[x264-devel] Re: Problems with PTS and b frames

[Bill May]

bond wrote:
> lo bill
> 
> are you saying x264's .mp4 output is not correct with b-pyramid? if yes can
> you plz post the exact cmdl to use to reproduce this, cause in my tests x264
> always produced the same pts/dts than mp4creator with any b-frame/pyramid
> settings

No.  I'm not saying that.  I'm saying that I'm using the i_pts output,
along with the formula, and not getting the correct output.

Bill
> 
> cheers
> bond
> 
> ----- Original Message ----- 
> From: "Bill May" <wmay at cisco.com>
> To: <x264-devel at videolan.org>
> Sent: Wednesday, December 21, 2005 7:25 PM
> Subject: [x264-devel] Re: Problems with PTS and b frames
> 
> 
>> Hi, again,
>>
>> A couple of days ago, I asked the below, in reference to the amount to add
>> to the output.i_pts to get the correct PTS.
>>
>> Bill May wrote:
>>> Is there a way to get max_reorder_buffer_size programmatically ?  It
> looks
>>> like p_mp4->i_init_delay is what I want - is that correct ?  It uses
>>> p_param->i_bframe ? (p_param->b_bframe_pyramid ? 2 : 1) : 0.
>> Without b frame pyramid, this works correctly.
>>
>> However with a b_frame_pyramid (set by:
>>     m_param.i_bframe = 3;
>> +  m_param.i_frame_reference = 5;
>> +  if (m_param.i_bframe >= 2)
>> +    m_param.b_bframe_pyramid = 1;
>>
>> the pts is not getting set correct.
>>
>> pts offsets: (should be values determined by mp4creator (Laurent's code)
>>
>> should be:      is:
>> 12000         12000   (IDR-I)
>> 30000         26400   (NIDR-P)
>> 12000         12000   (NIDR-B)
>> 0              1200   (NIDR-B)
>> 6000           8400   (NIDR-B)
>> 30000         30000   (NIDR-P)  ***
>> 12000         12000   (NIDR-B)
>> 0              1200   (NIDR-B)
>> 6000           4800   (NIDR-B)
>> 30000         30000   (NIDR-P)
>> 12000         12000   (NIDR-B)
>> 0              1200   (NIDR-B)
>> 6000           4800   (NIDR-B)
>> 30000         30000   (NIDR-P)
>> 12000         12000   (NIDR-B)
>> 0                 0   (NIDR-B)
>> 6000           8400   (NIDR-B)
>>
>> repeats from *** above.
>>
>> Any idea ?
>>
>> Thanks, and happy holidays.
>>
>> Bill May
>>
>> -- 
>> This is the x264-devel mailing-list
>> To unsubscribe, go to: http://developers.videolan.org/lists.html
>>
>>
>>
>> -- 
>> No virus found in this incoming message.
>> Checked by AVG Free Edition.
>> Version: 7.1.371 / Virus Database: 267.14.2/208 - Release Date: 20.12.2005
>>
>>
> 

-- 
This is the x264-devel mailing-list
To unsubscribe, go to: http://developers.videolan.org/lists.html