Mjf database

A database for the Montreux Jazz Archive

STAR 2013 - 18.01.2013

Jean Rossier

The MJF archive

• Concert recordings (circa 5000 hours of audio and video)

• BUT also: metadata• set lists (songs)• musicians on stage• instruments played• issues during a concert• ...

• Photos• Rights (legal)

2

The richness of the archive

• Digitalizing the concerts is the first step, but it is not sufficient

• Metadata is necessary to exploit the richness of the archive

• Any active and deep exploration of the archive relies on metadata

3

The richness of the archive - schema

4

instruments

musicians songs

concert

rightholders

source tapes

sourcequality

audio files

digital quality LTO

audio problems

video problems

video files

digitalization

Multiple partners

• Multiple partners are involved in the Montreux Jazz digital project

• Each partner provides some part of the metadata

• At beginning of the project, information was spread around

5

Spread metadata

6

instruments

musicians songs

concert

rightholders

source tapes

sourcequality

audio files

digital quality LTO

audio problems

video problems

video files

digitalization

Vectracom

Montreux Sounds EPFL

MJF

Single reference point

• We started to build a central database that would be the single reference point

• This database is hosted in EPFL and managed by the MMC team

• All existing information has been imported and consolidated in the central database

7

The MJF archive as a model

8

concert

songs

...

usage rights

...

recorded medias

audio files video files

quality

LTO

...

AV problems

...

instruments

...

rightholders

...

quality control

...

musicians

...

Multiple partners - single database

• Each partner of the project provides some part of the metadata

• Partners insert metadata independently• All metadata is aggregated in the

central database

9

Interacting with the database

• Constraints:• Partners access the database remotely• Inserted data must be constrained (to

avoid meaningless data)• Accesses must be monitored and

controlled

• We built a custom web application to access the database

10

Web application details

• Web application developed in Scala, using the Lift web framework

• Scala is a language created at EPFL• MySQL database

11

Application layers

12

MJF  Research  portal

other  apps

IT  infrastructure

MJF  Database

MJF  DB  API  (REST)

MJF  DigitalizaJon MJF  Rights

MySQL

ScalaWeb services

Partners interfaces

• Specific interfaces for each partner:• Insertion• Update• Search

• Partners only have access to some parts of the data

13

MJF  Research  portal

other  apps

MJF  DigitalizaJon MJF  Rights

Digitalization data

14

Quality control data

15

Rights data (legal)

16

Video previews

• Every digitalized video can be previewed via the web application

• A low quality format of each video is encoded

• The preview is played using Sublime Video, an HTML5 video player developed at EPFL

17

Demo: video preview

Photos

• We are currently working on integrating the photos into the database

• ~ 100’000 photos to archive• As other data, the photos are linked to

existing items in the database (e.g.: concerts, artists, rights)

18

Search interface

• Rich multi-criteria search interface

19

Demo: search tool

Live from the festival

• From 2012, the application is used during the festival

• The MJF team inserts data directly in the database

20

THANK YOU