log src/vamp-hostsdk/PluginInputDomainAdapter.cpp @ 444:7bab0c5422f4 vampipe

age author description
Thu, 18 Aug 2016 14:43:52 +0100 Chris Cannam Make single/double-precision selectable for input domain adapter windowing and FFTs. Double precision is necessary to pass Sonic Annotator regression tests, though in practice most real-world methods would be fine with single precision. vampipe
Tue, 16 Aug 2016 16:37:57 +0100 Chris Cannam Fix m_ri initialisation vampipe
Tue, 16 Aug 2016 16:04:09 +0100 Chris Cannam Switch from Cross FFT with option of FFTW build, to KissFFT only (code bundled). This is much faster than the default build and simpler than managing two options. vampipe
Thu, 14 Apr 2016 14:23:48 +0100 Chris Cannam Make this warning a #pragma message, as we don't want it to make compilation fail with -Werror set
Thu, 14 Apr 2016 11:49:12 +0100 Chris Cannam Fix compiler warnings, and fix potential mangling of utf8 when downcasing (this is still a nasty and incomplete way to do it though)
Wed, 14 Jan 2015 16:58:53 +0000 Chris Cannam Fine adjustment to timestamp to avoid rounding error
Thu, 16 Oct 2014 10:50:12 +0100 Chris Cannam Non-power-of-two is not an issue if we're built with fftw
Wed, 08 May 2013 10:36:02 +0100 Chris Cannam Emit warning when compiling with FFTW support (about the licence implications) vamp-plugin-sdk-v2.5