Question / Help Audio: Automatic Level Control ALC

[hptschupp]

I use OBS in a Windows 7 machine linked to a blackmagic Web Presenter, streaming to Youtube. Everything works fine except that the audio in OBS shows some undesired ALC. The audio meters on the Web Presenter act according to the microphone input level. The audio meters on OBS work in sync but it seems to me that some AGL is present. When I have a loud signal on the Web Presenter and then a sudden low signal, the OBS meter level goes down accordingly but then "slowly pumps up" to a 20-30 db higher level even though the Web Present is still on low audio. It seems to me that something acts to boost up the low signal in which case the noise floor level is now hearable. This reminds me of a camcorder where the ALC is engaged.

Since my streaming is mainly musical performances where dynamic differences should not be compensated by a AGL, the current streaming setup is not satisfactory.

Any suggestions or thoughts are welcome.

 

[Narcogen]

https://docs.microsoft.com/en-us/windows/win32/coreaudio/disabling-the-ducking-experience

https://docs.microsoft.com/en-us/windows/win32/coreaudio/stream-attenuation

 

[uwe]

I continue this thread because the headline is perfect.

For streaming from church I was looking for automatic level control to have automatic adaption to speech and music blocks.
Compressor was doing a great job but the background noise becomes much louder.
So I combined with an expander. This is my actual setting:
Gain with 12 dB
Expander with 1,5:1 below -35 dB
Compressor with 6:1 above -30 dB and 18 dB output gain
Limiter at -0,5 dB

An image with detailed setting values is attached.

Because I am a newbie with OBS, I would appreciate feedback from experienced users what they are using to make automatic audio level control behaviour with OBS. Thank you very much in advance.

 

[konsolenritter]

Hi Uwe,

if your gain and thresholds will do, depends mainly on the entering stage of your speech and music into that channels filter chain. It depends on the crest factor between incoming peaks and rms (or loudness) too.

Some people here have experienced much trouble when trying/using something like a automatic gain control/level. It's quite useful and "hearable" for speech, but horrible for music with (expected and wanted) dynamic range within. ALC or something alike shrinks the dynamic and brings pumping effects between loud and quiet parts of the music.

Aside of the snapshots, will you provide a youtube video showing us your outcome for speech/prayer on the one and music on the other hand, please?

 

[uwe]

Hi konsolenritter,

your are absolutely right, there is always a tradeoff necessary concerning the dynamics: musicians love the original range and people with hearing aid do not like more than approx. 6 dB.

The first time I used the compressor with the default settings, many listeners complained about poor speech intelligibility, because the level difference between speech and echo became too small. With a release time higher than 300 ms it became okay for them.

Here I have extracted some examples from our streaming of today. You may access it at
https://www.magentacloud.de/share/j8st9zwauz
I hope it is accessible from your location, otherwise please complain.
0:00-0:21 Pastor speaks at the ambo
0:21-1:10 Silence
1:10-2:25 Congregation prays
2:25-2:46 Pastor sings at the altar
2:46-3:31 Congregation sings
3:31-3:57 Pastor speaks at the altar

Do you know if it is possible to record the unfiltered microphone signal as a second track? There is a possibility to setup more tracks but I don't know if they are only unmixed or also unfiltered.

As soon as I have some spare time, I will also try the proposal of the following link with longer release time values:
https://obsproject.com/forum/posts/545136/

This is our setup with
cam: Tenveo VHD102U
mic: M-Audio Uber Mic


Regards

 

[konsolenritter]

Hi Uwe,

thanks for that great example given! Wir könnten auf deutsch umschwenken, aber auf englisch haben vielleicht mehr Threadleser was davon... :-)

One additional question regarding your mic position. Got it right that all audio is coming from that single large condenser hanging? As far as i can tell all the other mics (at the altar, lectern aso.) go into a room system like a Strässer Anlage (the column speakers at the outer walls)?

So all speech and singing into those mics just go indirect via the hanging condenser into your win7 machine?
Or is there a direct injection of the room p.a. going into your audio device as well?

 

[uwe]

Hi konsolenritter,

we are using only this one stereo condenser mic. It catches the original sound and the sound of the church amplifier from seis-akustik. There is also an invisible speaker, hidden by the column, left from the cam. No other audio input to the ubuntu laptop.

Regards

 

[konsolenritter]

Aaahm, okay. Are there possibilities to improve this? I think it would be a really big step forward if you could catch the signal at the seis-akustik stuff (even if it's only cinch-based) and go directly into your device. It costs an adapter cable if doing so, and if you fear hum and buzz loops throughout the system then the insert (between seis and your audio-device) of simple passive di boxes would help great! Look at thomann for "palmer pan 01" or "palmer pan 04" (for stereo). (Can't recommend the Millenium own-brand ones...)

That way the central hanging condenser mic steps back to a role it fits really, really good: being ambient and room mic, so it's even (1) a great backup and (2) fill for more silent passages during the service as follows:

Under normal conditions you would take the output from the seis-akustik system as primary, leaving the condenser about 10..14 dB back. So it fills a little bit with room acoustics and catches the congregation during hymn and singing. In pauses of speech your streaming doesn't fall silent, even if all of the seis-mics are gated within the seis-akustik.

And (backup): If something fails on the seis-akustik or the seis-mics are gated to heavily (preventing them from catchup speech further in the room), you can easily raise the central condenser up again with the needed 12 dB or so to take over the part until the seis-akustik steps in again at the next part at ambo or altar. Your compressor/compander settings for the central condenser don't have to be such extreme settings (which result in up-raisen background noise and pumping now). Think of the coughs, chomping with the feet, snivel, crying.

For your remote listeners, even people with hearing-loss will benefit mainly if the main (spoken) parts come thru the direct cable from microphones that are 40 or 50cm apart from the pastor than (now) seven or nine meters apart. Sound will be clearer with more SNR. Even then compressor settings will benefit, because you can restrict dynamics as you wish (but not too much) without false compromises by heavily risen noise.

With a good compressor with sidechain input (ducking function) you could even automate the kick-in of the central condenser: If the seis-akustik line keeps to silent, the hanging may rise automatically.

 

[uwe]

Thank you for your detailed description of the possibilities for improvement. I have unfortunately only now come to look again in the forum (Nativity play film, Christmas, caroling). Unfortunately, the camera equipment and the seis amplifier are located on different sides of the church. But I really like the idea of controlling the room mic automatically from the seis mic via sidechain. I will keep it in mind. Maybe an occasion will soon arise to drill holes to the sacristy and run audio cables around the sanctuary.
Again, big thanks and best wishes for 2022