ovbox:
Remote Music and Speech Communication
with Interactive Low-Delay Virtual Acoustics
using the Open Source Toolbox TASCAR
Giso Grimm and Volker Hohmann
Why?
- Allow for rehearsals during Covid19 lockdown
- All concerts cancelled
- Freelance musicians have lots of time (but no money)
- Allow for concerts during Covid19 lockdown
- Cost-efficient and climate-friendly rehearsals, now and in the future
- Research: Distributed laboratory for behavioural experiments during Covid19 lockdown
- Hearing- and hearing aid research
- Mostly elderly test participants
- Research evolved towards virtual reality for higher ecological validity
What is needed?
- Hardware:
- Soundcard
- Headphones
- Microphone
- Linux computer, e.g., Raspberry Pi
- Software:
- Sound card to network, network to sound card
(e.g., netjack, jacktrip, zita-njbridge)
- Mixing, panning, or virtual acoustics
(e.g., ardour, pd, TASCAR)
- Manual configuration of sessions by IP address,
or better a Configuration Frontend with Selective Forwarding Unit
Everything available as FLOSS on Linux!
The ovbox project
- We (ORLANDOviols) suffered from cancellations as well
- No/minimal support for freelancers: Help yourself!
- Birth of 'ORLANDOviols consort box' (=ovbox)
- March 29th: First steps (manual zita-njbridge connection to Norway)
- April 14th: First rehearsal with five musicians
- July 7th: First concert played from five locations
- Core concept: plug and play! - Simple interface, connect with others
- Share: https://github.com/gisogrimm/ovbox
- Alternative solutions: SPRAWL! (not many, lacking features, proprietary)
The Digital Stage
“The digital-stage.org is a
non-profit initiative and is developing an online conference
service that allows art, music and theater ensembles to connect,
rehearse and perform live on the Internet from different
locations.”
Close collaboration, digital stage box is based on ovbox.
Agenda of today
- Installation
- Unboxing a new Raspberry Pi
- While installing: Behind the scenes
- The configuration front-end
- Q & A
Installation
- Raspberry Pi:
dedicated image based on Raspberry Pi OS (headless),
or install from apt repository
- Ubuntu LTS:
Install from apt repository
- Planet CCRMA/Archlinux/MacOS:
Install from sources, install zita-njbridge from repo or sources
- Other systems:
Not yet tested
https://github.com/gisogrimm/ovbox/wiki/Installation
Behind the scenes
Example: Music ensemble at three locations.
Behind the scenes
Clients, 'rooms' and user register at the web interface for configuration.
Behind the scenes
Clients receive relay server address.
Behind the scenes
Clients exchange addresses, relay server acts as STUN server.
Behind the scenes
Clients exchange audio: 1 and 2 are in peer-to-peer mode, 5 not.
Behind the scenes
Lab application: Same as before, clients also exchange behavioural data.
Delay considerations
Delay considerations
Bandwidth requirements
|
4 clients:
peer-to-peer mode:
2.7 MBit/s upload
2.7 MBit/s download
n clients:
(n-1)·0.9 MBit/s
upload and download
|
Bandwidth requirements
|
4 clients:
server mode:
0.9 MBit/s upload
2.7 MBit/s download
n clients:
0.9 MBit/s upload
(n-1)·0.9 MBit/s download
|
Bandwidth requirements
|
1+3 clients:
mixed mode:
on peer-to-peer mode clients:
2.7 MBit/s upload
2.7 MBit/s download
on server mode clients:
0.9 MBit/s upload
2.7 MBit/s download
|
Spatialization
TASCAR - Toolbox for Acoustic Scene Creation and Rendering
https://tascar.org/
- Low-delay real-time interactive virtual acoustics
- Moving sources and receivers
- Image source model
- Algorithmic late reverberation
- Difusse sound field model
Core lab tool for some multi-channel labs at Carl von Ossietzky Universität Oldenburg
Examples: Gesture Lab Youtube Channel
Questions!
and
Answers?