On Jul 29, 2014, at 10:21 AM, Laurent Birtz <[email protected]> wrote:

> I see two problems with the patch.
> 
> It breaks the regression suite, in particular the HM-compatibility tests. 
> There must be some way to set vps_timing_info_present_flag back to 0. Add a 
> parameter to disable timing info in VPS, default to 0.
> 
> I'm not convinced the patch does what you think it does. 
> vps_num_units_in_tick and vps_time_scale specify how long a tick is in 
> seconds. It doesn't specify what to do with that information.
> 
> You specify vps_poc_proportional_to_timing_flag=0. According to the spec, 
> that means that there is no relation between the clock rate above and the 
> frame POCs.
> 
> Regards,
> Laurent
> --
> To unsubscribe visit http://f265.org
> or send a mail to [email protected].
> 

Hi,

I’ve made the patch more friendly towards regression compatibility. The use 
will have to specify “vps-timing=1” on the command-line to have the info 
written to the VPS.

Here’s the updated patch.

Thanks,
François Caron

Attachment: vps_timing.patch
Description: vps_timing.patch

Reply via email to