search for: roet

...'ve broadend the tests to try all the possibilities like VBR, perceptual enhancement, complexity and quality. The results are at www.rational.co.za:81/speex.csv > ----- Original Message ----- > From: "Jean-Marc Valin" <Jean-Marc.Valin@USherbrooke.ca> > To: "Nic Roets" <speex@rational.co.za> > Cc: <> > Sent: Thursday, September 22, 2005 2:12 PM > Subject: Re: Fw: [Speex-dev] Results of Automated Batch Tests > > > Hi Nic, > > While your tool may be useful to detect regressions (I'd include > fixed-point tests too w...

I've been browsing around the speex-1.10 source tree. Firstly I see 'float' is still used in many places. Can't we change them all to the appropriate type (spx_wordxx_t or similar) defined in arch.h ? Secondly I see there is a lot of duplication regarding real-valued constants. One way would be to say: #ifndef FIXED_POINT #define CONSTANT(x) (x) #else #define CONSTANT(x)

Sounds like a great idea, except for the part about only running on linux/cygwin. What would you need in order to make it cross-platform, as libspeex is? "Nic Roets" <speex@rational.co.za> wrote: > > I'm working on a patch that will batch test most of the code in > libspeex. Here's the process (Feel free to suggest changes) : > 1. Download the representative sample with wget, if it does not exist. > Currently it's www....

Firstly, running for more channels will not break my hack. All that's needed is to call RECOPLAY_MARK with different identifiers (say nb, wb or uwb) before doing the appropriate initialization. Secondly, my attempts to do the Gaussian in fixed point went like this : Define a new constant lag_factor_gauss that is manually set equal to exp(sqr(2*M_PI*lag_factor)/-2) by whoever changes the

The results are at www.rational.co.za/speex.csv Each of the 11 quality settings is tested 3 times (narrow, wide and ultra wide band). Strangely narrow band quality 11 outperforms all wide band tests, but it can be due to my subsampling or some other inaudible effect like delaying. You can import it into Excel and sort it by SNR or other value. Divide the bits by 24 to get the bps. The

I see there are still some floating point operations left in the codec init(ialization) code. Changing that code to fixed point is not only difficult (due to the trigonometric functions etc) but may also degrade the precision. Here is an idea whereby we can easily precompute (record) all those values on a powerful processor and then use (replay) them on an embedded processor / DSP. The only

I've implemented my RECOPLAY idea and also changed a few small things that used floating point in the current version (1.1.10). In the process I discovered a few bugs (see the diff). Then I #ifdef'ed out everything that is still requires floating point including : Preprocessing, Echo cancellation, Stereo, VBR, vocoder modes, wideband and ultrawideband modes that do not use Innovation