FFmpeg

Commit Graph

Author	SHA1	Message	Date
Justin Ruggles	e0cc66df61	ac3enc: split templated float vs. fixed functions into a separate file. Function pointers are used for templated functions instead of needlessly duplicating many functions.	14 years ago
Justin Ruggles	e754dfc0bb	ac3enc: dynamically allocate AC3EncodeContext fields windowed_samples and mdct This will allow the same struct to be used for both the fixed and float ac3 encoders.	14 years ago
Justin Ruggles	36151b3e31	ac3enc: use function pointer to choose between AC-3 and E-AC-3 header output functions.	14 years ago
Diego Biurrun	d3778972d8	ac3enc: Fix linking of AC-3 encoder without the E-AC-3 encoder. The AC-3 encoder unconditionally references some symbols from the E-AC-3 encoder; make those references conditional to fix linking.	14 years ago
Justin Ruggles	c8e9ea43d0	Move E-AC-3 encoder functions to a separate eac3enc.c file.	14 years ago
Justin Ruggles	787a13535a	ac3enc: remove convenience macro, #define DEBUG	14 years ago
Justin Ruggles	a8bd53402a	ac3enc: remove unused #define	14 years ago
Justin Ruggles	35d5cb1ab4	ac3enc: extract all exponents for the frame at once	14 years ago
Mans Rullgard	90da52f01f	ac3enc: fix LOCAL_ALIGNED usage in count_mantissa_bits() Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Justin Ruggles	6ca23db9cc	ac3enc: modify mantissa bit counting to keep bap counts for all values of bap instead of just 0 to 4. This does all the actual bit counting as a final step.	14 years ago
Justin Ruggles	1323828a0f	ac3enc: split mantissa bit counting into a separate function. No speed difference. This is to allow for more flexible bit counting.	14 years ago
Justin Ruggles	7743865ffc	ac3enc: store per-block/channel bap pointers by reference block in a 2D array rather than in the AC3Block struct. This will make it easier to access the bap values without having to chase the reference block pointers each time.	14 years ago
Mans Rullgard	7f7726c7a2	ac3enc: use LOCAL_ALIGNED macro Aligned local variables must use the LOCAL_ALIGNED macro to ensure correct alignment on all systems. The unusual size argument to memset is required since the implementation of LOCAL_ALIGNED may use a pointer to an array. Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Justin Ruggles	1a950da6a2	cosmetics: indentation and alignment after previous commit	14 years ago
Justin Ruggles	aa47c35dff	ac3enc: add support for E-AC-3 encoding. This adds basic stream format support and allows for arbitrary bit rates rather than just those supported in AC-3.	14 years ago
Justin Ruggles	b1bf8c7887	ac3enc: Move AC-3 AVOptions array to a separate file to make it easier to use only selected options for the different AC-3 encoder types.	14 years ago
Justin Ruggles	478455d66b	ac3enc: initialize all coefficients to zero. Uninitialized coefficients were being used to generate exponents, some of which actually ended up in the final stream. Even though, they were just extra exponents that are not used by any decoder, it is still better to have consistent output for testing. This also fixes valgrind errors.	14 years ago
Justin Ruggles	7f3a7b5c40	ac3enc: add channel coupling support Channel coupling is an optional AC-3 feature that increases quality by combining high frequency information from multiple channels into a single channel. The per-channel high frequency information is sent with less accuracy in both the frequency and time domains. This allows more bits to be used for lower frequencies while preserving enough information to reconstruct the high frequencies.	14 years ago
Justin Ruggles	4d7a421574	ac3enc: put the counting of stereo rematrixing bits in the same place to make the code easier to understand.	14 years ago
Justin Ruggles	257de756fb	ac3enc: clean up count_frame_bits() and count_frame_bits_fixed()	14 years ago
Anton Khirnov	b66752790a	AVOptions: make default_val a union, as proposed in AVOption2. This breaks API and ABI.	14 years ago
Alex Converse	1a5e4fd8c5	Replace strncpy() with av_strlcpy().	14 years ago
Justin Ruggles	177fed4e9b	ac3enc: do not store a bandwidth code for each channel. Although AC-3 allows it, it's not very useful. The encoder uses the same code for all full-bandwidth channels.	14 years ago
Justin Ruggles	a1d0f511fc	ac3enc: remove bandwidth reduction as fallback for bit allocation failure. It was only needed at low bitrates, which now already use a low bandwidth, so the bandwidth reduction is no longer needed.	14 years ago
Justin Ruggles	ba6bce5140	ac3enc: merge compute_exp_strategy_ch() into compute_exp_strategy()	14 years ago
Justin Ruggles	4142487d1c	ac3enc: return error if frame+exponent bits are too large instead of using av_assert2(). This can occur in some very rare cases with low bitrates.	14 years ago
Justin Ruggles	987fe2dc55	ac3enc: differentiate between current block and reference block in bit_alloc()	14 years ago
Justin Ruggles	6b2636bba6	ac3enc: simplify exponent_init() by calculating exponent_group_tab[] based on exponent group sizes.	14 years ago
Justin Ruggles	e0b33d479c	ac3enc: simplify stereo rematrixing decision options	14 years ago
Anton Khirnov	85770f2a26	AVOptions: make default_val a union, as proposed in AVOption2. This breaks API and ABI.	14 years ago
Justin Ruggles	8745e9c458	ac3enc: remove check for mismatching channels and channel_layout	14 years ago
Vitor Sessak	9d35fa520e	Add AVX FFT implementation. Signed-off-by: Reinhard Tartler <siretart@tauware.de>	14 years ago
Michael Niedermayer	1aeb88b77d	Correctly implement ac3 float/fixed encoder. There is no need to have 2 encoders, the input sample format can,does and should choose which is used Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	14 years ago
Mans Rullgard	2310ee4b1c	ac3enc: move extract_exponents inner loop to ac3dsp Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Justin Ruggles	e05a3ac713	ac3enc: select bandwidth based on bit rate, sample rate, and number of full-bandwidth channels. This reduces high-frequency artifacts and improves the quality of the lower frequency audio at low bit rates.	14 years ago
Mans Rullgard	79997def65	ac3enc: use generic fixed-point mdct This makes the AC3 encoder use the shared fixed-point MDCT rather than its own implementation. The checksum changes are due to different rounding in the MDCT. Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Mans Rullgard	7f6e05cdfd	ac3enc: simplify sym_quant() These expressions are equivalent since levels is always odd, and overflow is impossible due to the constraints set by the assert(). Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Mans Rullgard	52fd16a264	ac3enc: move compute_mantissa_size() to ac3dsp Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Mans Rullgard	6f718471ea	ac3enc: move mant_cnt and qmant_ptr out of AC3EncodeContext These fields are only used in quantize_mantissas() and reset on each call, no need to store them in the main context. Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Mans Rullgard	6d9f52b2cd	ac3: move ff_ac3_bit_alloc_calc_bap to ac3dsp Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Justin Ruggles	c4549bd66a	ac3enc: avoid memcpy() of exponents and baps in EXP_REUSE case by using exponent reference blocks.	14 years ago
Justin Ruggles	991f3de1bb	ac3enc: Add codec-specific options for writing AC-3 metadata.	14 years ago
Justin Ruggles	be18738801	Get audio_service_type for AC-3 based on bitstream mode in the AC-3 parser and decoder, and vice-versa for the AC-3 encoder.	14 years ago
Justin Ruggles	e6e9823488	Add apply_window_int16() to DSPContext with x86-optimized versions and use it in the ac3_fixed encoder.	14 years ago
Justin Ruggles	85ab290470	ac3enc: make sym_quant() branch-free	14 years ago
Mans Rullgard	2912e87a6c	Replace FFmpeg with Libav in licence headers Signed-off-by: Mans Rullgard <mans@mansr.com>	14 years ago
Justin Ruggles	0f999cfddb	ac3enc: add float_to_fixed24() with x86-optimized versions to AC3DSPContext and use in scale_coefficients() for the floating-point AC-3 encoder.	14 years ago
Michael Niedermayer	27293b840c	Revert "ac3enc: shift coefficients to 24-bit following MDCT rather than using an exponent offset." This reverts commit `7e0a284b9f`. revert at authors request due to better impementation Signed-off-by: Michael Niedermayer <michaelni@gmx.at>	14 years ago
Justin	323e6fead0	ac3enc: do not right-shift fixed-point coefficients in the final MDCT stage. This increases the accuracy of coefficients, leading to improved quality. Rescaling of the coefficients to full 25-bit accuracy is done rather than offsetting the exponent values. This requires coefficient scaling to be done before determining the rematrixing strategy. Also, the rematrixing strategy calculation must use 64-bit math to prevent overflow due to the higher precision coefficients.	14 years ago
Michael Niedermayer	2800e8d62e	Revert "ac3enc: use MUL64() to multiply fixed-point coefficients" Author says, this is unneeded. This reverts commit `67fc32ac61`.	14 years ago

1 2 3 4 5 ...

382 Commits (77f359670fdbfcd7ac2a464b3bc000b1ceb54763)