On 11/26/2011 10:51 AM, Luca Barbato wrote:

> On 26/11/11 16:07, Justin Ruggles wrote:
>> On 11/24/2011 03:34 AM, Luca Barbato wrote:
>>
>>> On 24/11/11 00:49, Justin Ruggles wrote:
>>>> ist->next_pts += ((int64_t)AV_TIME_BASE * decoded_frame->nb_samples) /
>>>> +                     avctx->sample_rate;
>>>
>>> ist->next_pts += ((int64_t)AV_TIME_BASE * decoded_frame->nb_samples) /
>>> +                     avctx->sample_rate;
>>>
>>> Is duration less reliable?
>>
>>
>> The original packet duration can be less accurate if it's based on
>> what's in the input stream. Or depending on what libavformat does, it
>> can be a made-up value that's just plain wrong.
> 
> Ok, put a @warning or a FIXME about it please ^^


ok. for audio, this is mostly needed for multiple frames in a packet. if
the input stream has pts, next_pts would not match packet pts only in
that case. if the stream doesn't have pts, it probably doesn't have
packet durations either.

-Justin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to