Skip to content

Latest commit

 

History

History
132 lines (109 loc) · 5.77 KB

audio-sidechain.md

File metadata and controls

132 lines (109 loc) · 5.77 KB
Raw

Audio-Sidechain

Audio Sidechain Compression

Allow audio output volumes to be automatically and temporarily reduced, based on audio input volumes (when temporarily above a threshold) -- similar to an audio side-chain compression.

Problem

In practice there are two challenges in vMix when it comes to audio:

  1. MICROPHONE DUCKING: Allow input devices (microphones, attached to the Master bus plus the "marker" bus Bus-A, but individually controlled) to be temporarily "ducked" (volume reduced) as long as output devices (callers and media, monitored on Bus-B) are active. This prevents nasty echos or even full loops.

  2. TRANSLATOR VOICE-OVER: Allow one or more people (usually remote translators, sitting on vMix Call inputs, receiving the program in Bus-A, and mixed onto the Master audio bus and additionally monitored on Bus-C) to speak over the program (usually received via NDI and mixed on the Master audio bus after being "dimmed" on Bus-B).

Solution

The solution to both challenges is to place those audio inputs onto a monitoring audio bus which should be observed (similar to the side-chain of a compressor) and the audio inputs which should be adjusted onto another audio bus (similar to the regular input of a compressor).

Usage

The recommended configurations are:

  1. MICROPHONE DUCKING:

    busMonitor           = "B"   (Notice: callers and media)
busAdjust            = "A"   (Notice: microphones)
busAdjustInputs      = true  (Notice: adjust the inputs)
busAdjustInputsExcl  = ""
busAdjustUnmutedOnly = false
volumeThreshold      = -42   (Notice: -42 dB FS)
volumeFull           = 100   (Notice: 100%)
volumeReduced        = 30    (Notice: 30% = reduce by -42 dB)
timeSlice            = 10
timeAwaitOver        = 20
timeAwaitBelow       = 150
timeFadeDown         = 50
timeFadeUp           = 50

  2. TRANSLATOR VOICE-OVER:

    busMonitor           = "C"   (Notice: translators)
busAdjust            = "B"   (Notice: original program)
busAdjustInputs      = false (Notice: adjust the bus)
busAdjustInputsExcl  = ""
busAdjustUnmutedOnly = false
volumeThreshold      = -42   (Notice: -42 dB FS)
volumeFull           = 100   (Notice: 100%)
volumeReduced        = 60    (Notice: 60% = reduce by -18 dB)
timeSlice            = 10
timeAwaitOver        = 10    (Notice: allow translators to have priority)
timeAwaitBelow       = 1500  (Notice: allow translators to breathe)
timeFadeDown         = 10    (Notice: allow translators to have priority)
timeFadeUp           = 400   (Notice: fade in program slowly)

Background

The audio volume science is a little bit hard to understand and vMix in addition also makes it even more complicated by using different scales. Here is some background on the Decibel (dB) based scales (like "volumeThreshold"), the formulas how the scales can be converted, and some examples:

Scales:
    Volume:      0 to 100    (used for UI volume bars, SetVolumeFade)
    Amplitude:   0 to 1      (used for API audio bus meter, @meterF1)
    Amplitude2:  0 to 100    (used for API input volume input, @volume)
    Decibels:    -oo to 0    (used in audio science)

Formulas:
    Amplitude    = Amplitude2 / 100
    Amplitude2   = Amplitude * 100
    Volume       = (Amplitude ^ 0.25) * 100
    Amplitude    = (Volume / 100) ^ 4
    Decibels     = 20 * Math.Log10(Amplitude)
    Amplitude    = 10 ^ (Decibels / 20)

Examples:
    ==== METER (measured) ===============      ==== CONTROL =============
    Volume (%) Amplitude  Decibel (dB FS)      Volume (%) Decibel (dB FS) *
    ---------- ---------- ---------------      ---------- ---------------
    100        1,0000        0,00              100        - 18,00
     95        0,8145     -  1,78               95        - 19,78
     90        0,6561     -  3,66               90        - 21,66
     85        0,5220     -  5,65               85        - 23,65
     80        0,4096     -  7,75               80        - 25,75
     75        0,3164     - 10,00               75        - 28,00
     70        0,2401     - 12,39               70        - 30,39
     65        0,1785     - 14,97               65        - 32,97
     60        0,1296     - 17,75               60        - 35,75
     55        0,0915     - 20,77               55        - 38,77
     50        0,0625     - 24,08               50        - 42,08
     45        0,0410     - 27,74               45        - 45,74
     40        0,0256     - 31,84               40        - 49,84
     35        0,0150     - 36,47               35        - 54,47
     30        0,0081     - 41,83               30        - 59,83
     25        0,0039     - 48,16               25        - 66,16
     20        0,0016     - 55,92               20        - 73,92
     15        0,0005     - 65,91               15        - 83,91
     10        0,0001     - 80,00               10        - 98,00
      5        0,0000     -104,08                5        -122,08
      1        0,0000     -160,00                1        -178,08
      0        0,0000     -    oo                0        -    oo

  (*) for a usual incoming voice signal of -18 dB FS

Attention: Keep in mind that 100% of a Volume Meter corresponds to 0 dB FS, while 100% of the Volume Bar corresponds to the full incoming signal (usually about the target loudness of -18 dB FS for voice). So, if you attenuate by targeting a Volume Bar of 50%, you are reducing your incoming signal BY(!) -24 dB relative, and NOT TO(!) -24 dB absolute.