MeRit 3.1 is an interactive multimedia instrument that was developed especially for live performances.
The goal was to create a system that has to be extremely flexible in terms of functionality and expandability, cost-effective, stable and reliable for show-use and has to guarantee to play back video files accurate, in high quality and with a negligible starting delay.
Overview
The whole system set-up consists of:
- a number of input devices (MIDI-instruments, MIDI-fied percussion instruments or sensors of any kind)
- a control / remote-computer to administer and control the system
- several video content / playback-computer, depending on the demands. Each of those keeps the whole set of video files used for a project and outputs up to four independent video streams through one single video output — either straight to a video projector or, if needed, through a video mixer first. The audio signals will probably be sent to a mixing console.
All the computers are connected to an ethernet hub and communicate via TCP/ IP.
fig. 1: exemplary system scheme for a four-screen set-up
As video playback computers, standard Apple desktop computers running OSX are required (recommended are at least double processor computers with min. 2GHz).
The videos are streamed directly from disk, therefore fast hard drives are recommended (preferably RAID 0, with two or more channels). Video files can have any Apple Quicktime® supported compression and format. Support for multichannel audio (up to 16 channels) is provided.
The software
The Software is entirely written in Max/MSP and Jitter, using Apple Quicktime® and OpenGL technology.
The basic concept of the software is rather simple: incoming trigger signals are stepping through or jumping to specific steps of preprogrammed lists, which are organised in scenes. Each step refers to a data set, holding in formation to specify the playback parameters of each video file.
All video and data content is redundant. In case of computer problems during a performance, the remote computer allows to reallocate each connected video server to the data sets of any of the screens.
The software is divided into three modules:
A) The Composer module
The Composer module can be run on any video content server and allows the user to configure and program the video content, layout and screen distribution for a whole show. Per scene 999 programmable steps are available. Each step stores in formation about:
- video file name and path
- video content of up until four independent sub-screens / layers per screen, allocable via separate browser window
- sub-screen x-/y-position and size
- MIDI control matrix (start/pause/re-start), MIDI input selection, MIDI transfer tables, MIDI-gate times etc.
- basic audio controls
The Composer module allows the user to automatically update the data sets of all connected video content servers and remote machines. The user interface of the Composer is designed to visualise the resulting appearance of a performance already during preproduction:
fig. 2: screen shot of the Composer module
B) The Player module
The Player module is a background application that is running on each video content server, allowing to in dependently play back up to four video files at the same time. Conceptually the number of video content
servers for a whole system is unlimited hence the number of simultaneous video streams is unlimited, too.
The Player module provides a minimal playback latency (< ~10 ms) and frame-precise synchrony of all running video files of all connected video content servers.
fig. 3: part of the Max-patch showing the independent video layers
C) The Remote module
The Remote module allows the user to administer and control all the connected video content servers. It gives visual feedback of any currently played video file (file names, playback position, next video file in list) and shows the file loading and network connectivity status of all servers.
fig. 4: screen shot of a part of the Remote module
The Remote software does not contain or play back video files, therefore a less powerful computer is sufficient. All input devices are connected (e.g. via a MIDI-interface) to the remote computer. Incoming trigger signals are getting translated and sent via TCP/ IP to all the connected video content servers.
Furthermore, the software gives the following controls:
- scene selection
- step number selection / manual override
- advanced MIDI-settings
- network configuration
- screen settings
- system performance settings
Outlook
Future releases of the software will include a live-input module to connect and control one ore more video input devices, buffering algorithms for video delay-lines, a time-control module which will allow a highly flexible manipulation of the playback speed for each video, soft edging for seamless blending of screens and sub-screens,
OSC-support as well as an entirely XML-based storage algorithm.
