Case 2.0 Config: Document device settings and reasoning

docs/case2.0-configuration/camera.md

@@ -0,0 +1,5 @@
+---
+title: Camera Configuration
+---
+
+TODO: Camera settings for recording, proxies, and TFTP upload

docs/case2.0-configuration/foh-module.md

@@ -0,0 +1,139 @@
+---
+title: FOH Module
+---
+
+## Audio Mixer
+
+The Case 2.0 uses an Allen & Heath CQ-18T audio mixer because it's compact and has all important features we need, especially an automatic microphone mixer (AMM).
+It takes all four microphones, the speaker laptop's audio, and optionally a remote participant's audio and outputs a signal to the room's loudspeakers (PA), a signal for the recording, and optionally an "n-1 mix" for the remote participant(s).
+
+### Basic Signal Routing
+
+As mentioned before, we need to produce two main audio mixes: The recording mix and the room PA.
+Those two are basically the same signal, but we want to have the opportunity to add additional EQ to the PA and have additional compression on the recording mix.
+Therefore we need two separate mixes and can't just "clone" the main mix for both.
+
+The remote participant output is the same as the main/ recording mix, but without the input from the remote participant itself.
+Otherwise we would have built a feedback loop.
+
+### Inputs and Outputs
+
+We're mainly using input 1-8 for our needs:
+
+- Input 1 and 2: Wireless headsets microphones
+- Input 3 and 4: Wireless handheld microphones
+- Input 5 and 6: Laptop audio
+- Input 7 and 8: Remote participant audio
+- TODO: Bluetooth/ USB In/ Stereo In
+
+Outputs:
+
+- Main Mix/ Main L+R
+- Out 5+6: Recording Mix
+- Out 3: n-1 Mix
+- Out 1+2: Unused
+- Out 4: Unused
+
+### Routing (Sends)
+
+- Everything goes to the PA
+- The "Recording Mix" is all post-fader
+    * TODO: Any channels not going to recording, like BT?
+- The "n-1 Mix" is all post-fader
+    * but the "remote participant input" is disabled/ pulled to `-infinity`
+
+We're using post-fader sends since we're only using the different mixes to remove some channels from a mix completely or to change the EQ and dynamics of the mix output.
+But we don't want different levels of the inputs in each mix (which would also make mixing a lot more complicated).
+So we're adjusting the input channel levels on the main mix and reuse them for the recording and "n-1" mixes.
+
+Note: Unfortunately it's not possible with the CQ-18T to disable a channel on the main mix only.
+If you really need to have something in the recording, but not on the PA, you should put the PA on a different output than Main output.
+(Otherwise you would need to use some pre-fader sends which is quite error-prone in our use case.)
+
+### Input Settings and Processing
+
+#### Microphone Channels
+
+The microphone inputs use:
+
+- Some EQ to adjust for the microphone characteristic, remove handling noise (low frequencies) and remove some harsh s-sounds (high frequencies)
+- Some decent compression to even out the speaker's volume and reduce plosives
+- No gate since it just sounds weird without atmo microphones and we have the AMM for "gating out" microphones without speech
+- TODO: Auto-gain should probably not be used since it creates more issues than it solves in practice
+
+![](img/cq18t-in-mic1-eq.jpg)
+This is only an example, different microphones (headset vs. handheld) need different EQ settings.
+The green area on the left shows the low pass filter which can be set at the input/ source settings page.
+
+![](img/cq18t-in-mic1-compressor.jpg)
+The gain of the compressor it set in a way, that an input level of 0 dB lands at an output level of 0 dB (yellow line intersects the crosspoint of the two zero lines).
+
+#### Laptop Channel
+
+The laptop channel uses:
+
+- No EQ, since this is not a microphone needing adjustment
+- Some slight compression to even out different input levels without much manual adjustment
+- A gate (TODO level) to mute the channel when no signal or just some low-level noise is present (fallback when the channel isn't muted)
+- No Auto-Gain, since the maximum level is known beforehand
+
+TODO: Approxmiate settings values
+
+#### Digital Inputs
+
+The whole signal chain on the CQ18-T is designed in a way that everything should be at the 0 dB level, but there's still 12 dB headroom left up to the "full scale" digital signal.
+But digital input signals don't really need additional headroom in the input since the maximum signal is strictly specified by the digital values.
+So those inputs will produce a +12 dB signal when fed with a maximum volume input.
+
+To keep the digital inputs (like Bluetooth and USB) at a comparable level to the analog inputs, each digital input shall be configured to a -12 dB input gain/ trim. Otherwise they'll be way too loud.
+
+### Output Settings and Processing
+
+#### Main Mix
+
+The main mix does not use additional compression, but some EQ and feedback assist can be used as needed to remove excessive reverb at certain frequencies as well as feedback frequencies.
+
+#### Recording Mix
+
+The recording mix uses no additional EQ, but some more compression to level out the volume swings of natural speech even more.
+![](img/cq18t-out-recording-compressor.jpg)
+The gain of the compressor it set in a way, that an input level of 0 dB lands at an output level of 0 dB (yellow line intersects the crosspoint of the two zero lines).
+
+We also apply some limiter at +3 dB with fast attack to prohibit clipping the camera input.
+![](img/cq18t-out-recording-limiter.jpg)
+
+Note: With this setting and the camera inputs set to 100% digital input level and the dial set to 7, the camera inputs should not clip!
+Alternately you can set the dials to 10 and the digital input level to 89 (TODO: check value).
+
+### AMM
+
+**The auto mic mixer (AMM) should always be used!**  
+From our experience the AMM works really well and produces way better sound than what you can do by muting and un-muting microphones all the time -- especially in a Q&A situation.
+
+The AMM is enabled in all microphone channels (inputs 1-4) and "follow fader" shall be enabled.
+(Without "follow fader" the AMM will still use the input of microphones which are pulled down using the fader. This may lead to unexpected results.)
+
+### Buttons
+
+The canonical configuration should have the following layout:
+
+- Left side (rotaries):
+    * channel gain (or smart rotary)
+    * headphone 1 volume (or smart rotary)
+    * headphone 2 volume (or smart rotary)
+- Right side (buttons):
+    * Enable/ disable smart rotaries (changes what the rotary knobs are doing)
+    * Channel mute (for selected channel)
+    * stream mix PAFL (or AMM in old config)
+
+## Audio De-Embedder
+
+The SDI audio de-embedder is used to extract the speaker laptop's audio from the "slides" SDI signal so that it's available in the audio mixer.
+This way we can:
+
+- inject the laptop audio to the loudspeakers in the room
+- adjust the audio level at a unified place (the audio mixer)
+
+## Wireless Microphones
+
+TODO: Gain at the transmitters

docs/case2.0-configuration/index.md

@@ -0,0 +1,16 @@
+---
+title: Overview
+order: 0
+---
+
+This section shall explain the configuration of the devices in the Case 2.0 setup to:
+
+1. Document the settings when a replacement device is needed
+2. Explain why the settings were chosen
+
+Notable devices are:
+
+- [ATEM Mini](./lectern-module.md#atem-mini)
+- [Audio Mixer](./foh-module.md#audio-mixer)
+- [Audio De-Embedder](./foh-module.md#audio-de-embedder)
+- [Camera](./camera.md)

docs/case2.0-configuration/lectern-module.md

@@ -0,0 +1,46 @@
+---
+title: Lectern Module
+---
+
+The Lectern Module contains an ATEM Mini as seamless switcher, an [Infobeamer](https://info-beamer.com/) Raspberry Pi, an HDMI splitter, a Blackmagic UpDownCross converter as HDMI to SDI converter and a network switch.
+This page will go into the details how the individual devices are configured.
+
+## ATEM Mini
+
+The ATEM Mini will mainly be configured by the [pygtk-atem-switcher](https://github.com/kunsi/pygtk-atem-switcher) which is installed on the mixer laptop and configured by the [VOC config management](https://forgejo.c3voc.de/voc/cm).
+
+It will configure the following things when it's started:
+
+- Input names
+- Output resulution and framerate (note: The ATEM Mini will reset itself to 1080p60 when it looses power)
+
+Not auto-configured are:
+
+- Still images
+- Audio levels of the HDMI inputs
+- Audio levels and muting of the 3.5mm stereo-jack inputs
+- Configuring the 3.5mm jacks to "line" level (instead of microphone)
+
+Manual default configuration:
+
+- Turn audio on input 1 (speaker's laptop) on and reset level ("reset" button)
+- Turn "MIC 1" (and "MIC 2") 3.5mm jacks off
+- Set analog level to 0 dB gain and 0 dB level (only possible using software control UI)
+
+![](img/lectern-audio-comp.jpeg)
+
+## UpDownCross
+
+The VOC records in 1080x1920px at 25fps, so all input signals to the recording shall also be at this resolution and frame rate.
+However, some projectors or displays don't like 25fps, so we set the output of the ATEM Mini to 50fps and use the Blackmagic UpDownCross for the framerate conversion as well as converting the signal from HDMI to SDI.
+
+The DIP switches should be set to 1080p25.
+(TODO: Image of the UDC DIP switches)
+
+## Infobeamer
+
+The Raspberry Pi is configured with the [Infobeamer](https://info-beamer.com/) SD card image and registered at the VOC account.
+Additional settings are:
+
+- Static IP (`10.73.x.4/16`, according to case number)
+- Router at `10.73.0.254`