W3C home > Mailing lists > Public > whatwg@whatwg.org > June 2009

[whatwg] Start position of media resources

From: Ian Hickson <ian@hixie.ch>
Date: Fri, 5 Jun 2009 09:21:41 +0000 (UTC)
Message-ID: <Pine.LNX.4.62.0906050909290.1648@hixie.dreamhostps.com>
On Fri, 1 May 2009, Silvia Pfeiffer wrote:
> On Fri, May 1, 2009 at 3:00 AM, Ian Hickson <ian at hixie.ch> wrote:
> > On Thu, 30 Apr 2009, David Singer wrote:
> >> At 16:45 ?+0000 30/04/09, Ian Hickson wrote:
> >> > On Thu, 30 Apr 2009, David Singer wrote:
> >> > >
> >> > > If the resource is 'seekable' then time is relevant, and I agree
> >> > > that time should be a normal play time and run from 0 to duration.
> >> >
> >> > That wouldn't address the use case of files that were split with
> >> > non-zero start times, though, where the author wants the original
> >> > range to be the one visible in the UI.
> >>
> >> The complexity of edited files is only really dealt with by embedded
> >> time-codes. ?A single segment is the beginning of a large can of worms;
> >> what do you want to have happen when there are two segments played
> >> consecutively? Following your logic, there would be a time-jump.
> >
> > Yeah I'm not really sure how to handle this case. (I imagine it could get
> > even worse with multiple frames having the same timecode). I guess time is
> > linearised starting from the first frame's timecode?
> 
> I would suggest handling this as a playlist. It is the same fundamental problem.
> If there is a time reset, the timeline should also reset. Then there
> can be a prev/next file button for scanning.
> We haven't really solved the problem of having playlist references in
> <video> tags, so this is a whole discussion to itself IMHO.

Since I definitely don't want to add playlist functionality yet, I don't 
think that this is necessarily the best way to solve this problem at this 
time... it seems that requiring the browser to ensure that frames are 
referenced by a time position that increases linearly from the start of 
the clip is the better solution (though for such clips it does mean that 
the underlying timecodes get lost).

I've added this requirement to the spec. In practice I don't expect this 
will affect many people.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Friday, 5 June 2009 02:21:41 UTC

This archive was generated by hypermail 2.4.0 : Wednesday, 22 January 2020 16:59:13 UTC