Seppo Ingalsuo
Seppo Ingalsuo
[FEATURE] Optimize audio processing components' source/sink buffer access, deprecate read/write frag
Changed IGO optimize priority to high (see issue #4970).
[FEATURE] Optimize audio processing components' source/sink buffer access, deprecate read/write frag
Next ones to optimize are KBP and Selector, unless other requested.
This PR tests idea of executing Octave reference code to create test vector to compare. Unfortunately it's not possible to call host shell correctly from xt-run environment so it's left...
> Looks good to me. And I'm curious about the rms between ref-win and win. I'm not sure if RMS is the best error criteria for this but it's simple....
The just updated version is using the old way pre-calculated data header files. The script ref_window.m was converted to export the headers instead of computing them in unit test runtime.
> > @ShriramShastry good for you ? > > Yes, it seems like the code is correct right now. In the present, as I return a second time, I see...
This push fixed the conflict in test/cmocka/src/math/CMakeLists.txt, no other changes.
The two codestyle checpatch fails are due to `FILE *fh;` The checker does not know the syntax. There's a debug option in unit test to output the numbers test data...
@EdwardAbramian2-SR The code doesn't build. I created a minimal kconfig and added sr_audio to CmakeLists.txt. I will try the block audio stream functions (to replace the read/write frag functions) with...
> @singalsu > > We have no problem building our code under tgl-013-drop-stable branch, assuming our SOF_COMP_SRAUDIO is defined in topology (see the attached build log). We used that branch...