Hi, there. I have an instalation where one operator is talking and her voice is out by a LCD display (with speakers). Above LCD, there's a mic (with a webcam) to allow the operator to hear the people who is in front the LCD, and talk to them

The problem is that the operator hears her voice.

I want to make an echo supressor. I thought to make something like this:

[adc~]
| \
| [delwrite~ echo-supressor-A 1000]
|
[delwrite~ echo-supressor-B 1000]

[1\ <-- here, I have a Hslider too, to synch, because there's some ms delay.
|
[vd~ echo_supressor-A]
|
[*~ -1]
|
[+~ 1] [vd~ echo_supressor-B 1]
| \ / |
| X |
| / [s~ $0-out-signal]
|
[dac~]

This is just a test at home, without the computer of the instalation (now, it is in an expo).

Doing this test, I can supress audio voice... it seems to be ok, but using:

[r~ $0-out-signal]
|
[env~]
|
[- 100]
|
[vu]

it appear as full volume... but there's no sound.

Why this is like this? I suppose that the [vu] should show no sound... but it's at full.

I think you need to get rid of the [+~ 1]. An amplitude of 1 is the equivalent of 0dBfs, but a constant of 1 isn't something you can hear. It's basically a DC of 1, but [env~] assumes DC is at 0.

Yes, I got rid of the [+~ 1] and, it seems to be ok. The sound dissappear and the [vu] shows no signal. This happens in this test, in this machine. I should-will test in the installation.

The main purpose of using [*~ -1] and [+~1] was because I try to make an antiphase (http://en.wikipedia.org/wiki/Active_noise_control) of the same sound. But it seems like I don't understand well some audio signal theory.

I thought that, if the sound goes from 0 to 1, multiplying to -1 and adding 1, I will have the inverted phase... but there is my mistake. Also, I have checked what DC offset is, to understand what you told me.

I'm always confused about dB, dBfs, DC, audio signal values (in Pd (from 0 to 1?)), and so...

Two questions:

Where should I learn about these topics? If spanish, better.
Getting rid of [+~ 1], the patch should make the anti-phase and should work ok?

Thank you.

To invert the phase, you just have to [*~ -1] !

@MarioMey said:

I thought that, if the sound goes from 0 to 1, multiplying to -1 and adding 1, I will have the inverted phase... but there is my mistake.

It's actually supposed to be -1 to 1, which is why [*~ -1] is all you need.

I'm always confused about dB, dBfs, DC, audio signal values (in Pd (from 0 to 1?)), and so...

dB is just a relative measure of the signal level, meaning any dB value you have only makes sense compared to another level. Saying something is -10 dB is meaningless on its own, but saying it is -10 dB relative to a certain level isn't. dBfs means "dB level at full scale", and full scale refers to the highest level allowable without clipping the dac (i.e., -1 and 1). It's pretty much standard for digital audio to go with 0 dBfs as the reference level, which is what [vu~] goes with. In that case, -10 dB then means something because it's just -10 dB lower than the highest allowable level. (This is also why I find [env~] kind of annoying, because it uses 100 dBfs instead of 0 dBfs).

DC refers to the level that 0 Hz is at, or what level the audio is oscillating around. Ideally, it should be 0 so you can make full use of the -1 to 1 range. So with the [+~ 1] in your patch, you're actually just adding a DC offset of 1, making the range of the audio now 0 to 2. [env~] assumes DC is at 0, so if it gets a flat value of 1, it's going to spit out 100 dB.

Where should I learn about these topics? If spanish, better.

pd-tutorial.com has a Spanish translation. Section 3 goes over this stuff a little.

Maelstorm, now, I understand so much more. Thank you very much.

Also, I read in Wikipedia about dB, dBfs and so (unfortunally, from dB Wikipedia spanish page, there's no link to DBFS, that does exist in the site. I don't know how to modify it).

Looking at [vu], I have a doubt. How many dBfs is the absolut silence? -99dBfs? I read that the middle is -6dBfs, but I also read that it is logaritmic. What about -12dBfs?

Now, I'm checking pd-tutorial.com.

Compared to 0dBfs, absolute silence is often referred to as -&#8734; dB. Since dB is a logarithmic representation of raw amplitude, there really isn't a hard level for absolute silence, it just gets smaller and smaller the closer you get to silence. So -&#8734; dB is used instead.

Keep in mind, though, that [env~] (as well as [rmstodb]) not only uses 100dB as full scale instead of 0, it also considers 0dB to be silence. This basically means that anything that is 100dB below full scale will just be considered silence by [env~], even if there actually is something there. That's pretty quiet, though, so most times it's not an issue.

A drop of 6dB is about the equivalent of cutting the raw amplitude in half. So a full value of 1 gives 0dB. Cut that in half to .5, and you get about -6dB. Cut it again to .25, and you get -12dB.

Thanks, Maelstorm.

Bringing back to life an old thread... I tried to make this patch work LIVE... and it is not as thought.

To make this patch work, I should find the perfect delay to suppress the audio. Using 44100, the delay time can vary in 0.0226ms. So... it's almost impossible to get that. I only acchieve a flanger, nothing else.

In the first Patch-Circle done in Argentina last Saturday, I talked with Pablo E. Riera about this and he told me about the adaptive external and the echo cancellation example... but there is no example like that (only intereference cancellation).

I will explain again: the operator has a microphone and it sounds in a LCD (far away). Above the LCD I have a Webcam with mic. This mic takes what the audience says and send to the operator headphones. The operator hears her own voice (with a delay). I want to supress this "return". It doesn't matter the delay, if there were no delay, I don't want the operator to hear her voice.

I would need something that learn from the main microphone and get rid of what the webcam mic takes.

I hope being clear in my explanation. Thanks in advace.