search for: opus_int16

...POSSIBILITY OF SUCH DAMAGE. current implementation silences by casting to unsigned. Enabling this should be safe in pretty much all cases, even though it is not technically C89-compliant. */ -#define USE_CELT_FIR 0 -void silk_LPC_analysis_filter( +void silk_LPC_analysis_filter_c( opus_int16 *out, /* O Output signal */ const opus_int16 *in, /* I Input signal */ const opus_int16 *B, /* I MA pre...

--- silk/macros.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/silk/macros.h b/silk/macros.h index 1ba614a..e1e05b9 100644 --- a/silk/macros.h +++ b/silk/macros.h @@ -48,14 +48,14 @@ POSSIBILITY OF SUCH DAMAGE. /* (a32 * (opus_int32)((opus_int16)(b32))) >> 16 output have to be 32bit int */ #if OPUS_FAST_INT64 -#define silk_SMULWB(a32, b32) (((a32) * (opus_int64)((opus_int16)(b32))) >> 16) +#define silk_SMULWB(a32, b32) ((opus_int32)(((a32) * (opus_int64)((opus_int16)(b32))) >> 16)) #else #define s...

...p; __STDC__ && __STDC_VERSION__ >= 199901L) || (defined(__GNUC__) && (defined(_STDINT_H) || defined(_STDINT_H_)) || defined (HAVE_STDINT_H)) -#include <stdint.h> +# include <stdint.h> + typedef int8_t opus_int8; + typedef uint8_t opus_uint8; typedef int16_t opus_int16; typedef uint16_t opus_uint16; typedef int32_t opus_int32; typedef uint32_t opus_uint32; + typedef int64_t opus_int64; + typedef uint64_t opus_uint64; + #elif defined(_WIN32) # if defined(__CYGWIN__) -# include <_G_config.h> - typedef _G_int32_t opus_int32; -...

Hello, I've been trying to understand how to use opus codec for quite some time. There is a thing that doesn't make sense to me. I will try to explain what it is below, In *opus_encode() *function, *opus_int16 pcm* is said to have a size of *frame_size * channels*, where *frame_size* is said to be one of* 20, 240, 480, 960, 1920, and 2880.* Let's say, frame_size is 960, then opus_int16 pcm's size in bytes would be *3840*. Documentation page says the recommended size for *output payload *should be...

...SUCH DAMAGE. #define opus_unlikely(x) (!!(x)) #endif -/* Set this if opus_int64 is a native type of the CPU. */ -#define OPUS_FAST_INT64 (defined(__x86_64__) || defined(__LP64__) || defined(_WIN64)) - /* This is an OPUS_INLINE header file for general platform. */ /* (a32 * (opus_int32)((opus_int16)(b32))) >> 16 output have to be 32bit int */ -- 2.4.9 (Apple Git-60)

...SUCH DAMAGE. #define opus_unlikely(x) (!!(x)) #endif +/* Set this if opus_int64 is a native type of the CPU. */ +#define OPUS_FAST_INT64 (defined(__x86_64__) || defined(__LP64__) || defined(_WIN64)) + /* This is an OPUS_INLINE header file for general platform. */ /* (a32 * (opus_int32)((opus_int16)(b32))) >> 16 output have to be 32bit int */ -#if defined(__x86_64__) || defined(__LP64__) || defined(_WIN64) +#if OPUS_FAST_INT64 #define silk_SMULWB(a32, b32) (((a32) * (opus_int64)((opus_int16)(b32))) >> 16) #else #define silk_SMULWB(a32, b32) ((((a32) >&g...

Hello all, after FOMS I decided to take a look at the opus library and I found that I could improve a bit the build system and cleanup the code a little bit. Most of the changes to the code has been suggested by my two tools cowstats and missingstatic (part of the ruby-elf gem if you care). HTH, Diego

.../* I Encoder State */ silk_nsq_state *NSQ, /* I/O NSQ state */ SideInfoIndices *psIndices, /* I/O Quantization Indices */ const opus_int16 x16[], /* I Input */ @@ -341,7 +341,7 @@ static OPUS_INLINE void silk_noise_shape_quantizer_del_dec( opus_int predictLPCOrder, /* I Prediction filter order */ opus_int...

NSQ_LPC_BUF_LENGTH is independent of DECISION_DELAY. --- silk/define.h | 4 ---- 1 file changed, 4 deletions(-) diff --git a/silk/define.h b/silk/define.h index 781cfdc..1286048 100644 --- a/silk/define.h +++ b/silk/define.h @@ -173,11 +173,7 @@ extern "C" #define MAX_MATRIX_SIZE MAX_LPC_ORDER /* Max of LPC Order and LTP order */ -#if( MAX_LPC_ORDER >

...r_prod_aligned_arm.h" #include "arm/LPC_analysis_filter_arm.h" #include "arm/LPC_inv_pred_gain_arm.h" #endif @@ -377,7 +378,7 @@ opus_int32 silk_inner_prod_aligned( ); -opus_int32 silk_inner_prod_aligned_scale( +opus_int32 silk_inner_prod_aligned_scale_c( const opus_int16 *const inVec1, /* I input vector 1 */ const opus_int16 *const inVec2, /* I input vector 2 */ const opus_int scale, /* I number of bi...

...4 --- a/silk/macros.h +++ b/silk/macros.h @@ -32,6 +32,10 @@ POSSIBILITY OF SUCH DAMAGE. #include "config.h" #endif +#ifdef ARM5E_ASM +#include "macros_arm5e.h" +#else /* Generic macro */ + /* This is an inline header file for general platform. */ /* (a32 * (opus_int32)((opus_int16)(b32))) >> 16 output have to be 32bit int */ @@ -134,5 +138,7 @@ static inline opus_int32 silk_CLZ32(opus_int32 in32) (*((Matrix_base_adr) + ((row)+(M)*(column)))) #endif +#endif + #endif /* SILK_MACROS_H */ diff --git a/silk/macros_arm5e.h b/silk/macros_arm5e.h new file mode 10064...

...r, If I try to make a simple example of opus_encode followed by and immediate opus_decode I do not get reasonable values compared to the input. I'm wondering if I'm somehow misunderstanding how the codec works. Could somebody tell me if my approach could work? int frame_size = 160; opus_int16 test_data_size = frame_size*channels*sizeof(opus_int16); opus_int32 max_data_bytes = 100; opus_int16 test_data[test_data_size]; opus_int16 *p_test_data = test_data; unsigned char encoded_buffer[max_data_bytes]; unsigned char *p_encoded_buffer = encoded_buffer; opus_int16 de...

...bufferIndex < numberOfBuffers ; ++bufferIndex) { AudioBuffer audioBuffer = bufferList->mBuffers[bufferIndex]; int frameSize = (samplingRate/1000) * sampleDuration; // sampling rate is 8000, sample duration is 20ms size_t count = frameSize*channels / sizeof(opus_int16); __block opus_int16 *pcmInput = calloc(count, sizeof(opus_int16)); memcpy(pcmInput, audioBuffer.mData, count * sizeof(opus_int16)); [_encoder encodeWithPCMInput:pcmInput frameSize:frameSize outputBlock:^(unsigned char *payloadOutput, opus_int32 payloadLength) {...

.../main_FIX.h | 11 +++++++++++ silk/x86/x86_silk_map.c | 2 ++ 2 files changed, 13 insertions(+) diff --git a/silk/fixed/main_FIX.h b/silk/fixed/main_FIX.h index ffeb4f3..375b5eb 100644 --- a/silk/fixed/main_FIX.h +++ b/silk/fixed/main_FIX.h @@ -97,6 +97,17 @@ void silk_prefilter_FIX( const opus_int16 x[] /* I Speech signal */ ); +void silk_warped_LPC_analysis_filter_FIX_c( + opus_int32 state[], /* I/O State [order + 1] *...

...e silk_biquad_alt_c_MulSingleAQ28() for your information. It uses 64-bit multiplications and is about 0.6% - 0.9% slower (for the whole encoder) than the original C code using 32-bit multiplications (for both strides and the same stride 2 unrolling). void silk_biquad_alt_c_MulSingleAQ28( const opus_int16 *in, /* I input signal */ const opus_int32 *B_Q28, /* I MA coefficients [3] */ const opus_int32 *A_Q28, /* I AR coefficients...

.../main_FIX.h | 11 +++++++++++ silk/x86/x86_silk_map.c | 2 ++ 2 files changed, 13 insertions(+) diff --git a/silk/fixed/main_FIX.h b/silk/fixed/main_FIX.h index ffeb4f3..375b5eb 100644 --- a/silk/fixed/main_FIX.h +++ b/silk/fixed/main_FIX.h @@ -97,6 +97,17 @@ void silk_prefilter_FIX( const opus_int16 x[] /* I Speech signal */ ); +void silk_warped_LPC_analysis_filter_FIX_c( + opus_int32 state[], /* I/O State [order + 1] *...

thanks for all your feedback. here's the revised patch: On Wed, Oct 11, 2017 at 2:20 PM Timothy B. Terriberry <tterribe at xiph.org> wrote: > Jean-Marc Valin wrote: > > I think you'll want something like: > > (opus_int16)((unsigned)demixing_matrix[2*i+1] << 8) > > (though you might want to check it too) > > FWIW, we use the construct > int s = buf[2*i + 1] << 8 | buf[2*i]; > s = ((s & 0xFFFF) ^ 0x8000) - 0x8000; > > to manually sign-extend the result in opus_compare (an...

--- configure.ac | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/configure.ac b/configure.ac index f52d2c2..e1a6e9b 100644 --- a/configure.ac +++ b/configure.ac @@ -190,7 +190,7 @@ AC_ARG_ENABLE([rtcd], [enable_rtcd=yes]) AC_ARG_ENABLE([intrinsics], - [AS_HELP_STRING([--disable-intrinsics], [Disable intrinsics optimizations for ARM(float) X86(fixed)])],, +

...mation. It uses > 64-bit multiplications and is about 0.6% - 0.9% slower (for the > whole encoder) than the original C code using 32-bit multiplications > (for both strides and the same stride 2 unrolling). > > void silk_biquad_alt_c_MulSingleAQ28( > const opus_int16 *in, /* I input > signal */ > const opus_int32 *B_Q28, /* I MA > coefficients [3] */ > const opus_int32 *A_Q28...

...gt; think is incorrect. > > Done. > 3) opus_projection_ambisonics_encoder_init() (which should be > OPUS_EXPORT) is missing from opus_projection.h > > Done. > 4) I believe mapping_matrix_get_size() should be returning > align(sizeof(MappingMatrix)) + rows * cols * sizeof(opus_int16) > rather than: > align(sizeof(MappingMatrix) + rows * cols * sizeof(opus_int16)) > to match what mapping_matrix_get_data() does > > Done. > 5) I'm not sure I understand the purpose of mapping_matrix_validate(). > Can the user really cause the matrix not to be valid? If y...