annotate src/ext/kissfft/CHANGELOG @ 196:da283326bcd3 tip master

Update plugin versions in RDF
author Chris Cannam <cannam@all-day-breakfast.com>
date Fri, 28 Feb 2020 09:43:02 +0000
parents 5ed6e970541b
children
rev   line source
c@174 1 1.3.0 2012-07-18
c@174 2 removed non-standard malloc.h from kiss_fft.h
c@174 3
c@174 4 moved -lm to end of link line
c@174 5
c@174 6 checked various return values
c@174 7
c@174 8 converted python Numeric code to NumPy
c@174 9
c@174 10 fixed test of int32_t on 64 bit OS
c@174 11
c@174 12 added padding in a couple of places to allow SIMD alignment of structs
c@174 13
c@174 14 1.2.9 2010-05-27
c@174 15 threadsafe ( including OpenMP )
c@174 16
c@174 17 first edition of kissfft.hh the C++ template fft engine
c@174 18
c@174 19 1.2.8
c@174 20 Changed memory.h to string.h -- apparently more standard
c@174 21
c@174 22 Added openmp extensions. This can have fairly linear speedups for larger FFT sizes.
c@174 23
c@174 24 1.2.7
c@174 25 Shrank the real-fft memory footprint. Thanks to Galen Seitz.
c@174 26
c@174 27 1.2.6 (Nov 14, 2006) The "thanks to GenArts" release.
c@174 28 Added multi-dimensional real-optimized FFT, see tools/kiss_fftndr
c@174 29 Thanks go to GenArts, Inc. for sponsoring the development.
c@174 30
c@174 31 1.2.5 (June 27, 2006) The "release for no good reason" release.
c@174 32 Changed some harmless code to make some compilers' warnings go away.
c@174 33 Added some more digits to pi -- why not.
c@174 34 Added kiss_fft_next_fast_size() function to help people decide how much to pad.
c@174 35 Changed multidimensional test from 8 dimensions to only 3 to avoid testing
c@174 36 problems with fixed point (sorry Buckaroo Banzai).
c@174 37
c@174 38 1.2.4 (Oct 27, 2005) The "oops, inverse fixed point real fft was borked" release.
c@174 39 Fixed scaling bug for inverse fixed point real fft -- also fixed test code that should've been failing.
c@174 40 Thanks to Jean-Marc Valin for bug report.
c@174 41
c@174 42 Use sys/types.h for more portable types than short,int,long => int16_t,int32_t,int64_t
c@174 43 If your system does not have these, you may need to define them -- but at least it breaks in a
c@174 44 loud and easily fixable way -- unlike silently using the wrong size type.
c@174 45
c@174 46 Hopefully tools/psdpng.c is fixed -- thanks to Steve Kellog for pointing out the weirdness.
c@174 47
c@174 48 1.2.3 (June 25, 2005) The "you want to use WHAT as a sample" release.
c@174 49 Added ability to use 32 bit fixed point samples -- requires a 64 bit intermediate result, a la 'long long'
c@174 50
c@174 51 Added ability to do 4 FFTs in parallel by using SSE SIMD instructions. This is accomplished by
c@174 52 using the __m128 (vector of 4 floats) as kiss_fft_scalar. Define USE_SIMD to use this.
c@174 53
c@174 54 I know, I know ... this is drifting a bit from the "kiss" principle, but the speed advantages
c@174 55 make it worth it for some. Also recent gcc makes it SOO easy to use vectors of 4 floats like a POD type.
c@174 56
c@174 57 1.2.2 (May 6, 2005) The Matthew release
c@174 58 Replaced fixed point division with multiply&shift. Thanks to Jean-Marc Valin for
c@174 59 discussions regarding. Considerable speedup for fixed-point.
c@174 60
c@174 61 Corrected overflow protection in real fft routines when using fixed point.
c@174 62 Finder's Credit goes to Robert Oschler of robodance for pointing me at the bug.
c@174 63 This also led to the CHECK_OVERFLOW_OP macro.
c@174 64
c@174 65 1.2.1 (April 4, 2004)
c@174 66 compiles cleanly with just about every -W warning flag under the sun
c@174 67
c@174 68 reorganized kiss_fft_state so it could be read-only/const. This may be useful for embedded systems
c@174 69 that are willing to predeclare twiddle factors, factorization.
c@174 70
c@174 71 Fixed C_MUL,S_MUL on 16-bit platforms.
c@174 72
c@174 73 tmpbuf will only be allocated if input & output buffers are same
c@174 74 scratchbuf will only be allocated for ffts that are not multiples of 2,3,5
c@174 75
c@174 76 NOTE: The tmpbuf,scratchbuf changes may require synchronization code for multi-threaded apps.
c@174 77
c@174 78
c@174 79 1.2 (Feb 23, 2004)
c@174 80 interface change -- cfg object is forward declaration of struct instead of void*
c@174 81 This maintains type saftey and lets the compiler warn/error about stupid mistakes.
c@174 82 (prompted by suggestion from Erik de Castro Lopo)
c@174 83
c@174 84 small speed improvements
c@174 85
c@174 86 added psdpng.c -- sample utility that will create png spectrum "waterfalls" from an input file
c@174 87 ( not terribly useful yet)
c@174 88
c@174 89 1.1.1 (Feb 1, 2004 )
c@174 90 minor bug fix -- only affects odd rank, in-place, multi-dimensional FFTs
c@174 91
c@174 92 1.1 : (Jan 30,2004)
c@174 93 split sample_code/ into test/ and tools/
c@174 94
c@174 95 Removed 2-D fft and added N-D fft (arbitrary)
c@174 96
c@174 97 modified fftutil.c to allow multi-d FFTs
c@174 98
c@174 99 Modified core fft routine to allow an input stride via kiss_fft_stride()
c@174 100 (eased support of multi-D ffts)
c@174 101
c@174 102 Added fast convolution filtering (FIR filtering using overlap-scrap method, with tail scrap)
c@174 103
c@174 104 Add kfc.[ch]: the KISS FFT Cache. It takes care of allocs for you ( suggested by Oscar Lesta ).
c@174 105
c@174 106 1.0.1 (Dec 15, 2003)
c@174 107 fixed bug that occurred when nfft==1. Thanks to Steven Johnson.
c@174 108
c@174 109 1.0 : (Dec 14, 2003)
c@174 110 changed kiss_fft function from using a single buffer, to two buffers.
c@174 111 If the same buffer pointer is supplied for both in and out, kiss will
c@174 112 manage the buffer copies.
c@174 113
c@174 114 added kiss_fft2d and kiss_fftr as separate source files (declarations in kiss_fft.h )
c@174 115
c@174 116 0.4 :(Nov 4,2003) optimized for radix 2,3,4,5
c@174 117
c@174 118 0.3 :(Oct 28, 2003) woops, version 2 didn't actually factor out any radices other than 2.
c@174 119 Thanks to Steven Johnson for finding this one.
c@174 120
c@174 121 0.2 :(Oct 27, 2003) added mixed radix, only radix 2,4 optimized versions
c@174 122
c@174 123 0.1 :(May 19 2003) initial release, radix 2 only