[FFmpeg-devel] [PATCH v6 0/4] Initial implementation of TTML encoding/muxing

Jan Ekström jeebjp at gmail.com
Thu Mar 4 15:52:13 EET 2021


On Thu, Mar 4, 2021 at 11:57 AM Moritz Barsnick <barsnick at gmx.net> wrote:
>
> On Wed, Mar 03, 2021 at 22:09:12 +0200, Jan Ekström wrote:
> > Thus, I will be pulling this set in Soon (probably tomorrow). Thus if
> > anyone still wishes to have a say, I recommend doing it earlier than
> > later.
>
> > + * Copyright (c) 2020 24i
>
> Is 24i a company? Just wondering.
>

Yes. In the case of this patch set this is the case.

> Also, I'm not sure whether it's usually advised to split libavformat
> and libavcodec changes into two separate commits.

Yes, in general if the feature is usable by itself it could be split.
In this case the tests can only be run after both parts are there, and
thus there is only limited usefulness, as currently reading a TTML
document from MXF is supported, and I have not seen such samples out
in the wild.

That said, if we really want to go in pedantic we could split this as follows:

1. Add the muxer, only for the remux use case (from MXF, which we
cannot test since there is no samples).
2. Add the encoder and modify the muxer to be able to handle the
paragraph-based packets.

Alternatively, we can pull in the encoder first, then the muxer. But
the encoder by itself can only be tested through the API or so.

I... just am not so sure in this case splitting the TTML writer and
the TTML encoder from each other makes sense.

Jan


More information about the ffmpeg-devel mailing list