width == 16
&& !resample->fp) {
gint16 *o = (gint16 *) out;
gfloat *i = (gfloat *) in;
gfloat tmp;
while (len) {
tmp = *i;
*o = (gint16) CLAMP (tmp * G_MAXINT16 + 0.5, G_MININT16, G_MAXINT16);
o++;
i++;
len--;
}
} else if (resample->width == 24 && !resample->fp) {
guint8 *o = (guint8 *) out;
gdouble *i = (gdouble *) in;
gdouble tmp;
while (len) {
tmp = *i;
GST_WRITE_UINT24 (o, (gint32) CLAMP (tmp * GST_MAXINT24 + 0.5,
GST_MININT24, GST_MAXINT24));
o += 3;
i++;
len--;
}
} else if (resample->width == 32 && !resample->fp) {
gint32 *o = (gint32 *) out;
gdouble *i = (gdouble *) in;
gdouble tmp;
while (len) {
tmp = *i;
*o = (gint32) CLAMP (tmp * G_MAXINT32 + 0.5, G_MININT32, G_MAXINT32);
o++;
i++;
len--;
}
} else {
g_assert_not_reached ();
}
} else {
if (gst_audio_resample_use_int && resample->width == 8 && !resample->fp) {
gint8 *i = (gint8 *) in;
gint16 *o = (gint16 *) out;
gint32 tmp;
while (len) {
tmp = *i;
*o = tmp << 8;
o++;
i++;
len--;
}
} else if (!gst_audio_resample_use_int && resample->width == 8
&& !resample->fp) {
gint8 *i = (gint8 *) in;
gfloat *o = (gfloat *) out;
gfloat tmp;
while (len) {
tmp = *i;
*o = tmp / G_MAXINT8;
o++;
i++;
len--;
}
} else if (!gst_audio_resample_use_int && resample->width == 16
&& !resample->fp) {
gint16 *i = (gint16 *) in;
gfloat *o = (gfloat *) out;
gfloat tmp;
while (len) {
tmp = *i;
*o = tmp / G_MAXINT16;
o++;
i++;
len--;
}
} else if (resample->width == 24 && !resample->fp) {
guint8 *i = (guint8 *) in;
gdouble *o = (gdouble *) out;
gdouble tmp;
guint32 tmp2;
while (len) {
tmp2 = GST_READ_UINT24 (i);
if (tmp2 & 0x00800000)
tmp2 |= 0xff000000;
tmp = (gint32) tmp2;
*o = tmp / GST_MAXINT24;
o++;
i += 3;
len--;
}
} else if (resample->width == 32 && !resample->fp) {
gint32 *i = (gint32 *) in;
gdouble *o = (gdouble *) out;
gdouble tmp;
while (len) {
tmp = *i;
*o = tmp / G_MAXINT32;
o++;
i++;
len--;
}
} else {
g_assert_not_reached ();
}
}
}
static void
gst_audio_resample_push_drain (GstAudioResample * resample)
{
GstBuffer *buf;
GstBaseTransform *trans = GST_BASE_TRANSFORM (resample);
GstFlowReturn res;
gint outsize;
guint out_len, out_processed;
gint err;
guint num, den, len;
guint8 *outtmp = NULL;
gboolean need_convert = FALSE;
if (!resample->state)
return;
need_convert = (resample->funcs->width != resample->width);
resample->funcs->get_ratio (resample->state, &num, &den);
out_len = resample->funcs->get_input_latency (resample->state);
out_len = out_processed = (out_len * den + num - 1) / num;
outsize = (resample->width / 8) * out_len * resample->channels;
if (need_convert) {
guint outsize_tmp =
(resample->funcs->width / 8) * out_len * resample->channels;
if (outsize_tmp <= resample->tmp_out_size) {
outtmp = resample->tmp_out;
} else {
resample->tmp_out_size = outsize_tmp;
resample->tmp_out = outtmp = g_realloc (resample->tmp_out, outsize_tmp);
}
}
res =
gst_pad_alloc_buffer_and_set_caps (trans->srcpad, GST_BUFFER_OFFSET_NONE,
outsize, GST_PAD_CAPS (trans->srcpad), &buf);
if (G_UNLIKELY (res != GST_FLOW_OK)) {
GST_WARNING_OBJECT (resample, "failed allocating buffer of %d bytes",
outsize);
return;
}
len = resample->funcs->get_input_latency (resample->state);
err =
resample->funcs->process (resample->state,
NULL, &len, (need_convert) ? outtmp : GST_BUFFER_DATA (buf),
&out_processed);
if (G_UNLIKELY (err != RESAMPLER_ERR_SUCCESS)) {
GST_WARNING_OBJECT (resample, "Failed to process drain: %s",
resample->funcs->strerror (err));
gst_buffer_unref (buf);
return;
}
if (G_UNLIKELY (out_processed == 0)) {
GST_WARNING_OBJECT (resample, "Failed to get drain, dropping buffer");
gst_buffer_unref (buf);
return;
}
/* If we wrote more than allocated something is really wrong now
* and we should better abort immediately */
g_assert (out_len >= out_processed);
if (need_convert)
gst_audio_resample_convert_buffer (resample, outtmp, GST_BUFFER_DATA (buf),
out_processed, TRUE);
GST_BUFFER_DURATION (buf) =
GST_FRAMES_TO_CLOCK_TIME (out_processed, resample->outrate);
GST_BUFFER_SIZE (buf) =
out_processed * resample->channels * (resample->width / 8);
if (GST_CLOCK_TIME_IS_VALID (resample->next_ts)) {
GST_BUFFER_OFFSET (buf) = resample->next_offset;
GST_BUFFER_OFFSET_END (buf) = resample->next_offset + out_processed;
GST_BUFFER_TIMESTAMP (buf) = resample->next_ts;
resample->next_ts += GST_BUFFER_DURATION (buf);
resample->next_offset += out_processed;
}
GST_LOG_OBJECT (resample,
"Pushing drain buffer of %u bytes with timestamp %" GST_TIME_FORMAT
" duration %" GST_TIME_FORMAT " offset %" G_GUINT64_FORMAT " offset_end %"
G_GUINT64_FORMAT, GST_BUFFER_SIZE (buf),
GST_TIME_ARGS (GST_BUFFER_TIMESTAMP (buf)),
GST_TIME_ARGS (GST_BUFFER_DURATION (buf)), GST_BUFFER_OFFSET (buf),
GST_BUFFER_OFFSET_END (buf));
res = gst_pad_push (trans->srcpad, buf);
if (G_UNLIKELY (res != GST_FLOW_OK))
GST_WARNING_OBJECT (resample, "Failed to push drain: %s",
gst_flow_get_name (res));
return;
}
static gboolean
gst_audio_resample_event (GstBaseTransform * base, GstEvent * event)
{
GstAudioResample *resample = GST_AUDIO_RESAMPLE (base);
switch (GST_EVENT_TYPE (event)) {
case GST_EVENT_FLUSH_STOP:
gst_audio_resample_reset_state (resample);
resample->next_offset = -1;
resample->next_ts = -1;
resample->next_upstream_ts = -1;
break;
case GST_EVENT_NEWSEGMENT:
gst_audio_resample_push_drain (resample);
gst_audio_resample_reset_state (resample);
resample->next_offset = -1;
resample->next_ts = -1;
resample->next_upstream_ts = -1;
break;
case GST_EVENT_EOS:
gst_audio_resample_push_drain (resample);
gst_audio_resample_reset_state (resample);
break;
default:
break;
}
return parent_class->event (base, event);
}
static gboolean
gst_audio_resample_check_discont (GstAudioResample * resample,
GstClockTime timestamp)
{
if (timestamp != GST_CLOCK_TIME_NONE &&
resample->next_upstream_ts != GST_CLOCK_TIME_NONE &&
timestamp != resample->next_upstream_ts) {
/* Potentially a discontinuous buffer. However, it turns out that many
* elements generate imperfect streams due to rounding errors, so we permit
* a small error (up to one sample) without triggering a filter
* flush/restart (if triggered incorrectly, this will be audible) */
GstClockTimeDiff diff = timestamp - resample->next_upstream_ts;
if (ABS (diff) > (GST_SECOND + resample->inrate - 1) / resample->inrate) {
GST_WARNING_OBJECT (resample,
"encountered timestamp discontinuity of %s%" GST_TIME_FORMAT,
(diff < 0) ? "-" : "", GST_TIME_ARGS ((GstClockTime) ABS (diff)));
return TRUE;
}
}
return FALSE;
}
static GstFlowReturn
gst_audio_resample_process (GstAudioResample * resample, GstBuffer * inbuf,
GstBuffer * outbuf)
{
guint32 in_len, in_processed;
guint32 out_len, out_processed;
gint err = RESAMPLER_ERR_SUCCESS;
guint8 *in_tmp = NULL, *out_tmp = NULL;
gboolean need_convert = (resample->funcs->width != resample->width);
in_len = GST_BUFFER_SIZE (inbuf) / resample->channels;
out_len = GST_BUFFER_SIZE (outbuf) / resample->channels;
in_len /= (resample->width / 8);
out_len /= (resample->width / 8);
in_processed = in_len;
out_processed = out_len;
if (need_convert) {
guint in_size_tmp =
in_len * resample->channels * (resample->funcs->width / 8);
guint out_size_tmp =
out_len * resample->channels * (resample->funcs->width / 8);
if (in_size_tmp <= resample->tmp_in_size) {
in_tmp = resample->tmp_in;
} else {
resample->tmp_in = in_tmp = g_realloc (resample->tmp_in, in_size_tmp);
resample->tmp_in_size = in_size_tmp;
}
gst_audio_resample_convert_buffer (resample, GST_BUFFER_DATA (inbuf),
in_tmp, in_len, FALSE);
if (out_size_tmp <= resample->tmp_out_size) {
out_tmp = resample->tmp_out;
} else {
resample->tmp_out = out_tmp = g_realloc (resample->tmp_out, out_size_tmp);
resample->tmp_out_size = out_size_tmp;
}
}
if (need_convert) {
err = resample->funcs->process (resample->state,
in_tmp, &in_processed, out_tmp, &out_processed);
} else {
err = resample->funcs->process (resample->state,
(const guint8 *) GST_BUFFER_DATA (inbuf), &in_processed,
(guint8 *) GST_BUFFER_DATA (outbuf), &out_processed);
}
if (G_UNLIKELY (in_len != in_processed))
GST_WARNING_OBJECT (resample, "Converted %d of %d input samples",
in_processed, in_len);
if (out_len != out_processed) {
if (out_processed == 0) {
GST_DEBUG_OBJECT (resample, "Converted to 0 samples, buffer dropped");
return GST_BASE_TRANSFORM_FLOW_DROPPED;
}
/* If we wrote more than allocated something is really wrong now
* and we should better abort immediately */
g_assert (out_len >= out_processed);
}
if (G_UNLIKELY (err != RESAMPLER_ERR_SUCCESS)) {
GST_ERROR_OBJECT (resample, "Failed to convert data: %s",
resample->funcs->strerror (err));
return GST_FLOW_ERROR;
} else {
if (need_convert)
gst_audio_resample_convert_buffer (resample, out_tmp,
GST_BUFFER_DATA (outbuf), out_processed, TRUE);
GST_BUFFER_DURATION (outbuf) =
GST_FRAMES_TO_CLOCK_TIME (out_processed, resample->outrate);
GST_BUFFER_SIZE (outbuf) =
out_processed * resample->channels * (resample->width / 8);
if (GST_CLOCK_TIME_IS_VALID (resample->next_ts)) {
GST_BUFFER_TIMESTAMP (outbuf) = resample->next_ts;
GST_BUFFER_OFFSET (outbuf) = resample->next_offset;
GST_BUFFER_OFFSET_END (outbuf) = resample->next_offset + out_processed;
resample->next_ts += GST_BUFFER_DURATION (outbuf);
resample->next_offset += out_processed;
}
GST_LOG_OBJECT (resample,
"Converted to buffer of %u bytes with timestamp %" GST_TIME_FORMAT
", duration %" GST_TIME_FORMAT ", offset %" G_GUINT64_FORMAT
", offset_end %" G_GUINT64_FORMAT, GST_BUFFER_SIZE (outbuf),
GST_TIME_ARGS (GST_BUFFER_TIMESTAMP (outbuf)),
GST_TIME_ARGS (GST_BUFFER_DURATION (outbuf)),
GST_BUFFER_OFFSET (outbuf), GST_BUFFER_OFFSET_END (outbuf));
return GST_FLOW_OK;
}
}
static GstFlowReturn
gst_audio_resample_transform (GstBaseTransform * base, GstBuffer * inbuf,
GstBuffer * outbuf)
{
GstAudioResample *resample = GST_AUDIO_RESAMPLE (base);
guint8 *data;
gulong size;
GstClockTime timestamp;
guint outsamples, insamples;
GstFlowReturn ret;
if (resample->state == NULL) {
if (G_UNLIKELY (!(resample->state =
gst_audio_resample_init_state (resample, resample->width,
resample->channels, resample->inrate, resample->outrate,
resample->quality, resample->fp))))
return GST_FLOW_ERROR;
resample->funcs =
gst_audio_resample_get_funcs (resample->width, resample->fp);
}
data = GST_BUFFER_DATA (inbuf);
size = GST_BUFFER_SIZE (inbuf);
timestamp = GST_BUFFER_TIMESTAMP (inbuf);
GST_LOG_OBJECT (resample, "transforming buffer of %ld bytes, ts %"
GST_TIME_FORMAT ", duration %" GST_TIME_FORMAT ", offset %"
G_GINT64_FORMAT ", offset_end %" G_GINT64_FORMAT,
size, GST_TIME_ARGS (timestamp),
GST_TIME_ARGS (GST_BUFFER_DURATION (inbuf)),
GST_BUFFER_OFFSET (inbuf), GST_BUFFER_OFFSET_END (inbuf));
/* check for timestamp discontinuities and flush/reset if needed */
if (G_UNLIKELY (gst_audio_resample_check_discont (resample, timestamp)
|| GST_BUFFER_IS_DISCONT (inbuf))) {
/* Flush internal samples */
gst_audio_resample_reset_state (resample);
/* Inform downstream element about discontinuity */
resample->need_discont = TRUE;
/* We want to recalculate the timestamps */
resample->next_ts = -1;
resample->next_upstream_ts = -1;
resample->next_offset = -1;
}
insamples = GST_BUFFER_SIZE (inbuf) / resample->channels;
insamples /= (resample->width / 8);
outsamples = GST_BUFFER_SIZE (outbuf) / resample->channels;
outsamples /= (resample->width / 8);
if (GST_CLOCK_TIME_IS_VALID (timestamp)
&& !GST_CLOCK_TIME_IS_VALID (resample->next_ts)) {
resample->next_ts = timestamp;
resample->next_offset =
GST_CLOCK_TIME_TO_FRAMES (timestamp, resample->outrate);
}
if (G_UNLIKELY (resample->need_discont)) {
GST_DEBUG_OBJECT (resample, "marking this buffer with the DISCONT flag");
GST_BUFFER_FLAG_SET (outbuf, GST_BUFFER_FLAG_DISCONT);
resample->need_discont = FALSE;
}
ret = gst_audio_resample_process (resample, inbuf, outbuf);
if (G_UNLIKELY (ret != GST_FLOW_OK))
return ret;
if (GST_CLOCK_TIME_IS_VALID (timestamp)
&& !GST_CLOCK_TIME_IS_VALID (resample->next_upstream_ts))
resample->next_upstream_ts = timestamp;
if (GST_CLOCK_TIME_IS_VALID (resample->next_upstream_ts))
resample->next_upstream_ts +=
GST_FRAMES_TO_CLOCK_TIME (insamples, resample->inrate);
return GST_FLOW_OK;
}
static gboolean
gst_audio_resample_query (GstPad * pad, GstQuery * query)
{
GstAudioResample *resample = GST_AUDIO_RESAMPLE (gst_pad_get_parent (pad));
GstBaseTransform *trans = GST_BASE_TRANSFORM (resample);
gboolean res = TRUE;
switch (GST_QUERY_TYPE (query)) {
case GST_QUERY_LATENCY:
{
GstClockTime min, max;
gboolean live;
guint64 latency;
GstPad *peer;
gint rate = resample->inrate;
gint resampler_latency;
if (resample->state)
resampler_latency =
resample->funcs->get_input_latency (resample->state);
else
resampler_latency = 0;
if (gst_base_transform_is_passthrough (trans))
resampler_latency = 0;
if ((peer = gst_pad_get_peer (trans->sinkpad))) {
if ((res = gst_pad_query (peer, query))) {
gst_query_parse_latency (query, &live, &min, &max);
GST_DEBUG_OBJECT (resample, "Peer latency: min %"
GST_TIME_FORMAT " max %" GST_TIME_FORMAT,
GST_TIME_ARGS (min), GST_TIME_ARGS (max));
/* add our own latency */
if (rate != 0 && resampler_latency != 0)
latency =
gst_util_uint64_scale (resampler_latency, GST_SECOND, rate);
else
latency = 0;
GST_DEBUG_OBJECT (resample, "Our latency: %" GST_TIME_FORMAT,
GST_TIME_ARGS (latency));
min += latency;
if (max != GST_CLOCK_TIME_NONE)
max += latency;
GST_DEBUG_OBJECT (resample, "Calculated total latency : min %"
GST_TIME_FORMAT " max %" GST_TIME_FORMAT,
GST_TIME_ARGS (min), GST_TIME_ARGS (max));
gst_query_set_latency (query, live, min, max);
}
gst_object_unref (peer);
}
break;
}
default:
res = gst_pad_query_default (pad, query);
break;
}
gst_object_unref (resample);
return res;
}
static const GstQueryType *
gst_audio_resample_query_type (GstPad * pad)
{
static const GstQueryType types[] = {
GST_QUERY_LATENCY,
0
};
return types;
}
static void
gst_audio_resample_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec)
{
GstAudioResample *resample;
resample = GST_AUDIO_RESAMPLE (object);
switch (prop_id) {
case PROP_QUALITY:
GST_BASE_TRANSFORM_LOCK (resample);
resample->quality = g_value_get_int (value);
GST_DEBUG_OBJECT (resample, "new quality %d", resample->quality);
gst_audio_resample_update_state (resample, resample->width,
resample->channels, resample->inrate, resample->outrate,
resample->quality, resample->fp);
GST_BASE_TRANSFORM_UNLOCK (resample);
break;
case PROP_FILTER_LENGTH:{
gint filter_length = g_value_get_int (value);
GST_BASE_TRANSFORM_LOCK (resample);
if (filter_length <= 8)
resample->quality = 0;
else if (filter_length <= 16)
resample->quality = 1;
else if (filter_length <= 32)
resample->quality = 2;
else if (filter_length <= 48)
resample->quality = 3;
else if (filter_length <= 64)
resample->quality = 4;
else if (filter_length <= 80)
resample->quality = 5;
else if (filter_length <= 96)
resample->quality = 6;
else if (filter_length <= 128)
resample->quality = 7;
else if (filter_length <= 160)
resample->quality = 8;
else if (filter_length <= 192)
resample->quality = 9;
else
resample->quality = 10;
GST_DEBUG_OBJECT (resample, "new quality %d", resample->quality);
gst_audio_resample_update_state (resample, resample->width,
resample->channels, resample->inrate, resample->outrate,
resample->quality, resample->fp);
GST_BASE_TRANSFORM_UNLOCK (resample);
break;
}
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
gst_audio_resample_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec)
{
GstAudioResample *resample;
resample = GST_AUDIO_RESAMPLE (object);
switch (prop_id) {
case PROP_QUALITY:
g_value_set_int (value, resample->quality);
break;
case PROP_FILTER_LENGTH:
switch (resample->quality) {
case 0:
g_value_set_int (value, 8);
break;
case 1:
g_value_set_int (value, 16);
break;
case 2:
g_value_set_int (value, 32);
break;
case 3:
g_value_set_int (value, 48);
break;
case 4:
g_value_set_int (value, 64);
break;
case 5:
g_value_set_int (value, 80);
break;
case 6:
g_value_set_int (value, 96);
break;
case 7:
g_value_set_int (value, 128);
break;
case 8:
g_value_set_int (value, 160);
break;
case 9:
g_value_set_int (value, 192);
break;
case 10:
g_value_set_int (value, 256);
break;
}
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
#if defined AUDIORESAMPLE_FORMAT_AUTO
#define BENCHMARK_SIZE 512
static gboolean
_benchmark_int_float (SpeexResamplerState * st)
{
gint16 in[BENCHMARK_SIZE] = { 0, }, out[BENCHMARK_SIZE / 2];
gfloat in_tmp[BENCHMARK_SIZE], out_tmp[BENCHMARK_SIZE / 2];
gint i;
guint32 inlen = BENCHMARK_SIZE, outlen = BENCHMARK_SIZE / 2;
for (i = 0; i < BENCHMARK_SIZE; i++) {
gfloat tmp = in[i];
in_tmp[i] = tmp / G_MAXINT16;
}
resample_float_resampler_process_interleaved_float (st,
(const guint8 *) in_tmp, &inlen, (guint8 *) out_tmp, &outlen);
if (outlen == 0) {
GST_ERROR ("Failed to use float resampler");
return FALSE;
}
for (i = 0; i < outlen; i++) {
gfloat tmp = out_tmp[i];
out[i] = CLAMP (tmp * G_MAXINT16 + 0.5, G_MININT16, G_MAXINT16);
}
return TRUE;
}
static gboolean
_benchmark_int_int (SpeexResamplerState * st)
{
gint16 in[BENCHMARK_SIZE] = { 0, }, out[BENCHMARK_SIZE / 2];
guint32 inlen = BENCHMARK_SIZE, outlen = BENCHMARK_SIZE / 2;
resample_int_resampler_process_interleaved_int (st, (const guint8 *) in,
&inlen, (guint8 *) out, &outlen);
if (outlen == 0) {
GST_ERROR ("Failed to use int resampler");
return FALSE;
}
return TRUE;
}
static gboolean
_benchmark_integer_resampling (void)
{
OilProfile a, b;
gdouble av, bv;
SpeexResamplerState *sta, *stb;
oil_profile_init (&a);
oil_profile_init (&b);
sta = resample_float_resampler_init (1, 48000, 24000, 4, NULL);
if (sta == NULL) {
GST_ERROR ("Failed to create float resampler state");
return FALSE;
}
stb = resample_int_resampler_init (1, 48000, 24000, 4, NULL);
if (stb == NULL) {
resample_float_resampler_destroy (sta);
GST_ERROR ("Failed to create int resampler state");
return FALSE;
}
/* Warm up cache */
if (!_benchmark_int_float (sta))
goto error;
if (!_benchmark_int_float (sta))
goto error;
/* Benchmark */
oil_profile_start (&a);
if (!_benchmark_int_float (sta))
goto error;
oil_profile_stop (&a);
/* Warm up cache */
if (!_benchmark_int_int (stb))
goto error;
if (!_benchmark_int_int (stb))
goto error;
/* Benchmark */
oil_profile_start (&b);
if (!_benchmark_int_int (stb))
goto error;
oil_profile_stop (&b);
/* Handle results */
oil_profile_get_ave_std (&a, &av, NULL);
oil_profile_get_ave_std (&b, &bv, NULL);
/* Remember benchmark result in global variable */
gst_audio_resample_use_int = (av > bv);
resample_float_resampler_destroy (sta);
resample_int_resampler_destroy (stb);
if (av > bv)
GST_INFO ("Using integer resampler if appropiate: %lf < %lf", bv, av);
else
GST_INFO ("Using float resampler for everything: %lf <= %lf", av, bv);
return TRUE;
error:
resample_float_resampler_destroy (sta);
resample_int_resampler_destroy (stb);
return FALSE;
}
#endif
static gboolean
plugin_init (GstPlugin * plugin)
{
GST_DEBUG_CATEGORY_INIT (audio_resample_debug, "audioresample", 0,
"audio resampling element");
#if defined AUDIORESAMPLE_FORMAT_AUTO
oil_init ();
if (!_benchmark_integer_resampling ())
return FALSE;
#endif
if (!gst_element_register (plugin, "audioresample", GST_RANK_PRIMARY,
GST_TYPE_AUDIO_RESAMPLE)) {
return FALSE;
}
return TRUE;
}
GST_PLUGIN_DEFINE (GST_VERSION_MAJOR,
GST_VERSION_MINOR,
"audioresample",
"Resamples audio", plugin_init, VERSION, "LGPL", GST_PACKAGE_NAME,
GST_PACKAGE_ORIGIN);
Reference in a new issue View git blame Copy permalink