瀏覽代碼

X32 stuff

HON95 1 年之前
父節點
當前提交
e31932daea
共有 4 個文件被更改,包括 138 次插入21 次删除
  1. 8 4
      audio/audio-basics.md
  2. 119 1
      audio/behringer-x32.md
  3. 5 8
      networking/juniper-junos.md
  4. 6 8
      networking/vyos.md

+ 8 - 4
audio/audio-basics.md

@@ -94,12 +94,16 @@ breadcrumbs:
     - Type C has flatter response than type A and is often used for audio calibration.
 - Feedback:
     - Happens when sound is fed from speakers back into a microphone (accidentally), at a high enough "loop gain" that the feedback noise level quickly escalates to annoying/damaging levels.
-    - Generally only happens at certain resonating frequencies, depending on the venue/room.
+    - Happens at certain resonating frequencies, depending on the micropone characteristics and venue acoustics.
+    - Keep in mind that room acoustics changes when the audience fills the room.
+    - Reverberations from e.g. untreated walls will contribute to causing feedback.
     - Preventing feedback:
-        - Avoid placing microphones in front of speakers.
-        - Use appropriate microphones, e.g. dynamic microphones pointing away from any (loud)speakers.
-        - Use an equalizer to reduce the level for feedback-inducing frequencies. To find the frequencies, test the setup at loud levels to try to induce it, then measure which frequency it's happening at.
+        - Avoid placing microphones in front of speakers. Place speakers in front of the stage instead of behind it. For stage mics in front of monitors, use cardoid mics pointed directly away from the monitor speakers.
+        - Use appropriate microphones. Use directional/cardioid mics (i.e. dynamic mics) whenever possible.
+        - Use gating and low cut on vocals to avoid contribution from silent/inactive microphones.
+        - Use an parametric equalizer to cut feedback-inducing frequencies. To find the frequencies, test the setup at loud levels to try to induce it, then measure which frequency it's happening at (called "ringing out the microhpone/feedback").
         - Don't use "feedback destroyers", they're crap.
+        - As a more permanent approach, find reflecting surfaces (walls, ceiling beams etc.) and cover them with acoustic treatment.
 - White and pink noise:
     - White noise has equal energy per frequency. The frequency-SPL curve is flat.
     - Pink noise has equal energy per octave. Since each octave has double the frequency of the previous (`log2(freq) = octave`), the frequency-SPL curve is decreasing. This makes pink noise appropriate for equalization for human hearing.

+ 119 - 1
audio/behringer-x32.md

@@ -5,6 +5,124 @@ breadcrumbs:
 ---
 {% include header.md %}
 
-**TODO**
+## Routing
+
+- "Channel" types:
+    - Channels (mono): Inputs, fed from physical inputs, stageboxes (AES50), sound cards etc.
+    - Mix bus (mono): General mix buses, fed by inputs or effects.
+    - Effect groups (stereo): Fed by **TODO**
+    - Main LR (stereo): Main stereo bus, fed by **TODO**
+    - Main M/C (mono): Main mono/center bus, fed by **TODO**
+    - Matrices (mono): Fed by buses (incl. main LR+M), can't be fed from channels.
+    - DCA groups: Doesn't carry audio but instead remote controls the level of a set of inputs.
+- Tap points:
+    - Tap points are used for both mixbuses (typically pre-fader) and outputs (typically post-fader).
+    - The "+M" tap variant mutes the tap too when the channel/bus is muted.
+    - In/LC (1): After preamp gain and low cut.
+    - Pre-EQ (2): Before EQ.
+    - Post-EQ (3): After EQ, before compressor if using post-EQ compressor.
+    - Pre-fader (4): Before fader.
+    - Post-fader (5): After fader.
+- The sampling rate (44.1k or 48k) must be set equally for everything connected, including AES50, USB and USB recordings.
+- Use the "L/R + Mono" panning mode, where you manually assign sources to LR and/or mono outputs and pan only applies to LR. If you have a center speaker/cluster, this allows you to assign vocals/talks to only the center to avoid combing filter effects and such which might be especially noticable for certain audio types. The "LCR" mode sets left, center, right side-by-side panning-wise, so that the source is only sent to the center when the pan knob is centered. LCR mode may appropriate for more "immersive" sound images.
+- Use "M/C depends on Main L/R" if you want the mono level to follow (depend on) the LR level.
+- For setting input channel sources individually (not blocks of 8), map the individual sources to user inputs and then select user input blocks for the channels.
+- "Sends on fader" works in both directions, you can select either an input channel or an output bus. The "bus sends" buttons/knobs are redundant (the compact console doesn't even have them).
+- Use the "mono bus" and "stereo bus" buttons and associate knobs to route channels to the main LR (stereo) and/or main M (mono/mid/center).
+- For the PA speakers, consider putting the full-range speakers and the subwoofers on separate matrices fed from the main bus. This has a few advantages:
+    - You can keep the main clean and put house EQ on the matrices instead.
+    - You to keep the main bus at zero and instead set the house volume at the matrices.
+    - You can use the matrices as the system processor, to handle speaker crossover EQ and stuff. In this case, you probably want a stereo pair for the full-range speakers and a mono or stereo pair for subs.
+- Use output delays if speakers are different distances away from the audience (if you're willing to do the math).
+- When setting up busses for e.g. monitors, it's a good idea to solo the bus and listen to it using your monitor headphones.
+- DCA groups:
+    - DCA groups are useful for controlling the volume for or muting a group of inputs (e.g. all vocals or all drum mics).
+    - To add channels to the DCA group, hold the DCA group's select button and press select on the channels to add or romove.
+
+### AES50
+
+- Supports 48 inputs and 48 outputs for each interconnect (A and B).
+- Typically used for stage boxes (e.g. an SD8, S32 or X32 Rack).
+- Use shielded EtherCON to avoid blowing out the AES50 ports in case of static discharge.
+- All interconnected devices must use the same clock to avoid audio glitches. To set the X32 to use the clock from a AES50 source (e.g. another X32), configure it in the settings.
+- Configuring an X32 as a pure stagebox:
+    - **TODO** details, I can't remember exactly. Patch AES50 directly to XLR out, patch inputs as directly as possible to AES50. Preamp settings must still be configured on the X32 stagebox.
+
+## Gating
+
+- Enable gating on microphones to avoid picking up nois and contributing to feedback.
+- Use a gain reduction (low "range") of e.g. -9dB to not silence it completely, which can sound weird.
+
+## Equalization
+
+- Use a graphical EQ to tune the main speakers to the venue. Measurement microphones and frequency sweeps may help give some indication on what to tweak, but relying solely on that may generally yield a pretty bad-sonding EQ. Test with the intended sound/music instead for fine tuning.
+- For inputs like microphones and instruments, add a low shelf (low cut) at an appropriate frequency to avoid noise from frequencies the source is not supposed to generate. For vocals, add a low cut/shelf somewhere in the 60-80Hz range.
+- For full-range speaker outputs, you may optionally add a low shelf and a high shelf if you know which frequencies the speakers are not able to play. This may help protect some old/bad speakers, but is typically not needed.
+- If feeding different busses/matrices to the full-range speakers and the subwoofers, consider adding some shelves for better controlling the crossover frequency range. Or don't if you want to at some point turn the subwoofers completely off (for some reason).
+- For crossover, you can use the LR24 mode which correctly crosses over if you use the same frequency at each side.
+- Some modes like LR24 eats up two EQ bands/slots.
+
+### Ringing out Feedback
+
+- See the notes in [Audio Basics](/audio/audio-basics/), mostly only X32-specific stuff is noted here.
+- Use gating and low cut on vocals to avoid contribution from silent/inactive microphones.
+- Use a parametrics EQ for the input channel that's causing feedback.
+- Use the built-in RTA inside the input channel EQ to see what's happening while tuning the EQ.
+- While talking/playing into the mic, very slowly turn up the channel level until you notice feedback. Be prepared to quickly turn down the level. Watch the RTA and cut the feedback frequency a bit. Repeat.
+- Avoid EQ-ing too much, it will obviously affect how the mic sounds. At most 3 frequencies, as narrow and shallow as possible.
+
+## Effects
+
+- The X32 has 8 stereo effect banks, where all may be used as inserts but only the first 4 may be used as FX returns that can be fed back into mix busses.
+- The first 4 are by default sourced as mono from busses 13–16, with stereo returns (8 stereo-linked busses).
+- FX 1 is reverb by default.
+- Mute groups can come in handy to mute effects between songs etc.
+
+## Talkback
+
+- The X32 supports two talkbacks (A and B).
+- Use the build-in microphone below the display or plug in an external one into the dedicated XLR input.
+
+## User Interface
+
+- Assign controls:
+    - The "assign" knobs (4) and buttons (8) allow you to assign functions or links to them, e.g. to modify FX stuff, mute channels or go to a certain view.
+    - Three sets can be configured (A, B and C).
+    - The X32 Compact doesn't have the strip displays, four knobs or the three buttons to select sets. Switching sets can be done via the view display instead.
+
+## Presets
+
+- **TODO**
+
+## Remote Control
+
+- For Windows/MacOS/Linux, use X32Edit (Arch Linux: `yay/x32edit`).
+- Use the "Mixing Station" app on Android and iOS for remote control. It's better than the official app.
+- Behringer X-Touch can be used as a remote control surface, connected over the local network. X-Touch remote must be enabled in the X32 settings.
+
+## Recording and Playback
+
+### X-USB Module
+
+- Supports 32 24-bit channels of input and output.
+- Allows for virtual sound checks using the module as the channel sources if the physical/original sound checks were recorded.
+
+### X-Live SD-Card Module
+
+- Supports 32-channel recording.
+- Has a USB port for connecting to a PC, like the default USB modile.
+
+### USB Drive
+
+- Record and play stereo audio using a USB drive.
+- The drive should be USB 2, "fast" (class 10) and must be formatted as FAT32. Some large USB 3 drives work too. Kingston and Sandisk are rumored to be OK.
+- Don't remove the USB drive while the access light is lit.
+- If you have a lot of headroom in the input mix, increase the recording trim until it reaches an appropriate level.
+
+## Tips and Tricks
+
+- On the Compact version, select two input layers at the same time to show the first on the left channel strips and the second on the right channel strips.
+- Enable color for the RTA.
+- Bring a 5GHz wireless router to set up your own out-of-band network for remote control. Consider hooking the WAN port up to the local network for internet access, if required. Also consider connecting the tablet using an RJ45 Ethernet adapter to avoid relying on WLAN.
 
 {% include footer.md %}

+ 5 - 8
networking/juniper-junos.md

@@ -440,7 +440,7 @@ Wait for the "The operating system has halted." text before pulling the power, s
 
 Note: USB3 drives may not work properly. Use USB2 drives.
 
-1. Make sure the drive is formatted as FAT32 (MS-DOS) (or something else supported).
+1. Make sure the drive is MBR-partitioned and that the partition is formatted as FAT32 (`mkfs.fat -F32`).
 1. To make it easier to find the device path in Junos, don't insert the USB drive just yet.
 1. Show current storage devices: `ls -l /dev/da*`
 1. Insert the drive. It should print a few lines to the console.
@@ -452,16 +452,13 @@ Note: USB3 drives may not work properly. Use USB2 drives.
 
 ### Upgrade Junos
 
-#### Preparations
-
-1. (Info) For virtualized boxes like EX4600 and QFX5100, skip the `request system snapshot` parts as these boxes are built differently wrt. Junos.
-1. Cleanup old files: `request system storage cleanup`
-1. Make sure the alternate partition contains a working copy of the current version: See [Validate the Partitions](#validate-the-partitions).
-
 #### Normal Method
 
 This should work in most cases and is the most streamlined version, but may not work for major version hops and stuff.
 
+1. (Info) For virtualized boxes like EX4600 and QFX5100, skip the `request system snapshot` parts as these boxes are built differently wrt. Junos.
+1. Cleanup old files: `request system storage cleanup`
+1. Make sure the alternate partition contains a working copy of the current version: See [Validate the Partitions](#validate-the-partitions).
 1. If downloading from a remote location:
     1. Get the file: `file copy <remote-url> /var/tmp/`
         - If it says it ran out of space, add `staging-directory /var/tmp`. By defaults it's buffered on the root partition, which may be tiny.
@@ -475,7 +472,7 @@ This should work in most cases and is the most streamlined version, but may not
     1. Copy the file to internal storage: `cp /var/tmp/usb0/jinstall* /var/tmp/`
     1. Unmount and remove the USB drive: `umount /var/tmp/usb0 && rmdir /var/tmp/usb0`
     1. Enter operational CLI again: `exit` (or `cli`)
-1. Prepare upgrade: `request system software add /var/tmp/<file> no-copy unlink reboot [force-host]` (supports auto-complete)
+1. Prepare upgrade: `request system software add /var/tmp/<file> no-copy unlink reboot [no-validate] [force-host]` (supports auto-complete)
     - `no-copy` prevents copying the file first (in this case it's pointless).
     - `unlink` removes the file afterwards.
     - `reboot` reboots the device, so the upgrade can begin when booting.

+ 6 - 8
networking/vyos.md

@@ -19,7 +19,9 @@ A Debian-based router OS, forked from Vyatta. Junos-like CLI.
 
 See [Installation (VyOS)](https://docs.vyos.io/en/latest/install.html).
 
-1. (Recommended) Disable Intel Hyper-Threading.
+1. For physical instances, consider disabling Intel Hyper-Threading (it does little for memory-intensive applications like packet routing).
+1. If running as VM inside a hypervisor like Proxmox, disable memory ballooning/memory sharing.
+    - VyOS does not use swapping so accidental overprovisioning that could starve the VyOS VM could cause errors.
 1. Download the latest rolling release (free) or LTS release (paid) ISO.
 1. Burn and boot from it (it's a live image).
 1. Log in using user `vyos` and password `vyos`.
@@ -109,13 +111,9 @@ An example of a full configuration. Except intuitive stuff I forgot to mention.
     1. (Note) Newlines must be escaped with `\n`.
     1. Set pre-login banner: `set system login banner pre-login ""` (disable)
     1. Set post-login banner: `set system login banner post-login ""`
-1. (Optional) Tuning (bare metal):
-    - **TODO** This can be done in the interface ethernet configs instead. Except RX/TX buffer sizes? Does VyOS automatically maximize now?
-    - (Old) See the Linux router notes.
-    - (Old) Enable GRO (example): `ethtool -K <if> gro on`
-    - (Old) Increase RX/TX buffer sizes (example): `ethtool -G <if> tx 4096 rx 4096`
-    - (Old) Enable scatter/gather aka vectored I/O (example): `ethtool -K <if> sg on`
-    - (Old) Make any ethtool (e.g.) commands permanent by adding them to `/config/scripts/vyos-postconfig-bootup.script`.
+1. Hardware tuning (bare metal):
+    - (Note) VyOS automatically sets large RX/TX buffers (always?) and provides a config interface for other options now, so no need to add `ethtool` stuff to `vyos-postconfig-bootup.script` anymore.
+    - (Note) For background info
 1. Commit and save: `commit` and `save`.
 
 ## General Configuration