search for: speex_preprocess_set_echo_state

Hi, when I use preprocessor with AEC, VC++ alerts me about heap corruption. I have protected speex_echo_playback, speex_echo_capture and speex_preprocess_run with mutexes, to avoid echo_state being used at the same time, but it still happens. Any help about this ussue? Thank you.

>jesus: > Now I see that it happens when I specify a framesize in > speex_echo_state_init greater than the one used in > speex_preprocess_state_init. > > Next combination works: > codec framesize = 160 > preprocessor framesize = 160 > eco framesize = 160 > > > Next combination DOESNT work (produces heap corruption): > codec framesize = 160 >

...ize, aec_tail, 2, 2); speex_echo_ctl(ec_state, SPEEX_ECHO_SET_SAMPLING_RATE, &sample_rate); preprocess_state_left = speex_preprocess_state_init(frame_size, sample_rate); preprocess_state_right = speex_preprocess_state_init(frame_size, sample_rate); speex_preprocess_ctl(preprocess_state_left , SPEEX_PREPROCESS_SET_ECHO_STATE, ec_state ); speex_preprocess_ctl(preprocess_state_right , SPEEX_PREPROCESS_SET_ECHO_STATE, ec_state ); AEC: speex_echo_cancellation(ec_state, mic, echo_ref, output); speex_preprocess_run(preprocess_state_left, output); speex_preprocess_run(preprocess_state_right, output); Is there anything wron...

> No, the 1.2beta1 preprocessor works rather well. I have a different opinion, but glad you like it anyway :-) > In fact, I very much welcomed the change that the preprocessor in > 1.2beta1 now starts out with the assumption of there being no speech, > and only activating and adapting AGC when where there is new activity > detected above that. As a result, with 1.2beta1, there is

I am trying to compile speexclient.c. I downloaded all the files to ~/speex- 1.2beta1/include When I call ./compile.sh I get the following errors, speexclient.c: In function 'main': speexclient.c:164: error: 'SPEEX_PREPROCESS_SET_ECHO_STATE' undeclared (first use in this function) speexclient.c:164: error: (Each undeclared identifier is reported only once speexclient.c:164: error: for each function it appears in.) speexclient.c:216: error: too few arguments to function 'speex_echo_capture' speexclient.c:223: warning: impli...

Aymeric Moizard a ?crit : > > I'm working on the echo canceller with the SVN, with Windows Mobile 5 > PocketPC. > > Obviously, I can't make it run correctly. What happens/doesn't happen? > My question: If I use the SPEEX_PREPROCESS_SET_ECHO_STATE, I have > changes in the sound: noise seems to be removed. Noise should be removed even without SPEEX_PREPROCESS_SET_ECHO_STATE as long as you use the preprocessor. > I want to know if I have to use the preprocessor AND that option > to make the echo canceller to work or if it's opti...

...to be used later on capture (look on capture) Echo and preprocess states are conveniently initiated using speex_echo_state_init(fpb,fpb*25) and speex_echo_ctl(es,SPEEX_ECHO_SET_SAMPLING_RATE,&sr) (and, of course speex_preprocess_state_init) Echo state is associated with preprocess state using SPEEX_PREPROCESS_SET_ECHO_STATE What is wrong?

...c_state, SPEEX_ECHO_SET_SAMPLING_RATE, &sample_rate); > > preprocess_state_left = speex_preprocess_state_init(frame_size, > sample_rate); > preprocess_state_right = speex_preprocess_state_init(frame_size, > sample_rate); > > speex_preprocess_ctl(preprocess_state_left , > SPEEX_PREPROCESS_SET_ECHO_STATE, ec_state ); > speex_preprocess_ctl(preprocess_state_right , > SPEEX_PREPROCESS_SET_ECHO_STATE, ec_state ); > > AEC: > > speex_echo_cancellation(ec_state, mic, echo_ref, output); > speex_preprocess_run(preprocess_state_left, output); > speex_preprocess_run(preprocess_state_r...

...the same as the echo data? My chain is like this: 1. get playback and capture data 2. down sample both to codec rate 3. do echo cancellation 4. do pre processing (using speex. The preprocessor also does residual echo cancelling because i called "speex_preprocess_ctl(m_pSpeexPreprocessState, SPEEX_PREPROCESS_SET_ECHO_STATE, m_EchoState)") 5. encode and send data over network. -- Regards, Niels Werensteijn

...* #define DURATION (20) * SPEEX_MODEID_NB */ _eco_state = speex_echo_state_init(_encframe_size, 10*_encframe_size); speex_echo_ctl(_eco_state, SPEEX_ECHO_SET_SAMPLING_RATE, &_sample_rate); _den_state = speex_preprocess_state_init(_encframe_size, _sample_rate); speex_preprocess_ctl(_den_state, SPEEX_PREPROCESS_SET_ECHO_STATE, _eco_state); tmp=1; speex_preprocess_ctl(_den_state, SPEEX_PREPROCESS_SET_DENOISE, &tmp); tmp=1; speex_preprocess_ctl(_den_state, SPEEX_PREPROCESS_SET_AGC, &tmp); ftmp=24000; // actually default is 8000(0,32768),here make it louder for voice is not loudy enough by default. 8000 speex_...

...th my already-working code that calls the AEC and preprocessor) to get the echo suppressor working? (Some declarations omitted, to save labor.) // m_EchoState = speex_echo_state_init( ... ); // m_PreProc = speex_preprocess_state_init( ... ); speex_preprocess_ctl( m_PreProc, SPEEX_PREPROCESS_SET_ECHO_STATE, m_EchoState ); //optional: INT32 maxEchoAttenuationDB = -40; //use negative; default is -40 speex_preprocess_ctl( m_PreProc, PREPROCESS_SET_ECHO_SUPPRESS, &maxEchoAttenuationDB ); INT32 maxNearEndEchoAttenuationDB = -15; //use negative; default is -15 speex_prepr...

As for me - speex_echo_cancellation is a better choise. Try using it in capture thread instead of those speex_echo_capture and speex_echo_playback functions. And please, describe your problem in details. Cause the fact that you "didn get echo cancellation" doesn't mean you are doing smth wrong. Regards, Anton A. Shpakovsky -----Original Message----- From: speex-dev-bounces at

...and preprocessor) to get the echo suppressor working? (Some > declarations omitted, to save labor.) > > > // m_EchoState = speex_echo_state_init( ... ); > > // m_PreProc = speex_preprocess_state_init( ... ); > > speex_preprocess_ctl( > m_PreProc, SPEEX_PREPROCESS_SET_ECHO_STATE, m_EchoState ); > > //optional: > INT32 maxEchoAttenuationDB = -40; //use negative; default is -40 > speex_preprocess_ctl( > m_PreProc, PREPROCESS_SET_ECHO_SUPPRESS, &maxEchoAttenuationDB > ); > > INT32 maxNearEndEchoAttenuationDB = -15; //use...

...try this test data on the origin speex test echo cancellation application. The echo was cancelled but the stream was damaged (some of non echo parts was cancelled to) I try this with my own test application again, but that does not work. Then I find out why. When I try to call preprocessing with SPEEX_PREPROCESS_SET_ECHO_STATE I got the same results as speex origin test application. This leads me to conclusion that echo was not cancelled by echo canceller but it was just removed as a noise. Well my question is why my echo cancellation does not work? 1. does speex echo canceller support PCM 16 audio or must I use the s...

...pprox. 30% CPU). Q.: would it work OK when the repetitive calls to speex_preprocess_run(), or only speex_echo_cancellation() are interrupted during idle? Or is another way to realize the desired behaviour as being preferred, e.g. copying the speex state as backup and restore using ctl-call SPEEX_PREPROCESS_SET_ECHO_STATE? Is this an intended feature to allow resume of idle connections? Further I'd wish to have the speex control agility be reduced when once already converged. Reason: less susceptibility to disturbing artifacts (e.g. short microphone overdrive) is suggested to happen. Any pointers to app...

Hi, We are using your speex codec in our VirtuvalClassRoom product. It's working good but we have problem using your echo canceller. Can you send me some code samples which use the echo cancelation? thanks suresh A Bring your gang together. Do your thing. Find your favourite Yahoo! group at http://in.promos.yahoo.com/groups/ -------------- next part -------------- An HTML

Hi All, I would like to communicate speech over a network compressed using speex. However, I do not want to communicate a whole Ogg-formatted file. I am interested in only the speech frames. I invoked: speexenc input-file-name - > raw-speech This, I am thinking gives the speech frames only. To play this back at the receiving end, do I need to format it into Ogg formatted file before I can

...er_ctl(enc_state, SPEEX_SET_COMPLEXITY, &quality); echo_state = speex_echo_state_init(20,320*10); speex_echo_ctl(echo_state,SPEEX_ECHO_SET_SAMPLING_RATE,&hz); preprocess_state = speex_preprocess_state_init(20 /*frame size in ms*/, hz); speex_preprocess_ctl(preprocess_state, SPEEX_PREPROCESS_SET_ECHO_STATE,echo_state); /* int value = -226; if (preprocess_state != NULL) { speex_preprocess_ctl(preprocess_state, SPEEX_PREPROCESS_SET_ECHO_SUPPRESS, &value); speex_preprocess_ctl(preprocess_state, SPEEX_PREPROCESS_SET_ECHO_SUPPRESS_ACTIVE, &value); speex_preprocess_ctl(preproc...

HA! Just when I was getting ready to drop myself from this list, along comes a ham radio question! Shoving digital voice down a 100 Hz pipe would be mighty darned tough/impossible. N0YMV has a good suggestion. Try those WSJT modes. They use narrow bandwidth and are fun to play with. There are documented instances where moon-bounce QSOs have been completed with these modes. No, it's

...er_ctl(enc_state, SPEEX_SET_COMPLEXITY, &quality); echo_state = speex_echo_state_init(20,320*10); speex_echo_ctl(echo_state,SPEEX_ECHO_SET_SAMPLING_RATE,&hz); preprocess_state = speex_preprocess_state_init(20 /*frame size in ms*/, hz); speex_preprocess_ctl(preprocess_state, SPEEX_PREPROCESS_SET_ECHO_STATE,echo_state); /* int value = -226; if (preprocess_state != NULL) { speex_preprocess_ctl(preprocess_state, SPEEX_PREPROCESS_SET_ECHO_SUPPRESS, &value); speex_preprocess_ctl(preprocess_state, SPEEX_PREPROCESS_SET_ECHO_SUPPRESS_ACTIVE, &value); speex_preprocess_ctl(preproc...