Translating at remote Events (rC3/DiVOC)
This page collects the c3lingo information about how translations work at remote talks.
c3lingo Team Meeting
We're meeting ahead of the event to talk over the details and distribute talks.
Lessons Learned from the original DiVOC are noted in a feedback pad.
The main takeaways were:
Mumble works well for this
The low-latency stream was so far ahead that the translation was streamed before the original audio
A decent microphone makes a big difference in audibility
Try to set up a quiet environment: Close doors and windows, phone on silent, etc.
Speaker notes/slides are still important and should be collected
Always record your translation locally (using mumble) in case the VOC messes it up (happened often)
Use Rocket.Chat to coordinate (e.g. quick shift replacements)
Languages
Most translations are en↔de as always. For details always see the relevant Fahrplan/schedule.
More languages are possible, theoretically there is no limit on translation channels for remote events.
How the interpreting will work
The VOC produces a stream like always, but for remote events, the source of the stream is an OBS setup, with speakers and heralds feeding into it remotely. Here are more details.
The interpreters watch and listen to a special low-latency stream provided by the VOC, and transmit the translated audio to the c3lingo mumble server using a mumble client.
To switch between interpreters, since they can't see each other, shorthands like “s?” (switch soon?) and “s!” (switch now!) are used in the mumble chat. This can be extended with some clarification if there is time, like “s? at the next slide?”, “s? if you want” or “s!!”. To summarise:
Signals | As current interpreter | As standby |
s? | “I'd like you to take over after this section” | “Do you want me to take over soon?” |
s! | “Please take over now” | “I'll take over now” |
Details on Mumble setup are on a separate page.
The VOC will pick up the audio from the mumble server and feed it into the stream. This means viewers will be able to simply switch their audio track to the second language in the media.ccc.de interface.
Viewers can also listen to the translation directly on our mumble server and its web interface at mumble.c3lingo.org.
Prepare as a translator
If you'd like any help with these, before (big) digital events the tech team often organises practice sessions where they help people like you get their technical setup right and practice remote translation. To see how you can practice by yourself or with a partner, see below.
To be able to translate, make sure:
You have a decent microphone/headphones/headset combination, see
below for recommendations.
You have a computer or phone you can reliably use to connect to our mumble server at mumble.c3lingo.org. Test this as soon as possible.
Make sure you are allowed to speak in the right channel. See below on how to get authorised.
You have VLC installed and working well, so you can watch the low-latency stream.
Optionally, set up macros/hotkeys in mumble for actions like mute/unmute, push-to-talk/push-to-mute, and quick chat messages like “switch soon?” and “switch now!”. You can find details at
Mumble setup.
Immediately before translating a talk, make sure that:
Your room is as quiet as possible (close doors/windows, set phone on silent)
The low-latency stream works in VLC
You can quickly mute/unmute and send switching signals
You set mumble to record your channel as a backup
You have a bottle of water available. Stay hydrated :)
The
CheatSheet collects intro/outro texts and perhaps more useful details.
Technical Setup
To take part as an interpreter, you will need a bit of kit:
Your computer should have enough power to play an HD video in VLC without stuttering. Any laptop from the last 5 years should be able to do that.
It might be possible to do this on a phone or tablet, but you will need to watch a video stream (for the original audio and video), and simultaneously use the Mumble client, and read the text chat in Mumble. If you need to go this route, make sure to test this properly before a talk!
Your internet connection should be stable and have a downlink rate of at least 5 Mbit/s (check with a speed test).
The upload rate should be at least 0.25Mbit/s.
If push comes to shove, a 4G/5G connection might work; it might not be as stable.
A decent audio setup:
(A list of good gear can be found on the
c3voc wiki)
Best: a professional headset as used in TV and radio production. Typically €250 or more
Good: a gaming headset that has the microphone on a small arm that positions it close to your mouth. No-name models can be had from around €25.
Adequate: headphones and a separate microphone connected to your computer. The microphone should be positioned as close to your head as possible. For example, use a stack of books in front of your laptop to prop up the microphone to ge tit closer to your mouth.
Not recommended: headphones and the microphone built into your laptop.
The microphone will pick up a lot of ambient noise from your laptop, you typing, you arranging things on the table, etc.
Also, there will be distortion from room reverb, so you would need to try to pad your room with as much cloth as possible (blankets, duvets, etc.) Any masonry, wood, metal or glass surface reflects sounds, and negatively affects the sound quality, which in turn makes it harder for listeners to understand you.
To be avoided: using speakers instead of headphones.
You can also use software filters to reduce echo (from your voice reverberating on walls and other surfaces) and noise (from your mic or from the ambient sound in your room). See Noise reduction in Mumble or Noise reduction in Pulseaudio.
Try to test your setup ahead of time, so you can tune things for the best possible sound.
Practicing
To practice remote translation you can either show up for the practice sessions that are announced on the mailing list and on rocket chat, or you can practice by yourself or with a partner. For the latter, we recommend the following setup:
-
Use
Watch2Gether to create a room and enter the link to the youtube video there.
Watch2Gether allows you to sync the playback between all members so your source material stays synchronised between all translators.
Join the mumble server as described, find a room to practice in and go on translating while listening and giving feedback to each other!
You can also hit the red 'record' button to record your practice session to review it later.
Get authorised to speak
The hall channels are configured to only allow members of the “team” group to be allowed to speak. As an interpreter, you need to register your username on the mumble server and have one of the admins add you to the group. Do this:
Connect to mumble.c3lingo.org with your mumble client.
In the channel list, right-click your username and select “Register” or choose Self → Register from the menu.
Contact
the admins in the
#lingo Rocket.Chat channel or via email, giving them your mumble username.
After being added to the group, go to one of the hall channels. Check the icons on your username: there should be no mute icons when you've not muted yourself. If the notification list shows “You were suppressed”, then you don't have the rights yet. This is what it should look like: