diff options
Diffstat (limited to 'media/kiss_fft')
| -rw-r--r-- | media/kiss_fft/CHANGELOG | 123 | ||||
| -rw-r--r-- | media/kiss_fft/COPYING | 11 | ||||
| -rw-r--r-- | media/kiss_fft/README | 134 | ||||
| -rw-r--r-- | media/kiss_fft/README.simd | 78 | ||||
| -rw-r--r-- | media/kiss_fft/README_MOZILLA | 8 | ||||
| -rw-r--r-- | media/kiss_fft/_kiss_fft_guts.h | 164 | ||||
| -rw-r--r-- | media/kiss_fft/kiss_fft.c | 408 | ||||
| -rw-r--r-- | media/kiss_fft/kiss_fft.h | 124 | ||||
| -rw-r--r-- | media/kiss_fft/kiss_fftr.c | 159 | ||||
| -rw-r--r-- | media/kiss_fft/kiss_fftr.h | 46 | ||||
| -rw-r--r-- | media/kiss_fft/moz.build | 17 | ||||
| -rwxr-xr-x | media/kiss_fft/update.sh | 34 |
12 files changed, 1306 insertions, 0 deletions
diff --git a/media/kiss_fft/CHANGELOG b/media/kiss_fft/CHANGELOG new file mode 100644 index 000000000..2dd360375 --- /dev/null +++ b/media/kiss_fft/CHANGELOG @@ -0,0 +1,123 @@ +1.3.0 2012-07-18 + removed non-standard malloc.h from kiss_fft.h + + moved -lm to end of link line + + checked various return values + + converted python Numeric code to NumPy + + fixed test of int32_t on 64 bit OS + + added padding in a couple of places to allow SIMD alignment of structs + +1.2.9 2010-05-27 + threadsafe ( including OpenMP ) + + first edition of kissfft.hh the C++ template fft engine + +1.2.8 + Changed memory.h to string.h -- apparently more standard + + Added openmp extensions. This can have fairly linear speedups for larger FFT sizes. + +1.2.7 + Shrank the real-fft memory footprint. Thanks to Galen Seitz. + +1.2.6 (Nov 14, 2006) The "thanks to GenArts" release. + Added multi-dimensional real-optimized FFT, see tools/kiss_fftndr + Thanks go to GenArts, Inc. for sponsoring the development. + +1.2.5 (June 27, 2006) The "release for no good reason" release. + Changed some harmless code to make some compilers' warnings go away. + Added some more digits to pi -- why not. + Added kiss_fft_next_fast_size() function to help people decide how much to pad. + Changed multidimensional test from 8 dimensions to only 3 to avoid testing + problems with fixed point (sorry Buckaroo Banzai). + +1.2.4 (Oct 27, 2005) The "oops, inverse fixed point real fft was borked" release. + Fixed scaling bug for inverse fixed point real fft -- also fixed test code that should've been failing. + Thanks to Jean-Marc Valin for bug report. + + Use sys/types.h for more portable types than short,int,long => int16_t,int32_t,int64_t + If your system does not have these, you may need to define them -- but at least it breaks in a + loud and easily fixable way -- unlike silently using the wrong size type. + + Hopefully tools/psdpng.c is fixed -- thanks to Steve Kellog for pointing out the weirdness. + +1.2.3 (June 25, 2005) The "you want to use WHAT as a sample" release. + Added ability to use 32 bit fixed point samples -- requires a 64 bit intermediate result, a la 'long long' + + Added ability to do 4 FFTs in parallel by using SSE SIMD instructions. This is accomplished by + using the __m128 (vector of 4 floats) as kiss_fft_scalar. Define USE_SIMD to use this. + + I know, I know ... this is drifting a bit from the "kiss" principle, but the speed advantages + make it worth it for some. Also recent gcc makes it SOO easy to use vectors of 4 floats like a POD type. + +1.2.2 (May 6, 2005) The Matthew release + Replaced fixed point division with multiply&shift. Thanks to Jean-Marc Valin for + discussions regarding. Considerable speedup for fixed-point. + + Corrected overflow protection in real fft routines when using fixed point. + Finder's Credit goes to Robert Oschler of robodance for pointing me at the bug. + This also led to the CHECK_OVERFLOW_OP macro. + +1.2.1 (April 4, 2004) + compiles cleanly with just about every -W warning flag under the sun + + reorganized kiss_fft_state so it could be read-only/const. This may be useful for embedded systems + that are willing to predeclare twiddle factors, factorization. + + Fixed C_MUL,S_MUL on 16-bit platforms. + + tmpbuf will only be allocated if input & output buffers are same + scratchbuf will only be allocated for ffts that are not multiples of 2,3,5 + + NOTE: The tmpbuf,scratchbuf changes may require synchronization code for multi-threaded apps. + + +1.2 (Feb 23, 2004) + interface change -- cfg object is forward declaration of struct instead of void* + This maintains type saftey and lets the compiler warn/error about stupid mistakes. + (prompted by suggestion from Erik de Castro Lopo) + + small speed improvements + + added psdpng.c -- sample utility that will create png spectrum "waterfalls" from an input file + ( not terribly useful yet) + +1.1.1 (Feb 1, 2004 ) + minor bug fix -- only affects odd rank, in-place, multi-dimensional FFTs + +1.1 : (Jan 30,2004) + split sample_code/ into test/ and tools/ + + Removed 2-D fft and added N-D fft (arbitrary) + + modified fftutil.c to allow multi-d FFTs + + Modified core fft routine to allow an input stride via kiss_fft_stride() + (eased support of multi-D ffts) + + Added fast convolution filtering (FIR filtering using overlap-scrap method, with tail scrap) + + Add kfc.[ch]: the KISS FFT Cache. It takes care of allocs for you ( suggested by Oscar Lesta ). + +1.0.1 (Dec 15, 2003) + fixed bug that occurred when nfft==1. Thanks to Steven Johnson. + +1.0 : (Dec 14, 2003) + changed kiss_fft function from using a single buffer, to two buffers. + If the same buffer pointer is supplied for both in and out, kiss will + manage the buffer copies. + + added kiss_fft2d and kiss_fftr as separate source files (declarations in kiss_fft.h ) + +0.4 :(Nov 4,2003) optimized for radix 2,3,4,5 + +0.3 :(Oct 28, 2003) woops, version 2 didn't actually factor out any radices other than 2. + Thanks to Steven Johnson for finding this one. + +0.2 :(Oct 27, 2003) added mixed radix, only radix 2,4 optimized versions + +0.1 :(May 19 2003) initial release, radix 2 only diff --git a/media/kiss_fft/COPYING b/media/kiss_fft/COPYING new file mode 100644 index 000000000..2fc6685a6 --- /dev/null +++ b/media/kiss_fft/COPYING @@ -0,0 +1,11 @@ +Copyright (c) 2003-2010 Mark Borgerding + +All rights reserved. + +Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. + * Neither the author nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. diff --git a/media/kiss_fft/README b/media/kiss_fft/README new file mode 100644 index 000000000..03b2e7a9c --- /dev/null +++ b/media/kiss_fft/README @@ -0,0 +1,134 @@ +KISS FFT - A mixed-radix Fast Fourier Transform based up on the principle, +"Keep It Simple, Stupid." + + There are many great fft libraries already around. Kiss FFT is not trying +to be better than any of them. It only attempts to be a reasonably efficient, +moderately useful FFT that can use fixed or floating data types and can be +incorporated into someone's C program in a few minutes with trivial licensing. + +USAGE: + + The basic usage for 1-d complex FFT is: + + #include "kiss_fft.h" + + kiss_fft_cfg cfg = kiss_fft_alloc( nfft ,is_inverse_fft ,0,0 ); + + while ... + + ... // put kth sample in cx_in[k].r and cx_in[k].i + + kiss_fft( cfg , cx_in , cx_out ); + + ... // transformed. DC is in cx_out[0].r and cx_out[0].i + + free(cfg); + + Note: frequency-domain data is stored from dc up to 2pi. + so cx_out[0] is the dc bin of the FFT + and cx_out[nfft/2] is the Nyquist bin (if exists) + + Declarations are in "kiss_fft.h", along with a brief description of the +functions you'll need to use. + +Code definitions for 1d complex FFTs are in kiss_fft.c. + +You can do other cool stuff with the extras you'll find in tools/ + + * multi-dimensional FFTs + * real-optimized FFTs (returns the positive half-spectrum: (nfft/2+1) complex frequency bins) + * fast convolution FIR filtering (not available for fixed point) + * spectrum image creation + +The core fft and most tools/ code can be compiled to use float, double, + Q15 short or Q31 samples. The default is float. + + +BACKGROUND: + + I started coding this because I couldn't find a fixed point FFT that didn't +use assembly code. I started with floating point numbers so I could get the +theory straight before working on fixed point issues. In the end, I had a +little bit of code that could be recompiled easily to do ffts with short, float +or double (other types should be easy too). + + Once I got my FFT working, I was curious about the speed compared to +a well respected and highly optimized fft library. I don't want to criticize +this great library, so let's call it FFT_BRANDX. +During this process, I learned: + + 1. FFT_BRANDX has more than 100K lines of code. The core of kiss_fft is about 500 lines (cpx 1-d). + 2. It took me an embarrassingly long time to get FFT_BRANDX working. + 3. A simple program using FFT_BRANDX is 522KB. A similar program using kiss_fft is 18KB (without optimizing for size). + 4. FFT_BRANDX is roughly twice as fast as KISS FFT in default mode. + + It is wonderful that free, highly optimized libraries like FFT_BRANDX exist. +But such libraries carry a huge burden of complexity necessary to extract every +last bit of performance. + + Sometimes simpler is better, even if it's not better. + +FREQUENTLY ASKED QUESTIONS: + Q: Can I use kissfft in a project with a ___ license? + A: Yes. See LICENSE below. + + Q: Why don't I get the output I expect? + A: The two most common causes of this are + 1) scaling : is there a constant multiplier between what you got and what you want? + 2) mixed build environment -- all code must be compiled with same preprocessor + definitions for FIXED_POINT and kiss_fft_scalar + + Q: Will you write/debug my code for me? + A: Probably not unless you pay me. I am happy to answer pointed and topical questions, but + I may refer you to a book, a forum, or some other resource. + + +PERFORMANCE: + (on Athlon XP 2100+, with gcc 2.96, float data type) + + Kiss performed 10000 1024-pt cpx ffts in .63 s of cpu time. + For comparison, it took md5sum twice as long to process the same amount of data. + + Transforming 5 minutes of CD quality audio takes less than a second (nfft=1024). + +DO NOT: + ... use Kiss if you need the Fastest Fourier Transform in the World + ... ask me to add features that will bloat the code + +UNDER THE HOOD: + + Kiss FFT uses a time decimation, mixed-radix, out-of-place FFT. If you give it an input buffer + and output buffer that are the same, a temporary buffer will be created to hold the data. + + No static data is used. The core routines of kiss_fft are thread-safe (but not all of the tools directory). + + No scaling is done for the floating point version (for speed). + Scaling is done both ways for the fixed-point version (for overflow prevention). + + Optimized butterflies are used for factors 2,3,4, and 5. + + The real (i.e. not complex) optimization code only works for even length ffts. It does two half-length + FFTs in parallel (packed into real&imag), and then combines them via twiddling. The result is + nfft/2+1 complex frequency bins from DC to Nyquist. If you don't know what this means, search the web. + + The fast convolution filtering uses the overlap-scrap method, slightly + modified to put the scrap at the tail. + +LICENSE: + Revised BSD License, see COPYING for verbiage. + Basically, "free to use&change, give credit where due, no guarantees" + Note this license is compatible with GPL at one end of the spectrum and closed, commercial software at + the other end. See http://www.fsf.org/licensing/licenses + + A commercial license is available which removes the requirement for attribution. Contact me for details. + + +TODO: + *) Add real optimization for odd length FFTs + *) Document/revisit the input/output fft scaling + *) Make doc describing the overlap (tail) scrap fast convolution filtering in kiss_fastfir.c + *) Test all the ./tools/ code with fixed point (kiss_fastfir.c doesn't work, maybe others) + +AUTHOR: + Mark Borgerding + Mark@Borgerding.net diff --git a/media/kiss_fft/README.simd b/media/kiss_fft/README.simd new file mode 100644 index 000000000..b0fdac550 --- /dev/null +++ b/media/kiss_fft/README.simd @@ -0,0 +1,78 @@ +If you are reading this, it means you think you may be interested in using the SIMD extensions in kissfft +to do 4 *separate* FFTs at once. + +Beware! Beyond here there be dragons! + +This API is not easy to use, is not well documented, and breaks the KISS principle. + + +Still reading? Okay, you may get rewarded for your patience with a considerable speedup +(2-3x) on intel x86 machines with SSE if you are willing to jump through some hoops. + +The basic idea is to use the packed 4 float __m128 data type as a scalar element. +This means that the format is pretty convoluted. It performs 4 FFTs per fft call on signals A,B,C,D. + +For complex data, the data is interlaced as follows: +rA0,rB0,rC0,rD0, iA0,iB0,iC0,iD0, rA1,rB1,rC1,rD1, iA1,iB1,iC1,iD1 ... +where "rA0" is the real part of the zeroth sample for signal A + +Real-only data is laid out: +rA0,rB0,rC0,rD0, rA1,rB1,rC1,rD1, ... + +Compile with gcc flags something like +-O3 -mpreferred-stack-boundary=4 -DUSE_SIMD=1 -msse + +Be aware of SIMD alignment. This is the most likely cause of segfaults. +The code within kissfft uses scratch variables on the stack. +With SIMD, these must have addresses on 16 byte boundaries. +Search on "SIMD alignment" for more info. + + + +Robin at Divide Concept was kind enough to share his code for formatting to/from the SIMD kissfft. +I have not run it -- use it at your own risk. It appears to do 4xN and Nx4 transpositions +(out of place). + +void SSETools::pack128(float* target, float* source, unsigned long size128) +{ + __m128* pDest = (__m128*)target; + __m128* pDestEnd = pDest+size128; + float* source0=source; + float* source1=source0+size128; + float* source2=source1+size128; + float* source3=source2+size128; + + while(pDest<pDestEnd) + { + *pDest=_mm_set_ps(*source3,*source2,*source1,*source0); + source0++; + source1++; + source2++; + source3++; + pDest++; + } +} + +void SSETools::unpack128(float* target, float* source, unsigned long size128) +{ + + float* pSrc = source; + float* pSrcEnd = pSrc+size128*4; + float* target0=target; + float* target1=target0+size128; + float* target2=target1+size128; + float* target3=target2+size128; + + while(pSrc<pSrcEnd) + { + *target0=pSrc[0]; + *target1=pSrc[1]; + *target2=pSrc[2]; + *target3=pSrc[3]; + target0++; + target1++; + target2++; + target3++; + pSrc+=4; + } +} diff --git a/media/kiss_fft/README_MOZILLA b/media/kiss_fft/README_MOZILLA new file mode 100644 index 000000000..a2fd35d6a --- /dev/null +++ b/media/kiss_fft/README_MOZILLA @@ -0,0 +1,8 @@ +The source from this directory was copied from the kissfft hg repository using +the update.sh script. The only changes made were those applied by update.sh +and the addition of moz.build and Makefile.in build files for the Mozilla build +system. + +The kissfft hg repository is: http://hg.code.sf.net/p/kissfft/code + +The hg revision ID used was fbe1bb0bc7b94ec252842b8b7e3f3347ec75d92f. diff --git a/media/kiss_fft/_kiss_fft_guts.h b/media/kiss_fft/_kiss_fft_guts.h new file mode 100644 index 000000000..ba6614440 --- /dev/null +++ b/media/kiss_fft/_kiss_fft_guts.h @@ -0,0 +1,164 @@ +/* +Copyright (c) 2003-2010, Mark Borgerding + +All rights reserved. + +Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. + * Neither the author nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +*/ + +/* kiss_fft.h + defines kiss_fft_scalar as either short or a float type + and defines + typedef struct { kiss_fft_scalar r; kiss_fft_scalar i; }kiss_fft_cpx; */ +#include "kiss_fft.h" +#include <limits.h> + +#define MAXFACTORS 32 +/* e.g. an fft of length 128 has 4 factors + as far as kissfft is concerned + 4*4*4*2 + */ + +struct kiss_fft_state{ + int nfft; + int inverse; + int factors[2*MAXFACTORS]; + kiss_fft_cpx twiddles[1]; +}; + +/* + Explanation of macros dealing with complex math: + + C_MUL(m,a,b) : m = a*b + C_FIXDIV( c , div ) : if a fixed point impl., c /= div. noop otherwise + C_SUB( res, a,b) : res = a - b + C_SUBFROM( res , a) : res -= a + C_ADDTO( res , a) : res += a + * */ +#ifdef FIXED_POINT +#if (FIXED_POINT==32) +# define FRACBITS 31 +# define SAMPPROD int64_t +#define SAMP_MAX 2147483647 +#else +# define FRACBITS 15 +# define SAMPPROD int32_t +#define SAMP_MAX 32767 +#endif + +#define SAMP_MIN -SAMP_MAX + +#if defined(CHECK_OVERFLOW) +# define CHECK_OVERFLOW_OP(a,op,b) \ + if ( (SAMPPROD)(a) op (SAMPPROD)(b) > SAMP_MAX || (SAMPPROD)(a) op (SAMPPROD)(b) < SAMP_MIN ) { \ + fprintf(stderr,"WARNING:overflow @ " __FILE__ "(%d): (%d " #op" %d) = %ld\n",__LINE__,(a),(b),(SAMPPROD)(a) op (SAMPPROD)(b) ); } +#endif + + +# define smul(a,b) ( (SAMPPROD)(a)*(b) ) +# define sround( x ) (kiss_fft_scalar)( ( (x) + (1<<(FRACBITS-1)) ) >> FRACBITS ) + +# define S_MUL(a,b) sround( smul(a,b) ) + +# define C_MUL(m,a,b) \ + do{ (m).r = sround( smul((a).r,(b).r) - smul((a).i,(b).i) ); \ + (m).i = sround( smul((a).r,(b).i) + smul((a).i,(b).r) ); }while(0) + +# define DIVSCALAR(x,k) \ + (x) = sround( smul( x, SAMP_MAX/k ) ) + +# define C_FIXDIV(c,div) \ + do { DIVSCALAR( (c).r , div); \ + DIVSCALAR( (c).i , div); }while (0) + +# define C_MULBYSCALAR( c, s ) \ + do{ (c).r = sround( smul( (c).r , s ) ) ;\ + (c).i = sround( smul( (c).i , s ) ) ; }while(0) + +#else /* not FIXED_POINT*/ + +# define S_MUL(a,b) ( (a)*(b) ) +#define C_MUL(m,a,b) \ + do{ (m).r = (a).r*(b).r - (a).i*(b).i;\ + (m).i = (a).r*(b).i + (a).i*(b).r; }while(0) +# define C_FIXDIV(c,div) /* NOOP */ +# define C_MULBYSCALAR( c, s ) \ + do{ (c).r *= (s);\ + (c).i *= (s); }while(0) +#endif + +#ifndef CHECK_OVERFLOW_OP +# define CHECK_OVERFLOW_OP(a,op,b) /* noop */ +#endif + +#define C_ADD( res, a,b)\ + do { \ + CHECK_OVERFLOW_OP((a).r,+,(b).r)\ + CHECK_OVERFLOW_OP((a).i,+,(b).i)\ + (res).r=(a).r+(b).r; (res).i=(a).i+(b).i; \ + }while(0) +#define C_SUB( res, a,b)\ + do { \ + CHECK_OVERFLOW_OP((a).r,-,(b).r)\ + CHECK_OVERFLOW_OP((a).i,-,(b).i)\ + (res).r=(a).r-(b).r; (res).i=(a).i-(b).i; \ + }while(0) +#define C_ADDTO( res , a)\ + do { \ + CHECK_OVERFLOW_OP((res).r,+,(a).r)\ + CHECK_OVERFLOW_OP((res).i,+,(a).i)\ + (res).r += (a).r; (res).i += (a).i;\ + }while(0) + +#define C_SUBFROM( res , a)\ + do {\ + CHECK_OVERFLOW_OP((res).r,-,(a).r)\ + CHECK_OVERFLOW_OP((res).i,-,(a).i)\ + (res).r -= (a).r; (res).i -= (a).i; \ + }while(0) + + +#ifdef FIXED_POINT +# define KISS_FFT_COS(phase) floor(.5+SAMP_MAX * cos (phase)) +# define KISS_FFT_SIN(phase) floor(.5+SAMP_MAX * sin (phase)) +# define HALF_OF(x) ((x)>>1) +#elif defined(USE_SIMD) +# define KISS_FFT_COS(phase) _mm_set1_ps( cos(phase) ) +# define KISS_FFT_SIN(phase) _mm_set1_ps( sin(phase) ) +# define HALF_OF(x) ((x)*_mm_set1_ps(.5)) +#else +# define KISS_FFT_COS(phase) (kiss_fft_scalar) cos(phase) +# define KISS_FFT_SIN(phase) (kiss_fft_scalar) sin(phase) +# define HALF_OF(x) ((x)*.5) +#endif + +#define kf_cexp(x,phase) \ + do{ \ + (x)->r = KISS_FFT_COS(phase);\ + (x)->i = KISS_FFT_SIN(phase);\ + }while(0) + + +/* a debugging function */ +#define pcpx(c)\ + fprintf(stderr,"%g + %gi\n",(double)((c)->r),(double)((c)->i) ) + + +#ifdef KISS_FFT_USE_ALLOCA +// define this to allow use of alloca instead of malloc for temporary buffers +// Temporary buffers are used in two case: +// 1. FFT sizes that have "bad" factors. i.e. not 2,3 and 5 +// 2. "in-place" FFTs. Notice the quotes, since kissfft does not really do an in-place transform. +#include <alloca.h> +#define KISS_FFT_TMP_ALLOC(nbytes) alloca(nbytes) +#define KISS_FFT_TMP_FREE(ptr) +#else +#define KISS_FFT_TMP_ALLOC(nbytes) KISS_FFT_MALLOC(nbytes) +#define KISS_FFT_TMP_FREE(ptr) KISS_FFT_FREE(ptr) +#endif diff --git a/media/kiss_fft/kiss_fft.c b/media/kiss_fft/kiss_fft.c new file mode 100644 index 000000000..465d6c97a --- /dev/null +++ b/media/kiss_fft/kiss_fft.c @@ -0,0 +1,408 @@ +/* +Copyright (c) 2003-2010, Mark Borgerding + +All rights reserved. + +Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. + * Neither the author nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +*/ + + +#include "_kiss_fft_guts.h" +/* The guts header contains all the multiplication and addition macros that are defined for + fixed or floating point complex numbers. It also delares the kf_ internal functions. + */ + +static void kf_bfly2( + kiss_fft_cpx * Fout, + const size_t fstride, + const kiss_fft_cfg st, + int m + ) +{ + kiss_fft_cpx * Fout2; + kiss_fft_cpx * tw1 = st->twiddles; + kiss_fft_cpx t; + Fout2 = Fout + m; + do{ + C_FIXDIV(*Fout,2); C_FIXDIV(*Fout2,2); + + C_MUL (t, *Fout2 , *tw1); + tw1 += fstride; + C_SUB( *Fout2 , *Fout , t ); + C_ADDTO( *Fout , t ); + ++Fout2; + ++Fout; + }while (--m); +} + +static void kf_bfly4( + kiss_fft_cpx * Fout, + const size_t fstride, + const kiss_fft_cfg st, + const size_t m + ) +{ + kiss_fft_cpx *tw1,*tw2,*tw3; + kiss_fft_cpx scratch[6]; + size_t k=m; + const size_t m2=2*m; + const size_t m3=3*m; + + + tw3 = tw2 = tw1 = st->twiddles; + + do { + C_FIXDIV(*Fout,4); C_FIXDIV(Fout[m],4); C_FIXDIV(Fout[m2],4); C_FIXDIV(Fout[m3],4); + + C_MUL(scratch[0],Fout[m] , *tw1 ); + C_MUL(scratch[1],Fout[m2] , *tw2 ); + C_MUL(scratch[2],Fout[m3] , *tw3 ); + + C_SUB( scratch[5] , *Fout, scratch[1] ); + C_ADDTO(*Fout, scratch[1]); + C_ADD( scratch[3] , scratch[0] , scratch[2] ); + C_SUB( scratch[4] , scratch[0] , scratch[2] ); + C_SUB( Fout[m2], *Fout, scratch[3] ); + tw1 += fstride; + tw2 += fstride*2; + tw3 += fstride*3; + C_ADDTO( *Fout , scratch[3] ); + + if(st->inverse) { + Fout[m].r = scratch[5].r - scratch[4].i; + Fout[m].i = scratch[5].i + scratch[4].r; + Fout[m3].r = scratch[5].r + scratch[4].i; + Fout[m3].i = scratch[5].i - scratch[4].r; + }else{ + Fout[m].r = scratch[5].r + scratch[4].i; + Fout[m].i = scratch[5].i - scratch[4].r; + Fout[m3].r = scratch[5].r - scratch[4].i; + Fout[m3].i = scratch[5].i + scratch[4].r; + } + ++Fout; + }while(--k); +} + +static void kf_bfly3( + kiss_fft_cpx * Fout, + const size_t fstride, + const kiss_fft_cfg st, + size_t m + ) +{ + size_t k=m; + const size_t m2 = 2*m; + kiss_fft_cpx *tw1,*tw2; + kiss_fft_cpx scratch[5]; + kiss_fft_cpx epi3; + epi3 = st->twiddles[fstride*m]; + + tw1=tw2=st->twiddles; + + do{ + C_FIXDIV(*Fout,3); C_FIXDIV(Fout[m],3); C_FIXDIV(Fout[m2],3); + + C_MUL(scratch[1],Fout[m] , *tw1); + C_MUL(scratch[2],Fout[m2] , *tw2); + + C_ADD(scratch[3],scratch[1],scratch[2]); + C_SUB(scratch[0],scratch[1],scratch[2]); + tw1 += fstride; + tw2 += fstride*2; + + Fout[m].r = Fout->r - HALF_OF(scratch[3].r); + Fout[m].i = Fout->i - HALF_OF(scratch[3].i); + + C_MULBYSCALAR( scratch[0] , epi3.i ); + + C_ADDTO(*Fout,scratch[3]); + + Fout[m2].r = Fout[m].r + scratch[0].i; + Fout[m2].i = Fout[m].i - scratch[0].r; + + Fout[m].r -= scratch[0].i; + Fout[m].i += scratch[0].r; + + ++Fout; + }while(--k); +} + +static void kf_bfly5( + kiss_fft_cpx * Fout, + const size_t fstride, + const kiss_fft_cfg st, + int m + ) +{ + kiss_fft_cpx *Fout0,*Fout1,*Fout2,*Fout3,*Fout4; + int u; + kiss_fft_cpx scratch[13]; + kiss_fft_cpx * twiddles = st->twiddles; + kiss_fft_cpx *tw; + kiss_fft_cpx ya,yb; + ya = twiddles[fstride*m]; + yb = twiddles[fstride*2*m]; + + Fout0=Fout; + Fout1=Fout0+m; + Fout2=Fout0+2*m; + Fout3=Fout0+3*m; + Fout4=Fout0+4*m; + + tw=st->twiddles; + for ( u=0; u<m; ++u ) { + C_FIXDIV( *Fout0,5); C_FIXDIV( *Fout1,5); C_FIXDIV( *Fout2,5); C_FIXDIV( *Fout3,5); C_FIXDIV( *Fout4,5); + scratch[0] = *Fout0; + + C_MUL(scratch[1] ,*Fout1, tw[u*fstride]); + C_MUL(scratch[2] ,*Fout2, tw[2*u*fstride]); + C_MUL(scratch[3] ,*Fout3, tw[3*u*fstride]); + C_MUL(scratch[4] ,*Fout4, tw[4*u*fstride]); + + C_ADD( scratch[7],scratch[1],scratch[4]); + C_SUB( scratch[10],scratch[1],scratch[4]); + C_ADD( scratch[8],scratch[2],scratch[3]); + C_SUB( scratch[9],scratch[2],scratch[3]); + + Fout0->r += scratch[7].r + scratch[8].r; + Fout0->i += scratch[7].i + scratch[8].i; + + scratch[5].r = scratch[0].r + S_MUL(scratch[7].r,ya.r) + S_MUL(scratch[8].r,yb.r); + scratch[5].i = scratch[0].i + S_MUL(scratch[7].i,ya.r) + S_MUL(scratch[8].i,yb.r); + + scratch[6].r = S_MUL(scratch[10].i,ya.i) + S_MUL(scratch[9].i,yb.i); + scratch[6].i = -S_MUL(scratch[10].r,ya.i) - S_MUL(scratch[9].r,yb.i); + + C_SUB(*Fout1,scratch[5],scratch[6]); + C_ADD(*Fout4,scratch[5],scratch[6]); + + scratch[11].r = scratch[0].r + S_MUL(scratch[7].r,yb.r) + S_MUL(scratch[8].r,ya.r); + scratch[11].i = scratch[0].i + S_MUL(scratch[7].i,yb.r) + S_MUL(scratch[8].i,ya.r); + scratch[12].r = - S_MUL(scratch[10].i,yb.i) + S_MUL(scratch[9].i,ya.i); + scratch[12].i = S_MUL(scratch[10].r,yb.i) - S_MUL(scratch[9].r,ya.i); + + C_ADD(*Fout2,scratch[11],scratch[12]); + C_SUB(*Fout3,scratch[11],scratch[12]); + + ++Fout0;++Fout1;++Fout2;++Fout3;++Fout4; + } +} + +/* perform the butterfly for one stage of a mixed radix FFT */ +static void kf_bfly_generic( + kiss_fft_cpx * Fout, + const size_t fstride, + const kiss_fft_cfg st, + int m, + int p + ) +{ + int u,k,q1,q; + kiss_fft_cpx * twiddles = st->twiddles; + kiss_fft_cpx t; + int Norig = st->nfft; + + kiss_fft_cpx * scratch = (kiss_fft_cpx*)KISS_FFT_TMP_ALLOC(sizeof(kiss_fft_cpx)*p); + + for ( u=0; u<m; ++u ) { + k=u; + for ( q1=0 ; q1<p ; ++q1 ) { + scratch[q1] = Fout[ k ]; + C_FIXDIV(scratch[q1],p); + k += m; + } + + k=u; + for ( q1=0 ; q1<p ; ++q1 ) { + int twidx=0; + Fout[ k ] = scratch[0]; + for (q=1;q<p;++q ) { + twidx += fstride * k; + if (twidx>=Norig) twidx-=Norig; + C_MUL(t,scratch[q] , twiddles[twidx] ); + C_ADDTO( Fout[ k ] ,t); + } + k += m; + } + } + KISS_FFT_TMP_FREE(scratch); +} + +static +void kf_work( + kiss_fft_cpx * Fout, + const kiss_fft_cpx * f, + const size_t fstride, + int in_stride, + int * factors, + const kiss_fft_cfg st + ) +{ + kiss_fft_cpx * Fout_beg=Fout; + const int p=*factors++; /* the radix */ + const int m=*factors++; /* stage's fft length/p */ + const kiss_fft_cpx * Fout_end = Fout + p*m; + +#ifdef _OPENMP + // use openmp extensions at the + // top-level (not recursive) + if (fstride==1 && p<=5) + { + int k; + + // execute the p different work units in different threads +# pragma omp parallel for + for (k=0;k<p;++k) + kf_work( Fout +k*m, f+ fstride*in_stride*k,fstride*p,in_stride,factors,st); + // all threads have joined by this point + + switch (p) { + case 2: kf_bfly2(Fout,fstride,st,m); break; + case 3: kf_bfly3(Fout,fstride,st,m); break; + case 4: kf_bfly4(Fout,fstride,st,m); break; + case 5: kf_bfly5(Fout,fstride,st,m); break; + default: kf_bfly_generic(Fout,fstride,st,m,p); break; + } + return; + } +#endif + + if (m==1) { + do{ + *Fout = *f; + f += fstride*in_stride; + }while(++Fout != Fout_end ); + }else{ + do{ + // recursive call: + // DFT of size m*p performed by doing + // p instances of smaller DFTs of size m, + // each one takes a decimated version of the input + kf_work( Fout , f, fstride*p, in_stride, factors,st); + f += fstride*in_stride; + }while( (Fout += m) != Fout_end ); + } + + Fout=Fout_beg; + + // recombine the p smaller DFTs + switch (p) { + case 2: kf_bfly2(Fout,fstride,st,m); break; + case 3: kf_bfly3(Fout,fstride,st,m); break; + case 4: kf_bfly4(Fout,fstride,st,m); break; + case 5: kf_bfly5(Fout,fstride,st,m); break; + default: kf_bfly_generic(Fout,fstride,st,m,p); break; + } +} + +/* facbuf is populated by p1,m1,p2,m2, ... + where + p[i] * m[i] = m[i-1] + m0 = n */ +static +void kf_factor(int n,int * facbuf) +{ + int p=4; + double floor_sqrt; + floor_sqrt = floor( sqrt((double)n) ); + + /*factor out powers of 4, powers of 2, then any remaining primes */ + do { + while (n % p) { + switch (p) { + case 4: p = 2; break; + case 2: p = 3; break; + default: p += 2; break; + } + if (p > floor_sqrt) + p = n; /* no more factors, skip to end */ + } + n /= p; + *facbuf++ = p; + *facbuf++ = n; + } while (n > 1); +} + +/* + * + * User-callable function to allocate all necessary storage space for the fft. + * + * The return value is a contiguous block of memory, allocated with malloc. As such, + * It can be freed with free(), rather than a kiss_fft-specific function. + * */ +kiss_fft_cfg kiss_fft_alloc(int nfft,int inverse_fft,void * mem,size_t * lenmem ) +{ + kiss_fft_cfg st=NULL; + size_t memneeded = sizeof(struct kiss_fft_state) + + sizeof(kiss_fft_cpx)*(nfft-1); /* twiddle factors*/ + + if ( lenmem==NULL ) { + st = ( kiss_fft_cfg)KISS_FFT_MALLOC( memneeded ); + }else{ + if (mem != NULL && *lenmem >= memneeded) + st = (kiss_fft_cfg)mem; + *lenmem = memneeded; + } + if (st) { + int i; + st->nfft=nfft; + st->inverse = inverse_fft; + + for (i=0;i<nfft;++i) { + const double pi=3.141592653589793238462643383279502884197169399375105820974944; + double phase = -2*pi*i / nfft; + if (st->inverse) + phase *= -1; + kf_cexp(st->twiddles+i, phase ); + } + + kf_factor(nfft,st->factors); + } + return st; +} + + +void kiss_fft_stride(kiss_fft_cfg st,const kiss_fft_cpx *fin,kiss_fft_cpx *fout,int in_stride) +{ + if (fin == fout) { + //NOTE: this is not really an in-place FFT algorithm. + //It just performs an out-of-place FFT into a temp buffer + kiss_fft_cpx * tmpbuf = (kiss_fft_cpx*)KISS_FFT_TMP_ALLOC( sizeof(kiss_fft_cpx)*st->nfft); + kf_work(tmpbuf,fin,1,in_stride, st->factors,st); + memcpy(fout,tmpbuf,sizeof(kiss_fft_cpx)*st->nfft); + KISS_FFT_TMP_FREE(tmpbuf); + }else{ + kf_work( fout, fin, 1,in_stride, st->factors,st ); + } +} + +void kiss_fft(kiss_fft_cfg cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout) +{ + kiss_fft_stride(cfg,fin,fout,1); +} + + +void kiss_fft_cleanup(void) +{ + // nothing needed any more +} + +int kiss_fft_next_fast_size(int n) +{ + while(1) { + int m=n; + while ( (m%2) == 0 ) m/=2; + while ( (m%3) == 0 ) m/=3; + while ( (m%5) == 0 ) m/=5; + if (m<=1) + break; /* n is completely factorable by twos, threes, and fives */ + n++; + } + return n; +} diff --git a/media/kiss_fft/kiss_fft.h b/media/kiss_fft/kiss_fft.h new file mode 100644 index 000000000..64c50f4aa --- /dev/null +++ b/media/kiss_fft/kiss_fft.h @@ -0,0 +1,124 @@ +#ifndef KISS_FFT_H +#define KISS_FFT_H + +#include <stdlib.h> +#include <stdio.h> +#include <math.h> +#include <string.h> + +#ifdef __cplusplus +extern "C" { +#endif + +/* + ATTENTION! + If you would like a : + -- a utility that will handle the caching of fft objects + -- real-only (no imaginary time component ) FFT + -- a multi-dimensional FFT + -- a command-line utility to perform ffts + -- a command-line utility to perform fast-convolution filtering + + Then see kfc.h kiss_fftr.h kiss_fftnd.h fftutil.c kiss_fastfir.c + in the tools/ directory. +*/ + +#ifdef USE_SIMD +# include <xmmintrin.h> +# define kiss_fft_scalar __m128 +#define KISS_FFT_MALLOC(nbytes) _mm_malloc(nbytes,16) +#define KISS_FFT_FREE _mm_free +#else +#define KISS_FFT_MALLOC malloc +#define KISS_FFT_FREE free +#endif + + +#ifdef FIXED_POINT +#include <sys/types.h> +# if (FIXED_POINT == 32) +# define kiss_fft_scalar int32_t +# else +# define kiss_fft_scalar int16_t +# endif +#else +# ifndef kiss_fft_scalar +/* default is float */ +# define kiss_fft_scalar float +# endif +#endif + +typedef struct { + kiss_fft_scalar r; + kiss_fft_scalar i; +}kiss_fft_cpx; + +typedef struct kiss_fft_state* kiss_fft_cfg; + +/* + * kiss_fft_alloc + * + * Initialize a FFT (or IFFT) algorithm's cfg/state buffer. + * + * typical usage: kiss_fft_cfg mycfg=kiss_fft_alloc(1024,0,NULL,NULL); + * + * The return value from fft_alloc is a cfg buffer used internally + * by the fft routine or NULL. + * + * If lenmem is NULL, then kiss_fft_alloc will allocate a cfg buffer using malloc. + * The returned value should be free()d when done to avoid memory leaks. + * + * The state can be placed in a user supplied buffer 'mem': + * If lenmem is not NULL and mem is not NULL and *lenmem is large enough, + * then the function places the cfg in mem and the size used in *lenmem + * and returns mem. + * + * If lenmem is not NULL and ( mem is NULL or *lenmem is not large enough), + * then the function returns NULL and places the minimum cfg + * buffer size in *lenmem. + * */ + +kiss_fft_cfg kiss_fft_alloc(int nfft,int inverse_fft,void * mem,size_t * lenmem); + +/* + * kiss_fft(cfg,in_out_buf) + * + * Perform an FFT on a complex input buffer. + * for a forward FFT, + * fin should be f[0] , f[1] , ... ,f[nfft-1] + * fout will be F[0] , F[1] , ... ,F[nfft-1] + * Note that each element is complex and can be accessed like + f[k].r and f[k].i + * */ +void kiss_fft(kiss_fft_cfg cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout); + +/* + A more generic version of the above function. It reads its input from every Nth sample. + * */ +void kiss_fft_stride(kiss_fft_cfg cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout,int fin_stride); + +/* If kiss_fft_alloc allocated a buffer, it is one contiguous + buffer and can be simply free()d when no longer needed*/ +#define kiss_fft_free free + +/* + Cleans up some memory that gets managed internally. Not necessary to call, but it might clean up + your compiler output to call this before you exit. +*/ +void kiss_fft_cleanup(void); + + +/* + * Returns the smallest integer k, such that k>=n and k has only "fast" factors (2,3,5) + */ +int kiss_fft_next_fast_size(int n); + +/* for real ffts, we need an even size */ +#define kiss_fftr_next_fast_size_real(n) \ + (kiss_fft_next_fast_size( ((n)+1)>>1)<<1) + +#ifdef __cplusplus +} +#endif + +#endif diff --git a/media/kiss_fft/kiss_fftr.c b/media/kiss_fft/kiss_fftr.c new file mode 100644 index 000000000..b8e238b1e --- /dev/null +++ b/media/kiss_fft/kiss_fftr.c @@ -0,0 +1,159 @@ +/* +Copyright (c) 2003-2004, Mark Borgerding + +All rights reserved. + +Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. + * Neither the author nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +*/ + +#include "kiss_fftr.h" +#include "_kiss_fft_guts.h" + +struct kiss_fftr_state{ + kiss_fft_cfg substate; + kiss_fft_cpx * tmpbuf; + kiss_fft_cpx * super_twiddles; +#ifdef USE_SIMD + void * pad; +#endif +}; + +kiss_fftr_cfg kiss_fftr_alloc(int nfft,int inverse_fft,void * mem,size_t * lenmem) +{ + int i; + kiss_fftr_cfg st = NULL; + size_t subsize, memneeded; + + if (nfft & 1) { + fprintf(stderr,"Real FFT optimization must be even.\n"); + return NULL; + } + nfft >>= 1; + + kiss_fft_alloc (nfft, inverse_fft, NULL, &subsize); + memneeded = sizeof(struct kiss_fftr_state) + subsize + sizeof(kiss_fft_cpx) * ( nfft * 3 / 2); + + if (lenmem == NULL) { + st = (kiss_fftr_cfg) KISS_FFT_MALLOC (memneeded); + } else { + if (*lenmem >= memneeded) + st = (kiss_fftr_cfg) mem; + *lenmem = memneeded; + } + if (!st) + return NULL; + + st->substate = (kiss_fft_cfg) (st + 1); /*just beyond kiss_fftr_state struct */ + st->tmpbuf = (kiss_fft_cpx *) (((char *) st->substate) + subsize); + st->super_twiddles = st->tmpbuf + nfft; + kiss_fft_alloc(nfft, inverse_fft, st->substate, &subsize); + + for (i = 0; i < nfft/2; ++i) { + double phase = + -3.14159265358979323846264338327 * ((double) (i+1) / nfft + .5); + if (inverse_fft) + phase *= -1; + kf_cexp (st->super_twiddles+i,phase); + } + return st; +} + +void kiss_fftr(kiss_fftr_cfg st,const kiss_fft_scalar *timedata,kiss_fft_cpx *freqdata) +{ + /* input buffer timedata is stored row-wise */ + int k,ncfft; + kiss_fft_cpx fpnk,fpk,f1k,f2k,tw,tdc; + + if ( st->substate->inverse) { + fprintf(stderr,"kiss fft usage error: improper alloc\n"); + exit(1); + } + + ncfft = st->substate->nfft; + + /*perform the parallel fft of two real signals packed in real,imag*/ + kiss_fft( st->substate , (const kiss_fft_cpx*)timedata, st->tmpbuf ); + /* The real part of the DC element of the frequency spectrum in st->tmpbuf + * contains the sum of the even-numbered elements of the input time sequence + * The imag part is the sum of the odd-numbered elements + * + * The sum of tdc.r and tdc.i is the sum of the input time sequence. + * yielding DC of input time sequence + * The difference of tdc.r - tdc.i is the sum of the input (dot product) [1,-1,1,-1... + * yielding Nyquist bin of input time sequence + */ + + tdc.r = st->tmpbuf[0].r; + tdc.i = st->tmpbuf[0].i; + C_FIXDIV(tdc,2); + CHECK_OVERFLOW_OP(tdc.r ,+, tdc.i); + CHECK_OVERFLOW_OP(tdc.r ,-, tdc.i); + freqdata[0].r = tdc.r + tdc.i; + freqdata[ncfft].r = tdc.r - tdc.i; +#ifdef USE_SIMD + freqdata[ncfft].i = freqdata[0].i = _mm_set1_ps(0); +#else + freqdata[ncfft].i = freqdata[0].i = 0; +#endif + + for ( k=1;k <= ncfft/2 ; ++k ) { + fpk = st->tmpbuf[k]; + fpnk.r = st->tmpbuf[ncfft-k].r; + fpnk.i = - st->tmpbuf[ncfft-k].i; + C_FIXDIV(fpk,2); + C_FIXDIV(fpnk,2); + + C_ADD( f1k, fpk , fpnk ); + C_SUB( f2k, fpk , fpnk ); + C_MUL( tw , f2k , st->super_twiddles[k-1]); + + freqdata[k].r = HALF_OF(f1k.r + tw.r); + freqdata[k].i = HALF_OF(f1k.i + tw.i); + freqdata[ncfft-k].r = HALF_OF(f1k.r - tw.r); + freqdata[ncfft-k].i = HALF_OF(tw.i - f1k.i); + } +} + +void kiss_fftri(kiss_fftr_cfg st,const kiss_fft_cpx *freqdata,kiss_fft_scalar *timedata) +{ + /* input buffer timedata is stored row-wise */ + int k, ncfft; + + if (st->substate->inverse == 0) { + fprintf (stderr, "kiss fft usage error: improper alloc\n"); + exit (1); + } + + ncfft = st->substate->nfft; + + st->tmpbuf[0].r = freqdata[0].r + freqdata[ncfft].r; + st->tmpbuf[0].i = freqdata[0].r - freqdata[ncfft].r; + C_FIXDIV(st->tmpbuf[0],2); + + for (k = 1; k <= ncfft / 2; ++k) { + kiss_fft_cpx fk, fnkc, fek, fok, tmp; + fk = freqdata[k]; + fnkc.r = freqdata[ncfft - k].r; + fnkc.i = -freqdata[ncfft - k].i; + C_FIXDIV( fk , 2 ); + C_FIXDIV( fnkc , 2 ); + + C_ADD (fek, fk, fnkc); + C_SUB (tmp, fk, fnkc); + C_MUL (fok, tmp, st->super_twiddles[k-1]); + C_ADD (st->tmpbuf[k], fek, fok); + C_SUB (st->tmpbuf[ncfft - k], fek, fok); +#ifdef USE_SIMD + st->tmpbuf[ncfft - k].i *= _mm_set1_ps(-1.0); +#else + st->tmpbuf[ncfft - k].i *= -1; +#endif + } + kiss_fft (st->substate, st->tmpbuf, (kiss_fft_cpx *) timedata); +} diff --git a/media/kiss_fft/kiss_fftr.h b/media/kiss_fft/kiss_fftr.h new file mode 100644 index 000000000..72e5a5771 --- /dev/null +++ b/media/kiss_fft/kiss_fftr.h @@ -0,0 +1,46 @@ +#ifndef KISS_FTR_H +#define KISS_FTR_H + +#include "kiss_fft.h" +#ifdef __cplusplus +extern "C" { +#endif + + +/* + + Real optimized version can save about 45% cpu time vs. complex fft of a real seq. + + + + */ + +typedef struct kiss_fftr_state *kiss_fftr_cfg; + + +kiss_fftr_cfg kiss_fftr_alloc(int nfft,int inverse_fft,void * mem, size_t * lenmem); +/* + nfft must be even + + If you don't care to allocate space, use mem = lenmem = NULL +*/ + + +void kiss_fftr(kiss_fftr_cfg cfg,const kiss_fft_scalar *timedata,kiss_fft_cpx *freqdata); +/* + input timedata has nfft scalar points + output freqdata has nfft/2+1 complex points +*/ + +void kiss_fftri(kiss_fftr_cfg cfg,const kiss_fft_cpx *freqdata,kiss_fft_scalar *timedata); +/* + input freqdata has nfft/2+1 complex points + output timedata has nfft scalar points +*/ + +#define kiss_fftr_free free + +#ifdef __cplusplus +} +#endif +#endif diff --git a/media/kiss_fft/moz.build b/media/kiss_fft/moz.build new file mode 100644 index 000000000..798240b91 --- /dev/null +++ b/media/kiss_fft/moz.build @@ -0,0 +1,17 @@ +# -*- Mode: python; indent-tabs-mode: nil; tab-width: 40 -*- +# vim: set filetype=python: +# This Source Code Form is subject to the terms of the Mozilla Public +# License, v. 2.0. If a copy of the MPL was not distributed with this +# file, You can obtain one at http://mozilla.org/MPL/2.0/. + +EXPORTS.kiss_fft += [ + 'kiss_fft.h', + 'kiss_fftr.h', +] + +SOURCES += [ + 'kiss_fft.c', + 'kiss_fftr.c', +] + +FINAL_LIBRARY = 'xul' diff --git a/media/kiss_fft/update.sh b/media/kiss_fft/update.sh new file mode 100755 index 000000000..e11a7406d --- /dev/null +++ b/media/kiss_fft/update.sh @@ -0,0 +1,34 @@ +# Usage: sh update.sh <upstream_src_directory> +set -e + +if [ ! -d "$1" ]; then + echo "Usage: ./update.sh /path/to/kiss_fft" > 2 + exit 1 +fi + +FILES="CHANGELOG \ + COPYING \ + README \ + README.simd \ + _kiss_fft_guts.h \ + kiss_fft.c \ + kiss_fft.h \ + tools/kiss_fftr.c \ + tools/kiss_fftr.h" + +for file in $FILES; do + cp "$1/$file" . +done + +if [ -d "$1/.hg" ]; then + rev=$(cd "$1" && hg log --template='{node}' -r `hg identify -i`) +fi + +if [ -n "$rev" ]; then + version=$rev + sed -i.bak -e "/The hg revision ID used was/ s/[0-9a-f]\{40\}\./$version./" README_MOZILLA + rm README_MOZILLA.bak +else + echo "Remember to update README_MOZILLA with the version details." +fi + |