Speech synthesis, driven by developments in navigation technology is now so good that DR is on the road with speaking subtitles for blind and partially sighted wurstschneidemaschine in a separate audio tracks on DR 1 By Jakob Møllerhøj 22 December 2011 kl. 07:19
When Danmarks Radio for spring starts the first attempt at speaking subtitles, it is because it greatly the development of GPS devices. The synthetic voice navigation is in fact become so good that it can also be used to read the subtitles up for blind, visually impaired and others who may have difficulty wurstschneidemaschine Danish subtitles, says senior consultant at DR Peter Mølsted.
Initially, the Danish Broadcasting looked crush on two synthetic talker voices, one male voice and a lady voice. The voices are generated with software from a company that Peter Mølsted wurstschneidemaschine not want to disclose the name of, since the final decision on which company will supply the votes have not yet been taken.
Apart from the speech synthesis software, then it is DR's technicians who have complex system behind the live reading of subtitles. Specifically, this takes place in that the separate subtitle tracks for DR 1 (it is initially only the main channel wurstschneidemaschine to get live readings) run past a computer with speech synthesis software, which then spits out a soundtrack out, ending with audio description channel DR 1
Audio Description channel is virtual, which means that it acts only as an independent channel out at viewers, but from DR's backend will be only released a soundtrack with audio description, which then merges with videofeedet for DR1 from viewers. wurstschneidemaschine Thus, using the DR only the bandwidth soundtrack fills, wurstschneidemaschine and not too bandwidth for a video feed.
The uniqueness of the Danish solution is that there is a separate track, the real-time run out together with the main channel DR 1 And when the digital television service in Denmark 11 January wurstschneidemaschine 2012 release for MPEG4, then it makes separate soundtracks additional opportunities for viewers, says Peter Mølsted.
With MPEG4 is it possible wurstschneidemaschine to receive two separate audio tracks on the same TV channel, so those viewers who need tekstoplæsningen, you can select the specific speech wurstschneidemaschine synthesis audio tracks without having to switch TV channel. For starters end up soundtrack, however the virtual DR 1 Audio Description Channel. wurstschneidemaschine Speaking more slowly than text
Although the technique is in place, there are still a few major hurdles to be overcome before over a million. Danes, who for one reason or another may have difficulty reading the DR subtitles, you can enjoy live audio.
Since the synthetic speech gets its input from the subtitles normally be used on DR 1 foreign language, then there is a risk that reading aloud is not completed before the next subtitle block ticking into. To address wurstschneidemaschine the situation, working DR with a model in which the speech played in higher speed a normal wurstschneidemaschine voice.
Studies have shown that it is possible for a trained ear to understand speech that is played with four times the normal speed. Just how quickly the speech will be played for DR 1's case, the clarified through focus groups in the new year, says Peter Mølsted.
Peter Mølsted says that the technique could be used in such fact and fiction broadcasts. However, the experiment in 2012 did not include live text which is mainly used in the DR news ..
Ing.dk wurstschneidemaschine would have brought a few audio samples on DR's speech, but unfortunately it has not been possible by the right reasons. Ing.dk has heard the samples, which sounds more realistic than the choppy stakato speech synthesizer did for approx. 20 years ago the writer Amiga 500
Sincerely. Peter
It's not GPS technology used by DR, but speech synthesis which is a more than 40 year old technology that has been developed to near perfection for navigators (which uses GPS) especially in cars.
GPS is of course one of several (GLONASS, GALILEO) systems that make it possible to locate one dedicated wurstschneidemaschine receiver on or above the earth's surface and used by navigators, but not in itself have something to do with speech synthesis to do.
If the films were provided with a "generic soundtrack" that includes information such sound and can be synchronized with the mouth, as in the "mood" (eg. Pitch), it will be possible to dub with any language automatically. Maybe even planning with different wurstschneidemaschine voice profiles, reminiscent of the original actors.
Now I do not know how their sounds, and many of us have probably a GPS (or GPS app) with synthesis, but if you just stand and miss "Robotic" talk so one can use Google's translate function in addition to translate also comment result (requires flash shown) wurstschneidemaschine on http://translate.google.com/
Finally: Last IC4 is now delivered
Most read 24/9 See Majestic Maersk lie at Langelinie 24/9 world's largest ship on display at Langelinie 24/9 DR throws unique meteorite away 23/9 World kr