similar to: Updated resampler patch

.. 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 +

Guys, Program stack is getting overflowed when I create and destroy encoder instance continuously for around 30 times. How can I ensure stack is completely freed when I destroy the encoder instance? I tried with VAR_ARRAYS. But Its giving me errors while I try to compile. I suppose VAR_ARRAYS allocates local variables as arrays so that they could go on to stack and finally when we leave the

Hi, Here's the (hopefully) final version of the resampler, now always using st->mem as the buffer area. It only allocates buffers on the stack when it's necesarry to convert the output between int and float. -------------- next part -------------- Index: include/speex/speex_resampler.h =================================================================== ---

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

Hi, I just built a sample application with speex resampler in linux and I tried to resample 8K sine wave tone mono to 48k using speex_resample_process_int. I am using a tool called EAQUAL for audio quality. I find the quality of Speex resampler to be decreasing when I increase the quality q of the resampler init function. Can some one give me pointers regarding this?? As per the API, if the

Hi, I've noticed that the opus-tools is using a really old version of Speex's resampler code - a version that I've seen fail in the wild first-hand under low resource circumstances. I've actually submitted patches for some issues in the Speex resampler a while ago (and IIRC they were accepted): http://lists.xiph.org/pipermail/speex-dev/2009-November/007541.html ,

Attached is a snapshot of work-in-progress of a FFT based resampler. At the moment it works in floating point only, and only basic quality inspection has been done. Some benchmarks comparing the filter-based resampler at Q3 with the FFT resampler with overlap = in_len / 2, using 20ms chunks of data. (-O3 -ffast-math, FFTW3, gcc 4.3.0 on x86_64) 16=>48: 59us vs 19us 16=>44.1: 204us vs

Hi all, I have a couple of questions about the speex resampler: 1. Can a resampler state be used to process frames belonging to several audio streams? I would guess that there is information about the audio streams stored in the state so the answer is no, but it would be good to have some confirmation. 2. What is the best way to reset a resampler state? speex_resampler_reset_mem appears

Hi, *"That's in general not very reliable. You can get PEAQ to say all sorts of silly things." Can you provide me links for any more effective tools other than PEAQ? Which is more reliable for Speex resampler? * *"strongly suspect that it's just not compensating for the delay introduced by the resampler. Because higher quality means higher delay, you'd find that PEAQ

Hello Jean-Marc, I did some experiments with the fixed-point version of the resampler in SVN. Basically it works very well, great work! But I experienced a problem when upsampling audio data with slight clipping. This seems to cause an overflow somewhere, resulting in "cracks" in the output. I'm aware that the resampler hasn't been released yet, but I wanted to mention it. :-)

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

Hello Jean-Marc, thank you for your answer! > I'll look into this. There's basically no overflow prevention for now. > I'll think about how to add that without affecting performance too much > (on CPUs that don't have hardware saturation). I'm open to suggestions :-) I'm not sure if I can really help, but I did a few more tests. Reducing the volume of the input

Hi, These patches improve the resampler set_rate function. The first 2 patches do some cleanups in the GCD calculation and avoid an overflow when calculating the new phase. This patch could probably be simplified if we allowed 64bits operations in speexdsp. The 3rd patch avoid rounding errors in the phase calculation. The problem is that the new rate is calculated with the reduced rate, which

Hi, I just started to examine the DIV32_16 function (Blackfin ASM version), and wondered why the return value of the function inside 'fixed_bfin.h' is of type 'spx_word16_t', but the local variable 'res' which is returned by this function is of type 'spx_word32_t'. Is this a trick of optimization or a bug? (Same question for PDIV32_16 and MAX16, too!) best

Hi all, I am using Speex in a VoIP application and everything is working great. Now I am trying to integrate the Resampler in order to convert data input, especially in the Wide Band mode (16 Khz). I have seen in the doc that for mono the ChannelID parameter should be 0, but How one should call the resampler in the case of PCM stereo data ? It is also stated that "It is also possible to

Hi Jean-Marc, Thanks for taking a look. I've added an example program in the patches that changes the rate frequently. You can run test-resample2 >test.raw and open in audacity or so to look at the spectrum etc. I've attached a before/after screenshot. In theory, depending on the current phase and the rate changes that are applied, the error can be audible as a pop when changing

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

Hi Jean-Marc, didn't get a reply to my last post (see below) -- do you have no idea what happens here? After some more tests, I disabled the DIV32_16 Blackfin optimizations and now get good quality on the Blackfin. But when I have overdrive on the input, things become very bad -- I'm not sure if this is really a filter stability issue like I wrote some weeks ago. I use the speex

Hi, I tried to figure out what the problem is -- but it seems to be totally different from what I expected. My status at the moment is: - computing results for "generic" and "Blackfin ASM" versions of the DIV32_16 function are the same, there is no "algorithmic bug" - Instead, there seems some sort of memory corruption: When I comment out the DIV32_16 function

Jean-Marc, Thanks for the response and the helpful info. I am trying to get the library to build without using the pseudostack define, and use either VAR_ARRAYS or ALLOC, but it seems the global stack is not defined. Where do can I define this in my example? VR -----Original Message----- From: Jean-Marc Valin [mailto:jmvalin at jmvalin.ca] Sent: Tuesday, February 20, 2018 5:40 PM To: