#include "amrwbdata.h"
typedef struct {
+ AVFrame avframe; ///< AVFrame for decoded samples
AMRWBFrame frame; ///< AMRWB parameters decoded from bitstream
enum Mode fr_cur_mode; ///< mode index of current frame
uint8_t fr_quality; ///< frame quality index (FQI)
for (i = 0; i < 4; i++)
ctx->prediction_error[i] = MIN_ENERGY;
+ avcodec_get_frame_defaults(&ctx->avframe);
+ avctx->coded_frame = &ctx->avframe;
+
return 0;
}
/**
* Decode the frame header in the "MIME/storage" format. This format
- * is simpler and does not carry the auxiliary information of the frame
+ * is simpler and does not carry the auxiliary frame information.
*
* @param[in] ctx The Context
* @param[in] buf Pointer to the input buffer
}
/**
- * Decodes quantized ISF vectors using 36-bit indexes (6K60 mode only)
+ * Decode quantized ISF vectors using 36-bit indexes (6K60 mode only).
*
* @param[in] ind Array of 5 indexes
* @param[out] isf_q Buffer for isf_q[LP_ORDER]
}
/**
- * Decodes quantized ISF vectors using 46-bit indexes (except 6K60 mode)
+ * Decode quantized ISF vectors using 46-bit indexes (except 6K60 mode).
*
* @param[in] ind Array of 7 indexes
* @param[out] isf_q Buffer for isf_q[LP_ORDER]
}
/**
- * Apply mean and past ISF values using the prediction factor
- * Updates past ISF vector
+ * Apply mean and past ISF values using the prediction factor.
+ * Updates past ISF vector.
*
* @param[in,out] isf_q Current quantized ISF
* @param[in,out] isf_past Past quantized ISF
/**
* Interpolate the fourth ISP vector from current and past frames
- * to obtain a ISP vector for each subframe
+ * to obtain an ISP vector for each subframe.
*
* @param[in,out] isp_q ISPs for each subframe
* @param[in] isp4_past Past ISP for subframe 4
}
/**
- * Decode an adaptive codebook index into pitch lag (except 6k60, 8k85 modes)
- * Calculate integer lag and fractional lag always using 1/4 resolution
- * In 1st and 3rd subframes the index is relative to last subframe integer lag
+ * Decode an adaptive codebook index into pitch lag (except 6k60, 8k85 modes).
+ * Calculate integer lag and fractional lag always using 1/4 resolution.
+ * In 1st and 3rd subframes the index is relative to last subframe integer lag.
*
* @param[out] lag_int Decoded integer pitch lag
* @param[out] lag_frac Decoded fractional pitch lag
}
/**
- * Decode a adaptive codebook index into pitch lag for 8k85 and 6k60 modes
- * Description is analogous to decode_pitch_lag_high, but in 6k60 relative
- * index is used for all subframes except the first
+ * Decode an adaptive codebook index into pitch lag for 8k85 and 6k60 modes.
+ * The description is analogous to decode_pitch_lag_high, but in 6k60 the
+ * relative index is used for all subframes except the first.
*/
static void decode_pitch_lag_low(int *lag_int, int *lag_frac, int pitch_index,
uint8_t *base_lag_int, int subframe, enum Mode mode)
/**
* Find the pitch vector by interpolating the past excitation at the
- * pitch delay, which is obtained in this function
+ * pitch delay, which is obtained in this function.
*
* @param[in,out] ctx The context
* @param[in] amr_subframe Current subframe data
/**
* The next six functions decode_[i]p_track decode exactly i pulses
* positions and amplitudes (-1 or 1) in a subframe track using
- * an encoded pulse indexing (TS 26.190 section 5.8.2)
+ * an encoded pulse indexing (TS 26.190 section 5.8.2).
*
* The results are given in out[], in which a negative number means
- * amplitude -1 and vice versa (i.e., ampl(x) = x / abs(x) )
+ * amplitude -1 and vice versa (i.e., ampl(x) = x / abs(x) ).
*
* @param[out] out Output buffer (writes i elements)
* @param[in] code Pulse index (no. of bits varies, see below)
/**
* Decode the algebraic codebook index to pulse positions and signs,
- * then construct the algebraic codebook vector
+ * then construct the algebraic codebook vector.
*
* @param[out] fixed_vector Buffer for the fixed codebook excitation
* @param[in] pulse_hi MSBs part of the pulse index array (higher modes only)
}
/**
- * Decode pitch gain and fixed gain correction factor
+ * Decode pitch gain and fixed gain correction factor.
*
* @param[in] vq_gain Vector-quantized index for gains
* @param[in] mode Mode of the current frame
}
/**
- * Apply pitch sharpening filters to the fixed codebook vector
+ * Apply pitch sharpening filters to the fixed codebook vector.
*
* @param[in] ctx The context
* @param[in,out] fixed_vector Fixed codebook excitation
}
/**
- * Calculate the voicing factor (-1.0 = unvoiced to 1.0 = voiced)
+ * Calculate the voicing factor (-1.0 = unvoiced to 1.0 = voiced).
*
* @param[in] p_vector, f_vector Pitch and fixed excitation vectors
* @param[in] p_gain, f_gain Pitch and fixed gains
}
/**
- * Reduce fixed vector sparseness by smoothing with one of three IR filters
- * Also known as "adaptive phase dispersion"
+ * Reduce fixed vector sparseness by smoothing with one of three IR filters,
+ * also known as "adaptive phase dispersion".
*
* @param[in] ctx The context
* @param[in,out] fixed_vector Unfiltered fixed vector
/**
* Calculate a stability factor {teta} based on distance between
- * current and past isf. A value of 1 shows maximum signal stability
+ * current and past isf. A value of 1 shows maximum signal stability.
*/
static float stability_factor(const float *isf, const float *isf_past)
{
/**
* Apply a non-linear fixed gain smoothing in order to reduce
- * fluctuation in the energy of excitation
+ * fluctuation in the energy of excitation.
*
* @param[in] fixed_gain Unsmoothed fixed gain
* @param[in,out] prev_tr_gain Previous threshold gain (updated)
}
/**
- * Filter the fixed_vector to emphasize the higher frequencies
+ * Filter the fixed_vector to emphasize the higher frequencies.
*
* @param[in,out] fixed_vector Fixed codebook vector
* @param[in] voice_fac Frame voicing factor
}
/**
- * Conduct 16th order linear predictive coding synthesis from excitation
+ * Conduct 16th order linear predictive coding synthesis from excitation.
*
* @param[in] ctx Pointer to the AMRWBContext
* @param[in] lpc Pointer to the LPC coefficients
/**
* Upsample a signal by 5/4 ratio (from 12.8kHz to 16kHz) using
- * a FIR interpolation filter. Uses past data from before *in address
+ * a FIR interpolation filter. Uses past data from before *in address.
*
* @param[out] out Buffer for interpolated signal
* @param[in] in Current signal data (length 0.8*o_size)
/**
* Calculate the high-band gain based on encoded index (23k85 mode) or
- * on the low-band speech signal and the Voice Activity Detection flag
+ * on the low-band speech signal and the Voice Activity Detection flag.
*
* @param[in] ctx The context
* @param[in] synth LB speech synthesis at 12.8k
/**
* Generate the high-band excitation with the same energy from the lower
- * one and scaled by the given gain
+ * one and scaled by the given gain.
*
* @param[in] ctx The context
* @param[out] hb_exc Buffer for the excitation
}
/**
- * Calculate the auto-correlation for the ISF difference vector
+ * Calculate the auto-correlation for the ISF difference vector.
*/
static float auto_correlation(float *diff_isf, float mean, int lag)
{
/**
* Extrapolate a ISF vector to the 16kHz range (20th order LP)
- * used at mode 6k60 LP filter for the high frequency band
+ * used at mode 6k60 LP filter for the high frequency band.
*
* @param[out] out Buffer for extrapolated isf
* @param[in] isf Input isf vector
/**
* Conduct 20th order linear predictive coding synthesis for the high
- * frequency band excitation at 16kHz
+ * frequency band excitation at 16kHz.
*
* @param[in] ctx The context
* @param[in] subframe Current subframe index (0 to 3)
}
/**
- * Apply to high-band samples a 15th order filter
- * The filter characteristic depends on the given coefficients
+ * Apply a 15th order filter to high-band samples.
+ * The filter characteristic depends on the given coefficients.
*
* @param[out] out Buffer for filtered output
* @param[in] fir_coef Filter coefficients
}
/**
- * Update context state before the next subframe
+ * Update context state before the next subframe.
*/
static void update_sub_state(AMRWBContext *ctx)
{
LP_ORDER_16k * sizeof(float));
}
-static int amrwb_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
- AVPacket *avpkt)
+static int amrwb_decode_frame(AVCodecContext *avctx, void *data,
+ int *got_frame_ptr, AVPacket *avpkt)
{
AMRWBContext *ctx = avctx->priv_data;
AMRWBFrame *cf = &ctx->frame;
const uint8_t *buf = avpkt->data;
int buf_size = avpkt->size;
int expected_fr_size, header_size;
- float *buf_out = data;
+ float *buf_out;
float spare_vector[AMRWB_SFR_SIZE]; // extra stack space to hold result from anti-sparseness processing
float fixed_gain_factor; // fixed gain correction factor (gamma)
float *synth_fixed_vector; // pointer to the fixed vector that synthesis should use
float hb_exc[AMRWB_SFR_SIZE_16k]; // excitation for the high frequency band
float hb_samples[AMRWB_SFR_SIZE_16k]; // filtered high-band samples from synthesis
float hb_gain;
- int sub, i;
+ int sub, i, ret;
+
+ /* get output buffer */
+ ctx->avframe.nb_samples = 4 * AMRWB_SFR_SIZE_16k;
+ if ((ret = avctx->get_buffer(avctx, &ctx->avframe)) < 0) {
+ av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
+ return ret;
+ }
+ buf_out = (float *)ctx->avframe.data[0];
header_size = decode_mime_header(ctx, buf);
expected_fr_size = ((cf_sizes_wb[ctx->fr_cur_mode] + 7) >> 3) + 1;
if (buf_size < expected_fr_size) {
av_log(avctx, AV_LOG_ERROR,
"Frame too small (%d bytes). Truncated file?\n", buf_size);
- *data_size = 0;
+ *got_frame_ptr = 0;
return buf_size;
}
memcpy(ctx->isp_sub4_past, ctx->isp[3], LP_ORDER * sizeof(ctx->isp[3][0]));
memcpy(ctx->isf_past_final, ctx->isf_cur, LP_ORDER * sizeof(float));
- /* report how many samples we got */
- *data_size = 4 * AMRWB_SFR_SIZE_16k * sizeof(float);
+ *got_frame_ptr = 1;
+ *(AVFrame *)data = ctx->avframe;
return expected_fr_size;
}
.priv_data_size = sizeof(AMRWBContext),
.init = amrwb_decode_init,
.decode = amrwb_decode_frame,
+ .capabilities = CODEC_CAP_DR1,
.long_name = NULL_IF_CONFIG_SMALL("Adaptive Multi-Rate WideBand"),
.sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_FLT,AV_SAMPLE_FMT_NONE},
};
projects / ffmpeg / blobdiff