[FFmpeg-soc] [soc] libavsequencer [PATCH 01/08] Music module public API header file.

[Sebastian Vater]

Vitor Sessak a écrit :
> On 07/13/2010 10:15 PM, Sebastian Vater wrote:
>> Ronald S. Bultje a écrit :
>>> Hi,
>>>
>>> On Tue, Jul 13, 2010 at 6:04 AM, Vitor
>>> Sessak<vitor1001-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org>  wrote:
>>>
>>>> On 07/07/2010 10:46 PM, Sebastian Vater wrote:
>>>>
>>>>>     /** Duration of the module, in AV_TIME_BASE fractional
>>>>>        seconds. This is the total sum of all sub-song durations
>>>>>        this module contains.  */
>>>>>     uint64_t duration;
>>>>>
>>>> Is this ever written in the files or is it calculated by the
>>>> player? If the
>>>> later, it does not belong to the BSS.
>>>>
>>>
>>> Either way this belongs in AVFormatContext and the demuxer.
>>>
>>
>> Don't forget that the composer can edit this value in case if ever
>> needed. TuComposer has a synth-sound assember which should be a turing
>> machine making it impossible under certain situations to calculate
>> playback time correctly, so the user can edit and adjust that manually.
>>
>> Also some formats don't have even something close like a clear song-end
>> detection. In that case, the user can add it manually. That's the
>> purpose of this field.
>
> From your explanation this field has two, possible conflicting, meanings:
>
> a) The known duration of the file
> b) The time after which the song should be forced by the player to finish
>
> (a) should be in lavf, while (b) should have zero as default value,
> meaning never force the song to end.
>
> I think you should leave this field just for (b) (and document it
> accordingly) and set (a) in lavf. I'd even rename it
> forced_song_end_time.

I fully agree! forced_duration would be OK, though?

Another possibility would to mention that in the comment to this field.

-- 

Best regards,
                   :-) Basty/CDGS (-: