EU solidarity with Ukraine

Uppgifter om anbudsinfordringar

Rubrik:
Live Speech to Text and Machine Translation Tool for 24 Languages
Upphandlande myndighet:
European Parliament, DG Translation
Publiceringsdatum i TED:
06/08/2019
Sista anbudsdag:
02/07/2020
Status:
Stängd
Status
21/08/2019
28/08/2019
English (en)
Uppgifter om fråga
Real-time volume
In order to estimate workload, system capacity and availability, we would need some figures regarding expected volumes for speech-to-text and translation. Could the EP kindly share the anticipated number of hours per month of raw input (voice) to be transcribed and translated in real time please? Any other numbers regarding current performance figures (by human interpreters) would also help a great deal. Thank you in advance.
30/08/2019
In order to avoid any confusion with existing transcription services in the European Parliament, the European Parliament wants to clarify that the purpose of the contract is to provide in the first place "automatic live" transcription and "machine translation" services for plenary debates.
28/08/2019
The purpose of the contract is to provide in the first place transcription and translation services for plenary debates. As a general rule, the European Parliament meets for 12 plenary sittings a year in Strasbourg. Each of these sessions lasts for four days (from Monday to Thursday). Additional part-sessions are held in Brussels. The exact number of days and length of the sittings in hours varies depending on the political agenda of the Institution and the issues that are debated, therefore an exact amount of the input per month to be transcribed and translated is hard to estimate. As an example, plenary sessions were held on 54 days in 2018 and 51 days in 2019. The calendar of plenary sessions is available on the website of the European Parliament (http://www.europarl.europa.eu/plenary/en/meetings-search.html – see section "Parliament's calendar" in the right side menu). Please also see chapter 4 of the Technical Specifications which contains a rough estimation of the volume of audio recordings per language from past plenary sessions which can give an indication of the raw input to be handled per language. Original speeches, interpretation and verbatim records can be found here: http://www.europarl.europa.eu/plenary/en/debates-video.html?tabActif=tabLast#sidesForm