W3C home > Mailing lists > Public > public-html-media@w3.org > February 2014

Re: [MSE] About the buffered attribute

From: Aaron Colwell <acolwell@google.com>
Date: Wed, 5 Feb 2014 08:57:47 -0800
Message-ID: <CAA0c1bA-CpDvU3F8ghrY0PUUO9EEBH27cVuBUeC=EfJ8q=QHLw@mail.gmail.com>
To: Cyril Concolato <cyril.concolato@telecom-paristech.fr>
Cc: "public-html-media@w3.org" <public-html-media@w3.org>
Comments inline.


On Wed, Feb 5, 2014 at 8:45 AM, Cyril Concolato <
cyril.concolato@telecom-paristech.fr> wrote:

> Hi all,
>
> It is unclear to me how the SourceBuffer.buffered attribute should behave
> in two cases:
> - before the first segment has been appended,
> - when an initialisation segment containing no data has been appended.
>
> In the first case, the spec says "The buffer is empty when the
> SourceBuffer object is created." The term "empty" also used, in the MSE
> spec, in the description of the HTML5 buffered/seekable attributes. Is it
> an object with a single range (0,0) or with no range at all?
>

An empty range is a TimeRanges object with length = 0. It can't be a single
[0,0] range because that implies that we actually have data for time 0
since the ranges are inclusive.


>
> For the second case, since no data has been received, I would assume it's
> stays in the same state, but I couldn't find any text in the spec. I think
> it should be clarified.


Since the initialization segment doesn't have any media data, it seems
pretty clear that there can't be any modification of buffered time since
the buffered attribute reflects "a new static normalized TimeRanges object
for the *media segments* buffered."

Aaron


>
> Cyril
>
> --
> Cyril Concolato
> Multimedia Group / Telecom ParisTech
> http://concolato.wp.mines-telecom.fr/
> @cconcolato
>
>
>
Received on Wednesday, 5 February 2014 16:58:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 15:48:44 UTC