aboutsummaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
...
* Initiall support of stdin reading for windowsDaniil Cherednik2020-01-019-33/+393
| | | | Expected au(snd) format, 44100hz, 16bit, stereo or mono
* Stop encoding if case of pipe input and no writerDaniil Cherednik2019-12-092-26/+16
|
* Implementation of atrac3plus analysis pqfDaniil Cherednik2019-08-238-0/+919
|
* Fix some VS compilation warnings0.0.1Daniil Cherednik2019-05-1712-59/+80
|
* Support for MS Windows platformDaniil Cherednik2019-05-1559-140/+1485
| | | | | | - VS build (2017 tested) - Media Foundation Framework support to read/write pcm data (instead of libsndfile)
* [liboma] Write support for ATRAC3PLUS headerDaniil Cherednik2018-12-132-2/+28
|
* [atrac3] Remove tonal extraction codeDaniil Cherednik2018-12-053-90/+2
| | | | | | | Current implementation doesn't add notisible quality improvements, but often adds artifacts due to tonal bit allocation penalty. Proper implementation must use tonal component only if penalty of tonal allocation less than gain generic bit allocation.
* Remove extra copy/allocationDaniil Cherednik2018-07-282-1/+2
|
* Use lrint() + fesetround(FE_TONEAREST) instead of round() - a bit fasterDaniil Cherednik2018-07-284-4/+8
|
* Scale float values to (-1; 1) for internal processingDaniil Cherednik2018-07-047-15/+17
|
* atrac3: Tonal component coding improvements:Daniil Cherednik2018-07-024-54/+99
| | | | | | | | | - Move tonal coding in to bit allocation loop. - Tonal extraction pessimization (should be rewritten with proper psy). - Some tuning. Results: - Fixed high frequency cut while tonal component encoding enabled. - Reduce spectr holes.
* atrac3denc: isnan -> std::isnan.Valtteri Vuorikoski2018-03-311-1/+1
|
* Added validation of the -e option argument and updated help strings.Valtteri Vuorikoski2018-03-311-5/+12
|
* more strict assert - check only last BFUDaniil Cherednik2017-10-291-1/+2
|
* workaround problem with encoding extremely loud tracksDaniil Cherednik2017-10-291-8/+8
| | | | TODO: it looks like ugly hack, check it again
* Do not apply gain control if release in current frame was detected -Daniil Cherednik2017-10-291-1/+1
| | | | right now we can't do it.
* Find optimal BFU num during bit allocation.Daniil Cherednik2017-10-215-26/+38
|
* Fix bit allocationDaniil Cherednik2017-10-191-11/+6
|
* Do not start encoding if unrecognized option givenDaniil Cherednik2017-10-171-3/+4
|
* Simplest gain control - scale first part of frame if the attack transient ↵Daniil Cherednik2017-10-177-103/+167
| | | | was detected
* Refactoring: use normal (previous frame is first) overlap buffer layoutDaniil Cherednik2017-08-054-15/+73
|
* Allow aiff input file formatDaniil Cherednik2017-07-231-1/+7
|
* Fix strict aliasing warningDaniil Cherednik2017-07-231-6/+8
|
* Refactoring to be able to implement joint stereoDaniil Cherednik2017-07-236-85/+133
|
* Set default LP2 for ATRAC3Daniil Cherednik2017-06-151-1/+4
|
* Fix encode optarg optionJiCeheR2017-06-141-2/+2
| | | | | | | * Fix typo encode should 'e' optarg option * enable to set 'atrac3' argument to optarg Signed-off-by: JiCeheR <jiceher@free.fr>
* Fix compilation eror (missed include)Daniil Cherednik2017-02-121-2/+3
|
* Set LGPL v2.1 licenseDaniil Cherednik2017-02-1253-0/+954
|
* Support of big endian platformsDaniil Cherednik2016-11-272-43/+39
|
* Checking of input file format addedDaniil Cherednik2016-09-293-30/+61
|
* kiss_fft library added, MDCT reimplemented via FFT:Daniil Cherednik2016-09-2710-40/+908
| | | | | | - we need fft for some atrac3 features - we need fft to implement better psychoacoustics analysis - use single transform library
* fix compilation errors in case of building without utDaniil Cherednik2016-09-253-0/+20
|
* Change gaincontrol default: disableDaniil Cherednik2016-09-021-5/+5
| | | | It looks like gaincontrol feature should be reimplemented. So disable it for a while.
* Dirty implementation of atrac3 encoder:Daniil Cherednik2016-09-0247-405/+3024
| | | | | | | - no JS mode - constant quantiser for tonal components - gain controll implemented but produces some artifacts with real signals. - etc...
* Compilation warnings fixedDaniil Cherednik2016-07-177-38/+33
|
* Fix building with cmake < 3.1Daniil Cherednik2016-07-173-3/+2
|
* Fix memory corruption: wrong size of HPFBuffer.Daniil Cherednik2016-04-222-9/+8
|
* fix memory corruptionDaniil Cherednik2016-04-221-1/+1
|
* fix memory corruptionDaniil Cherednik2016-04-221-1/+1
|
* atracdenc ut added to CMake buildDaniil Cherednik2016-04-021-1/+1
|
* do not use compile time frame size specialisationDaniil Cherednik2016-03-084-8/+22
|
* a bit more generic TScalerDaniil Cherednik2016-03-087-30/+27
|
* autodetect saved fmt by file extension (WAV, AU, AIFF, RAW)Daniil Cherednik2016-03-072-10/+37
| | | | | | use AU format in case of "-" (stdout) output (something like "./atracdenc --decode -i /tmp/test.aea -o - | cvlc - " works) why AU? http://www.mega-nerd.com/libsndfile/FAQ.html#Q017
* old Makefiles removedDaniil Cherednik2016-03-072-20/+0
|
* remove suboptimal multiply readf (writef) libsndfile calls during reading or ↵Daniil Cherednik2016-03-073-48/+71
| | | | writing wav file
* refactoring:Daniil Cherednik2016-01-067-32/+63
| | | | | some changes to get possibility to create alternative compression algorithms and use different containers.
* fix: do not swap stereo channels during decodingDaniil Cherednik2016-01-021-1/+1
|
* split fixed bit allocation table for short and long blocksDaniil Cherednik2015-12-313-4/+18
|
* improve scaling accuracy, prevent overflowDaniil Cherednik2015-12-314-6/+16
|
* use libsndfile instead of own wav readerDaniil Cherednik2015-12-275-114/+70
|
generated by cgit v1.2.3 (git 2.25.1) at 2025-07-13 13:37:44 +0000