similar to: Resampler, memory only variant

.. And a version without the API changes. -------------- next part -------------- Index: libspeex/resample_sse.h =================================================================== --- libspeex/resample_sse.h (revision 0) +++ libspeex/resample_sse.h (revision 0) @@ -0,0 +1,128 @@ +/* Copyright (C) 2002-2008 Jean-Marc Valin + * Copyright (C) 2008 Thorvald Natvig + */ +/** + @file resample_sse.h +

From: Jyri Sarha <jsarha at ti.com> Semantics of inner_product_single have also been changed to contain the final right shift and saturation so it can also be implemented in the optimal way for the used platform. This change affects fixed point calculations only. I also added a new fixed point macro SATURATE32PSHR(x, shift, a). It does pretty much the same thing as SATURATE32(PSHR32(x,

Hi, I've attached a patch against SVN r14645 which adds GCC visibility information to all symbols exported from libspeex.so and libspeexdsp.so. It includes a configure.ac change to test that both the compiler flags and __attribute__((visibility)) works, and if so will #define EXPORT __attribute__((visibility("default"))) and if not #define EXPORT I've attached a diff output

Hi, I've attached a patch against SVN r14645 which adds GCC visibility information to all symbols exported from libspeex.so and libspeexdsp.so. It includes a configure.ac change to test that both the compiler flags and __attribute__((visibility)) works, and if so will #define EXPORT __attribute__((visibility("default"))) and if not #define EXPORT I've attached a diff output

From: Thorvald Natvig <slicer at users.sourceforge.net> Some optimizing compilers miscompile the current SSE optimizations when full optimizations are enabled. By using output value pointer instead of a return value, we can bypass this misbehaviour. --- libspeex/resample.c | 8 ++++---- libspeex/resample_sse.h | 24 ++++++++---------------- 2 files changed, 12 insertions(+), 20

I'm using resample.c from speex in my project, and was testing it by processing the same waveform twice, with a reset() call in the middle, as follows: short *input_data = read_waveform_as_linear(infileName, &rate, &num_samples, &num_channels); SpeexResamplerState *state = speex_resampler_init_frac(num_channels, in_rate, out_rate, in_rate, out_rate, quality, &error);

From: Jyri Sarha <jsarha at ti.com> I optimized Speex resampler for NEON capable ARM CPUs. The first patch should speed up resampling on any platform that can spare the increased memory usage. It would be nice to have these merged to the master branch. Please let me know if there is anything I can do to help the the merge. The patches have been rebased on top of master branch in

Actually opus-tools-0.2 has the same issue. Sincerely Forrest Zhang -------------- next part -------------- Add "#define FIXED_POINT 1" into the top of resample.c, outputs of resample are all zeros. Actually the data transform is required before/after calling speex_resampler_process_native(). diff -Naupr libopusenc-0.2.1-vanilla/src/resample.c libopusenc-0.2.1/src/resample.c ---

There is a bug in "speex_resampler_process_interleaved_int" and "speex_resampler_process_interleaved_float" in the resampler. It seems that when "speex_resampler_process_int" is called for the first channel, "in_len" is set to the actual number of samples used (when output limited), which you would think would be OK to use for all the subsequent calls

From: Jyri Sarha <jsarha at ti.com> This bug happens quite often when resampling from a low to a high sample-rate with big enough factor. Also the resampling call has to be limited by the output buffer size and some unused samples needs be left in the input buffer. Sometimes when up-sampling with a big factor the resampling function wants to peek one more sample from the input buffer to

Need to adjust scale for floating input. diff -NaurbB a/libspeex/resample.c b/libspeex/resample.c --- a/libspeex/resample.c 2013-12-20 17:40:27.326576921 +0500 +++ b/libspeex/resample.c 2013-12-20 17:40:16.918577087 +0500 @@ -915,7 +915,7 @@ if (in) { for(j=0;j<ichunk;++j) #ifdef FIXED_POINT - x[j+st->filt_len-1]=WORD2INT(in[j*istride_save]); +

Hello, some misc. cleanup patches for speexdsp, nothing big I'm not sure about how to submit patches, so this is a test balloon :) ultimately, I'd like to fix the FIXED_POINT issue, see http://lists.xiph.org/pipermail/speex-dev/2013-December/008465.html currently, I think the only way to find out how speexdsp has been compiled is to resample some bytes and observe the output; which is

Look, how about you start from the testresample.c file? Oh, and you seem to be reading 320 samples at a time and processing 640, so that can't be good (and even beyond that your code is wrong for other reasons). Jean-Marc On 11/6/19 1:56 AM, zhouyuchen at iauto.com wrote: > Hello, > I printed the log, in/out len is not truncated, which means that the > input and output are not

Well, I don't quite understand why AC_C_BIGENDIAN and the solution you're proposing is likely to break other big endian machines (the ones that don't have __BIG_ENDIAN__). Can you send a patch that addresses that (i.e. still uses AC_C_BIGENDIAN when it works)? Jean-Marc Peter Grayson a ?crit : > Hi all, > > Speex currently decides endianness at configure-time. This causes

First of all, bufout_len = 320 is correct since that value is the number of stereo samples. Using 640 would be wrong (and no wonder that it crashes). Secondly, "speex_resampler_process_interleaved_float" has a bug that keeps it from working when it is output limited, so if you ever set "buf_len" to any value greater than 1764 the resampler will stop functioning properly

Be aware that inlen is an in/out parameter. It tells you how many samples the resampler read. That number can be smaller than what you passed -- in which case it means you need to buffer them and give them to the resampler again in the next call. Jean-Marc On 11/5/19 8:27 AM, zhouyuchen at iauto.com wrote: > Hello, > I have encountered some problems. I want to use speex to convert the

Hi Jean-Marc, Hi Jim, I have also seen some problems with the 1.1.8 release on the C55x. So far I have boiled down the issues to the following: 1) We need our own "fixed_xx.h" header file. I don't know why, and haven't had time to investigate, but there is a definite improvement when I use the attached fixed_c55x.h file which has turned all the maths into inline functions.

Hi Dan, Thanks for the patch, I'll apply it to svn. One question, why do you need this: +void kiss_fftri2(kiss_fftr_cfg cfg,const kiss_fft_scalar *freqdata,kiss_fft_scalar *timedata); + Cheers, Jean-Marc Dan Everton a ?crit : > Hi, > > I'm one of the people working on the Rockbox project > (http://www.rockbox.org) which is an open source alternative firmware > for a

Hi With 32-bit Xubuntu-14.04. When I compile FFmpeg from git using speex-1.2rc1 it builds OK. When I compile FFmpeg from git using speex-git it fails to build. (speex-HEAD-78cce67) It's not a big deal, I don't mind using speex-1.2rc1. But am I doing something wrong or is there something wrong with speex-git? (I've put the details in a txt attachment to keep this email post

Dear Tom, Short answer, if your using spec.pgram(), use the smoothing kernel to get a better estimate at the frequency centered in the bandwidth. If your frequency bin of interest is wider than the bandwidth of the kernel, average across frequencies (I think). The estimate appears to be normalized already. If you are calculating your PSD independently, then oversample (e.g. 2, perhaps 4 or more