aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Fix compilation for FreeBSD and probably other OS without alloca.h.at3pghadevDaniil Cherednik2025-05-312-0/+12
|
* Fix compilation for WindowsDaniil Cherednik2025-05-315-11/+33
|
* Fix memory leak in dtmf example.Daniil Cherednik2025-05-221-0/+1
|
* Fix memory corruption bug. Appears in 8829fb918111Daniil Cherednik2025-05-221-1/+1
|
* Upsample signal before omega search. Use Hann window.Daniil Cherednik2025-05-223-8/+104
|
* Get rid of VLA usage, use alloca insteadDaniil Cherednik2025-05-151-64/+95
|
* Slightly faster. Do not touch work buffer each time.Daniil Cherednik2025-05-041-1/+5
|
* Attempt to speed up.Daniil Cherednik2025-04-261-17/+16
| | | | | * use float sin/cos calculation which is faster and probably still have enouth precision. * do not convert float to double inside O(n^2) loop. From other side it causes worse cache locality, so gain depends on data size and dimension.
* allow to set partial frame size during gha_adjust_infoDaniil Cherednik2024-10-204-23/+76
|
* [bugfix] do not call residual cb in case of unsuccessful multidimensional ↵Daniil Cherednik2024-10-061-1/+1
| | | | optimization result
* Remove resuidal cb from global ctx. Just add it into adjust_info functionDaniil Cherednik2024-09-064-38/+17
|
* Allow to set max magnitude, iteration count parameters for gha_adjust_infoDaniil Cherednik2024-09-014-6/+192
|
* allow to get generated pcm signalDaniil Cherednik2024-08-112-0/+7
|
* Make possible to use double and own fft lib.HEADmasterDaniil Cherednik2024-07-066-48/+67
| | | | | Projects which use the library may have own fft librray. Projects may use double precision floating point numbers.
* Update codeql-analysis.ymlDaniil Cherednik2023-10-171-3/+3
|
* bzero -> memset replacementU-zzzz-VAIO\developer2023-09-191-2/+2
|
* Add codeqlDaniil Cherednik2022-02-091-0/+70
|
* Workflow status badge addedDaniil Cherednik2021-11-041-0/+2
|
* Fix utDaniil Cherednik2021-11-032-8/+8
|
* Add github actionDaniil Cherednik2021-11-021-0/+38
|
* Fix warningsDaniil Cherednik2021-02-232-3/+4
|
* Merge pull request #1 from rogovsky/devDaniil Cherednik2019-10-071-1/+10
|\ | | | | Minor changes
| * Guards in header was added for beeing used with C++.Yury A. Rogovsky2019-10-061-0/+9
| | | | | | | | | | | | Guards in header file was added to prevent function's name mangling (i.e. altering the name of the funtion in object -or library- files) while beeing compiled with C++.
| * Missprint correction in header file.Yury A. Rogovsky2019-10-061-1/+1
|/
* Optimize gha_search_omega - do not call sin/cos each iterationDaniil Cherednik2018-12-021-6/+18
|
* Initiall implementation of multi dimentional Newton optimizationDaniil Cherednik2018-11-298-4/+4303
|
* Simple method to extract K sinusoids for one callDaniil Cherednik2018-11-043-4/+32
|
* A bit more clean apiDaniil Cherednik2018-11-034-29/+42
|
* Create README.mdDaniil Cherednik2018-11-031-0/+4
|
* Add function to extract analysed harmonic from given pcmDaniil Cherednik2018-11-037-43/+169
| | | | Added dtmf analysation example
* First commitdcherednik2018-11-0248-0/+5670
- Method to get parameters of one harmonic