[vlc-devel] [RFC-PATCH] core/audio_output: do not drop (late) audio blocks if it would cause silence

[Filip Roséen]

> In this case the duration is *only* an implementation detail of the decoder, 
> or possibly the demuxer.

Indeed, though I am not sure how to handle cases as the previosly mentioned
sample otherwise. The problem is of course that we discard more data than we
really should; and I (perhaps naively) thought that the patch could solve this
issue in a manner that would still aid user-experience.

 - Do you have an alternative idea that would yield the correct result? 

Writing code is the easy part, having it be theoretically and practically
correct is another. With my lack of experience I most often base my patches on
the practical issues they solve.

Once again; thanks for taking a look at another RFC from me! `:-)`

> Stating the obvious here, but if you want to cut at AOUT_MAX_PTS_DELAY, then 
> it´s not the length you should be looking at. It´s the start (pts) and the end 
> (pts + length).

Certainly, but since the drift (named "advance" in this particular piece of
code) is based on the PTS simply using the length together with the drift would
yield the same result.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/vlc-devel/attachments/20160512/d453f46c/attachment.html>