similar to: FFT Resampler

Alexander Chemeris wrote: > Hi, > > Here are some questions from user point of view. :) > > On 5/29/08, Thorvald Natvig <thorvald at natvig.com> wrote: > >> I've done listening tests when converting wb_male.wav to 44.1, 48 and 8khz, >> and there aren't any obvious artifacts. I also did a 16=>16 test, and the >> results are delayed by 10ms

On 5/29/08, Thorvald Natvig <thorvald at natvig.com> wrote: > Alexander Chemeris wrote: > > On 5/29/08, Thorvald Natvig <thorvald at natvig.com> wrote: > > > I've done listening tests when converting wb_male.wav to 44.1, 48 and 8khz, > > > and there aren't any obvious artifacts. I also did a 16=>16 test, and the > > > results are delayed

Hi, Here are some questions from user point of view. :) On 5/29/08, Thorvald Natvig <thorvald at natvig.com> wrote: > I've done listening tests when converting wb_male.wav to 44.1, 48 and 8khz, > and there aren't any obvious artifacts. I also did a 16=>16 test, and the > results are delayed by 10ms and within +/- 1 (basically, rounding errors > from the FFT). Do

>> Yes, I plan to use it in a VoIP environment if I can get latency reduced to >> an acceptable level :) >> The latency depends directly on the overlap parameter, which also controls >> the quality. Higher quality => higher latency. You could set the overlap to >> 0, but that would give you some nasty artifacts. >> You can also resample with smaller block

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

Ok. I did some quality tests. First off; never do quality tests with ints. I had serious problems interpreting my results until it dawned on me that the signal differences were just 0 or 1. So, after a lot of scratching my head, these are done comparing the result from the _float versions (which is how both resamplers work internally anyway). What I did was this: Load speex_wb.wav as one

On 5/29/08, Thorvald Natvig <thorvald at natvig.com> wrote: > > > Yes, I plan to use it in a VoIP environment if I can get latency reduced to > > > an acceptable level :) > > > The latency depends directly on the overlap parameter, which also controls > > > the quality. Higher quality => higher latency. You could set the overlap to > > > 0,

And here's a patch for Intel Math Kernel Library. This allows commercial users of Speex to use a high-speed FFT library that isn't GPL'd. (You do need to pay for it though). This is 3 times faster than the default FFT in speex, and also faster than FFTW3 since MKL has native support for the complex packing Speex uses. Since Intel hasn't supplied any pkg-config files, and which

Using the following MATLAB snippet: fnames={'chirp','perfect','block','filter'} for k=1:length(fnames) fn=fnames{k}; myfile=fopen([fn, '.fl'], 'r', 'ieee-le'); x=fread(myfile, Inf, 'float32', 0, 'ieee-le'); fclose(myfile); X=specgram(x,2048,1,kaiser(2048,14)); spectrogram_floor=-96;

Hi, Attached is a patch which enables FFTW3 to be used by libspeexdsp. Using SSE, the FFT code is more than twice as fast, and it's 50% faster even without SSE. To enable the library, you'll need to call configure with '--with-gpl-fftw3'. Based on the benchmarking we did, this also changes the default FFT to smallft for floating point implementations when not using FFTW3.

Alexander Chemeris wrote: > Hi, > > On 5/27/08, Thorvald Natvig <thorvald at natvig.com> wrote: > >> And here's a patch for Intel Math Kernel Library. This allows commercial >> users of Speex to use a high-speed FFT library that isn't GPL'd. (You do >> need to pay for it though). This is 3 times faster than the default FFT in >> speex, and

Hi, I'm moving my "metadata" which accompanies some of the speech packets into the speex bits. However, as usual, I'm a bit confused. The manual (PDF from current SVN), section 5.10 (bottom of page 21) states that: "Finally, applications may define custom in-band messages using mode 13. The size of the message in bytes is encoded with 5 bits, so that the decoder can

Oh, I forgot to mention one thing. I currently append a few bits of custom information to each speex packet which I fetch out with things like speex_bits_unpack_unsigned(&sjJitter.current_packet, 1); It would be very usefull if the jitter buffer didn't actually decode the packet, but instead returned a pointer to it (or NULL if you should play silence and -1 if you should repeat the

Hi, I see you're changing the jitter buffer around quite a bit. Could you let us know when it's ready for general testing? (At the moment it doesn't handle missing packets at all) Best wishes, Thorvald

Jean-Marc Valin wrote: >> I recently found a .. weird bug on Win32 SSE with GCC 4.2.1. >> >> In libspeex/cb_search_sse.h, the following union is used: >> >> union { >> float __a[4]; >> __m128 __v; >> } __u; >> >> For some odd reason, this particular version of GCC will not 16-byte >> align the union. IE; the alignment

.. 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 +

Hy, I'm totaly out of ideas now. here are links to the code I use. codec.cpp http://barvanjekode.gama.us/temp/1078354945.html codec.h http://barvanjekode.gama.us/temp/135707080.html Variables I use are: int samplerate 32000 uint quality 10 uint complexity = 2 I get that wierd noise after I use speex encoder/decoder. It's like there where empty spaces between each encoded

Current SVN compilation fails on Win32 (Intel C++ compiler 9.0): ..\..\libspeex\mdf.c(317): error: declaration may not appear after executable statement in block float adapt_rate; Just moved it up to the variable declaration block to solve it locally, but as it's not valid C, I thought I'd mention it ;) PS: With the same compiler, AGC fails horribly (Zlast goes to several

Thorvald, re: At 03:18 AM 1/6/2006, Thorvald Natvig wrote: >I just checked it in the debugger, and this was with gcc 3.4.4 (mingw)... >And the addresses were not properly aligned :( From a bit of googling, >this seems to be a thread problem, as the gcc just maintains 16-byte >alignment of the stack -- if the start function of the thread had >misaligned stack, the misalignment

Could a method be added to use the VAD from the preprocessor to send the packet with the 'dtx_enable' flag in it, so the decoder on the other end knows it's in dtx mode? (and SPEEX_GET_DTX_STATUS returns 1) At the moment, I can hack around it by doing something like: if (VAD for prev frame && ! VAD for this frame) { iArg = 0;