similar to: Leaking audio and AGC/VAD

Hi, I am working on a VOIP implementation were one of the key design goals is zero user configuration. Similar to Skype. What I've come to notice is that my soundcard (NForce4 based) leaks audio from the playback path to the recording path. This is probably not unique for my hardware at all and will happen on some percentage of all users of my software. (All of this is Win32) What happens is

agc_gain seemed to fit with the idea of what I wanted to do, it was easy to understand its units and behavior, and freezing it produced the desired results. Also I wanted to cap it, so that's done at the same place, and that definitely works. All I want to do is be able to freeze AGC adaptation and put an upper bound on the AGC (for example, 2x amplification). Both of these things seem

I've been using Speex's AGC in my VoIP program with mixed results. One of the problems I had was: At the beginning of the call, if the user hadn't spoken yet and there weren't any background sounds for the AGC to lock on to, most of the time the AGC would almost immediately jump way up to 200x gain (the default max). This is undesirable and usually results in nasty feedback. If

I think you'll have to modify Speex to get the functionality you're looking for. I've made a few simple modifications to the AGC to prevent it from 1) exceeding a specified level of amplification and 2) enable and disable adaptation, so I can freeze it at a certain level while speech is not detected. It's mostly just a matter of doing this at the end of speex_compute_agc():

Jean-Marc Valin wrote: >> I've had a small case with noise suppression and AGC. I have a fairly >> noisy environment here, and with the default parameters, noise >> suppression works fairly well while I talk. However, when I shut up, AGC >> starts slowly increasing the gain until it has amplified whatever noise >> is left to levels about equal to having no

>> Yes, after I stop speaking, the noise slowly starts climbing again, and >> if I peek at st->agc_gain, that's slowly climbing too. I think part of >> the trouble is that the noise in here isn't uniform white noise; there's >> traffic outside the window and people walking in the hallway outside my >> door. Each little event is enough to cause the AGC

> Yes, after I stop speaking, the noise slowly starts climbing again, and > if I peek at st->agc_gain, that's slowly climbing too. I think part of > the trouble is that the noise in here isn't uniform white noise; there's > traffic outside the window and people walking in the hallway outside my > door. Each little event is enough to cause the AGC to increase a little

My data is speech sampled as pcm mono 32khz. I want to use vad for regulating the agc gain mechanism. So... if voice is not detected I just reduce the agc value in order to not amplify the 'silence' otherwise we just increase the agc value if it is below the minimum. Thanks, Daniel ----- Original Message ----- De: Jean-Marc Valin <jean-marc.valin at usherbrooke.ca> Para: danflu at

Hi, I've been looking a bit at the VAD/AGC situation. I noticed a leftover variable declaration. The AGC used to use nb_loudness_adapt to count the number of frames so far, but it now uses nb_adapt which is shared by all the functions. However, it still allocates and updates nb_loudness_adapt. This patch saves 4 bytes of memory and a partial cycle of CPU power ;) -------------- next part

Steve, You're right. The AGC gain does not max out when using VAD (via the preprocessor). So instead of not transmitting when the AGC max gain is reached, I now do this instead: Start the call with VAD enabled and AGC disabled. When speech is detected, disable VAD (if 100% continuous transmission is desired for the call) and enable AGC. This seems to be working reliably so far. However,

Hi Jean Marc, I found that the AGC API and algorithm has changed in svn head trunk. Is it safe to use it? Or is it for testing purpose only? You also said that VAD in svn is broken in a previous post, is it related to the AGC change? or can we mix the old VAD algorithm with the new AGC safely? kind regards, fredo

I'm attempting to use speex preprocess for automatic gain control in an application I'm working on and could use some help. I'm using Opus as my codec. In order to keep the number of packets down, I'm using 60msec frames. I'm sampling at 48KHz as is recommended for Opus. So, the frame length is 2880 samples and the sampling rate is 48000. The source of the data is a

As I understand it, there are two separate ways to get VAD information from Speex: 1) Using the encoder. 2) Using speex_preprocess(). I present the following observations from an application developer's perspective. They may be wrong, in which case I would appreciate corrections. - The two VAD systems are implemented differently. - speex_preprocess()'s VAD provides more accurate

I'm attempting to use speex preprocess for automatic gain control in an application I'm working on and could use some help. I'm using Opus as my codec. In order to keep the number of packets down, I'm using 60msec frames. I'm sampling at 48KHz as is recommended for Opus. So, the frame length is 2880 samples and the sampling rate is 48000. The source of the data is a

Jesus, Unfortunately, FFT and magic algorithms don't work (yet?). You might want to try this if you're not satisfied with Speex VAD: http://lists.xiph.org/pipermail/speex-dev/2008-August/006860.html It won't perform any miracles, but I think it works pretty well and is easy to tweak. Tom >---- Original Message ---- >From: jmorion at toomeeting.com >To: speex-dev at

Hello, in my project im using speex 1.2rc1 and the preprocessor VAD seems to only separate complete silence from not complete silence frames. The Speex Manual, you can read "The voice activity detector (VAD) provided by the preprocessor is more advanced than the one directly provided in the codec." but if you go to the source code in preprocess.c line 995 "/* FIXME: This VAD

Hi, I've had a small case with noise suppression and AGC. I have a fairly noisy environment here, and with the default parameters, noise suppression works fairly well while I talk. However, when I shut up, AGC starts slowly increasing the gain until it has amplified whatever noise is left to levels about equal to having no filtering at all. As soon as I talk, AGC backs down fairly quick

Hi, After trying to use the prebuilt Windows 1.1.6 binaries I've founded that the DLL version doesn't export the mode variables properly, this could probably easily be fixed. Anyways, as for now I've downloaded the 1.1.6 source and built it myself and I'm linking speex statically to my application. I've written an experimental software which just packs 20 ms frames and sends

Hi all: I was analysising the speex code these days and learned much from it. Thank you so much! But I have a question about VAD: After preprocess(denoise and agc are off, vad is on),the input pcm speech was distorted seriously.The input speech include ten syllables(one~ ten), but the output speech lost one! Can you tell me why? Please see the attachment about input and output: (See attached

Jean-Marc Valin escreveu: > > Well, I can change the defaults and then people with different setups > (e.g. no noise and low input volume) will ask me to change it the other > way... In the end, the only fix is to improve the AGC adaptation > decision, including (but not only) the VAD. Could that be a run time variable? So it could be adjusted somehow by a user interface, for