ltfat Log

Commit Date  
[a5c381] (21.7 kB) by prusa prusa

Experimental C implementation of comp_filterbank routines.

2013-08-21 17:27:28 View
[36b77a] (19.6 kB) by prusa prusa

Major Wavelets rewrite.

2013-07-15 16:39:55 View
[ca1655] (14.9 kB) by prusa prusa

Fixed: Mex template system did not worked properly when one of the MEX inputs was either complex/real/single/double and the other one just real/single/double.

2013-06-13 13:08:13 View
[e0859d] (14.6 kB) by prusa prusa

Normalize lineendings

2013-05-06 10:25:16 View
[6dfb83] (14.6 kB) by Peter L. Søndergaard Peter L. Søndergaard

large patch: converted all files to Unix lineendings

2013-05-02 15:40:52 View
[8ad57d] (15.0 kB) by prusa prusa

Finished OCT interfaces for both single and double precision.
With compiled OCT files
test_all_ltfat for double passes,
test_all_ltfat for single seg. error in TEST_FRAMES after FRAMES ACCEL DUAL REC.

2013-04-30 17:31:51 View
[3c7726] (15.0 kB) by prusa prusa

Windows: LTFAT backend now links BLAS and LAPACK shipped with Matlab.
All mex interfaces which works with input arrays, now accepts single precision inputs and produce single pres. output.

2013-04-26 10:35:44 View
[46e035] (14.5 kB) by prusa prusa

Added assert_classname to all important zeros command, making the input and output data type correspond.
All tests now call ltfatdiditfail.
Some tests do not pass for the single input arguments (increase tolerance?).

mex/comp_col2diag.c and oct/ contain examples of how to do the type-generic calls to the ltfat backend.

Started transforming the backend code to the new naming convention. (MEX and OCT interfaces should not compile).

2013-04-24 12:32:35 View
[d936bb] (14.2 kB) by prusa prusa

Random commit

2013-04-18 17:46:47 View
[8e5a66] (6.3 kB) by prusa prusa

Merged changes in makefiles.

2013-04-04 11:46:34 View
[13d48d] (6.0 kB) by prusa prusa

Small changes.

2013-04-03 14:42:26 View
[cbb813] (4.8 kB) by prusa prusa

Wavelets C implementation is now in *filterbank_td.c files.
Added experimental C "template" functionality -> ltfat_mex_template_helper.h. It is now used in all *filterbank_td.c files to allow generic code for both float and double precision data types.
Added experimental MEX function input/output hook functions to allow handling input/output data reformatting -> ltfat_mex_arg_helper.h
Fixed Mingw64 makefiles.
Added experimental build file for 64bit windows systems using MinGW64 -> ltfatmex_mingw64.m

2013-03-26 13:44:07 View