Understanding The Mixer - Microphone Level 100

[Gaetano]

Hello,

the wiki page 'Understanding The Mixer' in the 'Setting Audio Levels -> Your Operating System' section reads as follow:

In Windows for example, Control Panel->Hardware and Sound->Sound, under the Recording tab, select the device in question (like your mic) and click Properties. In the Levels tab, right click the % value and choose 'decibels'. You want this to be at (or around) 0.0dB for the least amount of clipping. It will retain the setting when you switch back to 'percent'.

In my Win 10 version 21H1 there is not the choice of 'decibels'. Only value can be set is from 0 to 100 (as a matter of fact it is not even a 'percent' value).

The interface is the Focusrite Scarlett Solo.

Should I set the 'percent' to 100 in the Levels tab under the Recording tab to have a perfect match between the tricolor LED rings on the Focusrite and the Left Dot Input level (as well as the PPM) on the OBS mixer?

Interestingly enough, when I connect the audio interface the speaker volume goes to 100 so having the mic level at 100 should not alter nor distort the sound.

Currently I am using windows native audio drivers in place of Focurite drivers.

Thanks

 

[koala]

I cannot help with a specific setting, but I try to help you with understanding when clipping appears and how it can be avoided by applying attenuation (attenuation is what you do if you choose 0..100 or some negative dB value. Gain is if you apply some positive dB value).

Digital audio is carried by samples every few microseconds. One sample is the volume of the source at sampling time. One sample is stored in a integer variable, and one audio stream is an array of such variables. Each variable has a given fixed range, for example common audio has 16 bit values that go from 0 to 65535. The loudest is 65535 and silence is 0.
If you mix audio sources, their sample values are added, and the resulting sample is the mixed audio.

If by chance both sources have very high volume at the same time, so the summed sample value is above the maximum range (i. e. 65535) the value is clipped to 65535, and distortion appears.

If you always have one single source, there will never be clipping, unless you're applying gain. The common settings always attenuate (including all the Windows settings), so there cannot be clipping, so you can and should set all to 0 dB resp. 100.

However, if you're mixing 2 sources, you need to attenuate both sources by half to avoid the clipping.
An alternative to this is to not attenuate but applying a compressor or limiter filter to the mixed source, so only the loudest parts are attenuated and not clipped.

Try to avoid applying gain to any source, because this can shift the loudest parts above the allowed sample value range for louder parts and create clipping. To not require gain at some point, make sure you don't attenuate too much earlier in the chain. So try to have every setting at 0 dB or 100 and use lower values only for sources you intend to mix.

And try to mix multiple sources at the same time. If you first mix 2 sources and apply attenuation or compression, then mix the resulting mixed source with a 3rd source, you will process the already mixed sources a second time.

 

[Gaetano]

If you always have one single source, there will never be clipping, unless you're applying gain. The common settings always attenuate (including all the Windows settings), so there cannot be clipping, so you can and should set all to 0 dB resp. 100.

So the Level tabs is acting as a fader with 100 being nominal value (0 dB). It's just one single source (one Mic on the interface). I will use the gain knob on the interface to get the signal to a 'speech level' in the OBS meter.