[Ltru] Tagging of silent films
L.Gillam at surrey.ac.uk
Thu Sep 29 12:17:39 CEST 2005
An interesting discussion ......
If there is a silent movie in a different language, what does a lip reader do?
Would the audio identification be used to transform the lip movements in the
video stream? Certainly an interesting technology.
That aside, assuming a split of video and audio streams and a desire to analyse the
audio stream in isolation, what does one analyse in "silent English"? At this
level of tagging, I agree with Peter - "no linguistic content". "silent", for me,
feels like a description alongside "director" or "running time".
> > Peter Constable wrote:
> >> So we could potentially have two versions of the same film -- one
> >> that is "Silent-audio/French-text" and one that is
> >> "Silent-audio/English-text.")
> > So, it seems to me that 'audio = "no linguistic content" / text =
> > "French"' and 'audio = "no linguistic content" / text = "English"
> > would be appropriate metadata for this scenario.
> For a hearing person, maybe. But "not audible" is not the
> same as "no
> content". If an actor is mouthing, the content is definitely there. As
> Karen pointed out, a lip reader may be able to pick it up just fine.
> Stretching your reasoning I could say that "audio=no
> linguistic content"
> is appropriate metadata for a Chinese movie because I don't
> understand a
> word of Chinese <g>.
> In fact, I'd suggest something like
> "silent-English-audio/French-text", if only because it would help
> lipreaders in activating the appropriate "language module". You could
> even argue that the "silent" part is somewhat equivalent to the script
> subtag for written documents.
> Luc Pardon
> Ietf-languages mailing list
> Ietf-languages at alvestrand.no
More information about the Ietf-languages