similar to: Preprocessor denoise. Does it work?

Andy Ross a ?crit : > I'm having trouble with the preprocessor's noise reduction feature. > The basic issue is that it simply doesn't work very well. > > With my laptop (whose microphone is otherwise quite capable) I > routinely hear transient background noise, typing, and other "quiet" > sounds leaking through to the speex stream. Even worse, the AGC >

Jean-Marc Valin wrote: > The noise suppressor will only attempt to remove stationary noise, > such as thermal noise, fans, ... The AGC can indeed do strange > things in these cases, but it's been improved in svn (compared to > 1.2beta1). OK, then the problem is that I misunderstood the feature. I assumed that dynamic squelch was part of it, but it's really something more

There are many ways to implement a VAD. What you described is actually perfectly equivalent to the most trivial (and least robust) VAD algorithm. Jean-Marc Andy Ross wrote: > Ton Grandgent wrote: >> Andy Ross wrotte: >>> I wrote a trivial squelch feature* in 10 minutes that works >>> basically 100% of the time. >> Could you please explain how this differs from

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

Jean-Marc Valin wrote: > Andy Ross wrote: > > Uh, production applications almost always require squelch, no? > > Some do, some don't. In general, distinguishing between a keyboard > and a speech transient is next to impossible based only on a few ms > of speech. That is true for distinguishing it by waveform, but not by amplitude. As I mentioned, these transients are

Jean-Marc Valin wrote: >> gcc version 3.4.5 (mingw special) > > Can you try 4.0 or 4.1 just to be sure? mingw (native gcc for win32) doesn't officially support using gcc 4.0 and 4.1 yet (apparantly there have been some issues), so there are no binary packages. But if you think it helps, I can compile gcc 4.1 and give it a shot. >> Compile flags: >> DEFINES +=

Andy Ross <andy@plausible.org> wrote: > > I wrote a trivial squelch feature* in 10 minutes that works > basically 100% of the time. > > * Zero the sample data if the maximum sample in a frame is less than > 4% of saturation or 20% of the maximum sample yet seen. It's about > 8 lines of code. Could you please explain how this differs from VAD? Tom

> That is true for distinguishing it by waveform, but not by amplitude. > As I mentioned, these transients are objectively tiny. *Your* transients may be "tiny" and in any case, it doesn't help if you don't know the level you're recording at. I guess I'd be > curious as to which voice codec applications require no squelch (other > than trivial examples

Hi all, i use speex preprocessor features in this way: =================================== #define NN 160 /* 20msec di audio */ ... int tbc=0,c,d,ret; spx_int16_t TEMP_Buffer[NN]; char DLECODE; /* Inizializza il preprocessore Speex se non inizializzato */ if(Modem->speex_pp_state == NULL) { Modem->speex_pp_state = speex_preprocess_state_init(NN,AUDIO_SAMPLERATE); }

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():

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

Hello, I'm using Speex1.1.11.1 source code. I enabled dereverb by speex_preprocess_ctl(). I try different dereverb_level and dereverb_decay values and it seems to not work. I generate some test files in CoolEdit using reverb or echo and then preprocess it by speex_preprocess(preprocess, input, NULL); and output is the same as input:(. Denoising and VAD works very good. I tried dereverb with

Jean-Marc Valin wrote: > Andy Ross wrote: > > Not knowing how VAD works, I can't say for sure. > > There are many ways to implement a VAD. I meant "not knowing how speex's VAD works", of course, not VAD in general. If you would stop interpreting everything I say in the least charitable manner, this might be going more smoothly than it is. (Tom was right, by the

Hi all, as in subject, speex_preprocess inject noise in my data. Someone can help ? Here's the way that i'm using: #define NN 160 /* 20msec di audio */ #define AUDIO_SAMPLERATE 8000 spx_int16_t TEMP_Buffer[NN]; speex_pp_state = speex_preprocess_state_init(NN,AUDIO_SAMPLERATE); c = denoise; speex_preprocess_ctl(speex_pp_state, SPEEX_PREPROCESS_SET_DENOISE,&c); c = agc;

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

Andy Ross <andy@plausible.org> wrote: > > Tom Grandgent wrote: > > Andy Ross wrote: > > > I wrote a trivial squelch feature* in 10 minutes that works > > > basically 100% of the time. > > > > Could you please explain how this differs from VAD? > > Not knowing how VAD works, I can't say for sure. But enabling > VAD wasn't catching

I'm getting confused, what's the problem exactly? The AGC doesn't remove the DC or it adds one? Jean-Marc Aymeric Moizard a ?crit : > Hi Jean-Marc & all, > > I'm currently working on a wav file with very low volume. > This wav also contains a DC component. > > Using testdenoise with modification (AGC turned on with > 20000.0 value) I'm able to

Andras Kadinger wrote: > I have now found a testcase where SVN AGC starts out good, but then it > keeps increasing the gain but never decreasing it, and eventually goes > out through the roof, during the same single recording. I've been doing massive changes in the svn preprocessor, so I'm not too surprised I broke the AGC, because I haven't updated it. Did it also do that in

What is the current status of the residual echo suppressor in the preprocessor? I ask this because I have occasional problems with the AEC, in which it fails to cancel echo. I think the problems have to do with samples being dropped by the audio IO system, which is effectively impossible to detect and counteract reliably. A colleague recommended using an echo suppression technique (which he

Hi The latest SVN trunk for speex has changed the SpeexPreprocessState to an opaque structure, for jolly good software engineering reasons. However, the Analogue AGC (AAGC) feature of iaxclient (in audio_enode.c) relies on some members of this. It uses speech_prob to detect when there is enough speech to consider AAGC and then loudness2 to decide how to adjust the input mixer. We want to use