Subtitling to support
an inclusive society
More than 51 million Europeans suffer from partial or complete loss of hearing. The European Parliament has initiated legislation that will force all public broadcasters to insert subtitles into all their programs, step by step.
The initial plan back in 2003 was to have all public broadcasters’ content subtitled by 2020. We are not even close, but great progress is being made due to new available technologies.
It makes commercial
sense to offer subtitles
About 20% of all viewers use same language closed captioning, and 80% of these viewers are not deaf or hard of hearing. As this percentage is expected to rise, it is not only an obligation for public broadcasters to increase the percentage of programs being ubtitled.
It also makes commercial sense for commercial broadcasters and on-demand providers.
Subtitling is a slow and
expensive manual process
The traditional workflow for same language closed captioning is a laborious process. It consists of several steps. First step is for transcribers to view and listen to content and to translate what they hear into written form. The next step is for a trained language expert to correct spelling mistakes and make improvements. The last step is spotting, the process of timecoding the sentences and
making them fit, following several spotting rules. In Western Europe one processed subtitling minute costs about 8 euro.
Making the process
60 to 80 percent more
efficient and cheaper
This way of working is a thing of the past. Our Subtitling Platform has made the process, depending where you are, 60% to 80% more efficient and more affordable, by using smart Natural Language Processing / Artificial Intelligence technology. Using this technology eliminates two out of the three steps of the traditional subtitling process.
Eliminating the manual
Our speech to text AI technology “listens” to your content and automatically transcribes into written word what it hears. The model has been trained to understands 31 languages, and transcribes whole tv shows in seconds to a maximum of a few minutes.
The technology is not perfect. The Platform makes mistakes, especially with names or when there is a lot of background noise.
There is an error margin of maximum 5%, and you will still need your language expert to check what has been written and make some corrections. The platform itself will point out to your experts what is likely wrong, saving you time to make the necessary improvements. The 5% error margin can be further reduced by “teaching” the platform names of public figures.
Eliminating the manual
The subtitling platform will also automate the process of spotting. It uses your specific spotting rules to do so: how long should sentences be, may the next sentence appear during a shot cut and any other spotting rule you may have. The automated spotting process is extremely accurate; it detects exactly when an actor starts a new sentence, and closed captioning uses this timecode.
Advantages of ai generated
Eliminating transcription and spotting, but keeping in place human intervention is exactly what automation should do. Slow, repetitive and time-consuming processes are removed in favor of more quality control. Using our platform will save you 60% to 80% of the costs per processed minute. Existing staff can be easily retrained from being a transcriber or a spotter to doing quality control. Enabling you to use closed captioning for an even higher percentage of your content, at substantially lower costs.
AI for subtitling
of foreign languages
Our subtitling platform is not limited to same language. Once foreign language content is transcribed and checked, we can use AI driven translation to translate foreign subtitles into your own language. Although with subtitling there is also an error margin, the results are still remarkably good and will definitely save you even more budget compared to translation services.