TV

Sound and Vision currently archives all public service TV broadcasts. A selection of commercial broadcasts are also archived, and twice a year we archive an entire week of Dutch radio and television (three public service broadcasters and six commercial broadcasters). Some regional TV is also archived.

Something to bear in mind when looking at the graphs on this page is that they show the number of programmes. However, a programme can vary from a 30s advert to the week-long recording of the 'Week of Dutch television'. This means that the proportions may be distorted - for example advertising can seem more dominant than it really is.

Depending on your browser, the graphs on this page may take a few seconds to load

Our TV collection starts with the advent of Dutch TV, although we unfortunately do not have the very first broadcasts, as programmes were not archived in the early years. Up until 2005, a selection of TV programmes was made for archival at Sound and Vision. Since 2005, every Dutch programme broadcast by a public broadcaster is automatically archived. As of 2010, subtitles are also preserved. Nowadays, a large selection of the programmes are analysed to recognise faces and voices, leading to ever richer descriptions of the programmes.

What sort of programmes are on TV? News is predominant, followed by current affairs, chat shows and reports. Genre was not considered important in early years, and while all series with more than ten programmes have had a genre applied retrospectively, other series have not, which is why a significant proportion of programmes have no series genre registered. Note that programmes can have more than one genre, so the sum of the parts adds up to more than the whole. This means that you cannot make hard comparisons of the numbers, but they do give an indication.

As news is so dominant, we will split the TV collection into news and non-news for further examination

TV news

If we look at TV news over time, something strange happens. There are large peaks in the 60's and the start of the 90's. What is going on?

Looking at the sort of programmes we have, more than half are 'international news exchange'. What is this?

Looking at a sample of the programme titles, we can see that these are news fragments from abroad, such as the new Italian cabinet being sworn in, a commemoration in Poland, and fighting in Syria:

  • Beëdiging van nieuwe kabinet Italië onder leiding van premier Enrico Letta
  • Begin schaaktweekamp Carlsen - Karjakin
  • Hevige gevechten van opstandelingen in Syrische stad Homs
  • Toespraken van paus Franciscus en president Abbas tijdens pausbezoek aan Bethlehem (Palestina)
  • Herdenking 75 jaar Opstand van Warschau in Polen
  • Ierland telt stemmen na referendum over wetgeving abortus
  • Massagraf ontdekt in Republiek Congo-Kinshasa
  • Republikeinen in senaat beschuldigigen Democraten van veroorzaken sluiting federale instellingen
  • Overzicht Dag Hammarskjöld als secretaris-generaal VN

International news exchange is a scheme whereby countries exchange news fragments. To concentrate on Dutch news, we exclude this from the further visualisations.

Now we see a much more logical progression over time, with a steep increase when automatic import of TV programmes started. As news programmes have several broadcasts per day, a slight change in schedules, such as the addition or elimination of an afternoon TV bulletin, or special breaking news programmes, can have a strong effect on the number of programmes in the archive, so the smaller variations over time do not necessarily have a particular cause.

A large proportion of news has already been digitised. Note that some programmes that fall under 'unknown/no carrier' have not been archived as a complete programme, but have selected segments that have been archived, and may have also been digitised.

News programmes often have little description. However, speech recognition technology gives us the possibility to create transcripts of the programmes, making it far easier to search them, and allowing researchers to do advanced analysis. Sound and Vision is currently carrying out speech recognition on the archive, giving priority to news items. Already two-thirds of programmes have transcripts available for researchers. The most recent content has been done first, so researchers looking further into the past will have to wait a while. 'Waiting' indicates programmes that are available in digital form and are waiting to be processed by the speech recogniser. Content still on physical carriers must first be digitally ingested, before speech recognition can occur, highlighting the importance of continued digitisation.

Word clouds can be a way of giving a quick idea of the content of the programme, either from speech transcripts or metadata. This word cloud shows the words that occur most frequently in the summaries, descriptions and item titles of a sample of 2000 news programmes

Word cloud for a sample of summaries, descriptions and item titles of 2000 news programmes (frequently occurring Dutch words removed)

The growth of new technology for metadata is clearly visible if we plot the percentage of programmes for which people, locations and organisations have been automatically recognised.

TV non-news

Excluding news, the remaining TV is a diverse mix of different genres. 'Other' includes genres as far-ranging as 'animation', 'medical information' and 'fairy tale'. Note that programmes can have more than one genre, so the sum of the parts adds up to more than the whole. This means that you cannot make hard comparisons of the numbers, but they do give an indication.

The biggest broadcaster is IP - advertising on commercial broadcasters. This may seem odd, given that the focus of Sound and Vision is on public broadcasters. The reason is that Sound and Vision once received a copy of the entire IP database. The public broadcaster adverts ('STER') are recorded in the normal way, and so in previous times their metadata had to be manually entered. This metadata is not complete for all the adverts, meaning they are underrepresented in the pie chart.

The remaining programmes are distributed over many broadcasters. Note that name changes and mergers between broadcasters are not yet accurately reflected here.

As might be expected, TV programmes are mainly from the public service networks, and shared fairly evenly over the main three channels (channel name changes have been accounted for). What may seem odd is that a large number of programmes have no channel. This is partly because advertising is not linked to a particular channel, and partly because in the early days of television it was unnecessary to register a channel - as there was only one. For the last 15 years the channel has been consistently registered in the metadata.

What sort of series are broadcast on TV? Looking at a selection of some of the most frequent, we can see some perennial favourites:

  • STUDIO SPORT
  • KRO KINDERTIJD
  • NETWERK
  • HET KLOKHUIS
  • EENVANDAAG
  • MAN BIJT HOND
  • 2 VANDAAG
  • POLITIEKE PARTIJEN
  • NOVA/DEN HAAG VANDAAG
  • DE WERELD DRAAIT DOOR
  • SPANGAS
  • NEDERLAND IN BEWEGING!
  • VAN GEWEST TOT GEWEST
  • WNL GOEDEMORGEN NEDERLAND
  • ACHTER HET NIEUWS

A large proportion of the programmes are digital, or have been digitised. Note that some programmes that fall under 'unknown/no carrier' have not been archived as a complete programme, but have selected segments that have been archived, and even digitised.

Some programmes have quite detailed descriptions, others do not. However, speech recognition technology gives us the possibility to create transcripts of the programmes, making it far easier to search them, and allowing researchers to do advanced analysis. Sound and Vision is currently carrying out speech recognition on the archive, giving priority to news items. This is visible in the low percentage of non-news programmes that have transcripts available. 'Waiting' indicates programmes that are available in digital form and are waiting to be processed by the speech recogniser. Most of these will be processed in due course, however, some programmes, such as music, would not benefit from transcripts. Content still on physical carriers must first be digitally ingested, before speech recognition can occur, highlighting the importance of continued digitisation.