[vls-devel] Re: Patch to regulate network output

[Jean-Paul Saman]

Marian Durkovic wrote:
> Hi all,
> 
> 
>   uff, first of all there's a fatal bug in WaitSendDate() which results
> in m_dSlope being treated as (s64) which means values like 0.9 are
> considered zero and all packets are outtputted immediately.
> 
>   The first line of WaitSendDate() should be changed to:
> 
>   s64 iSendDate = m_iLastTime + m_iDeltaClock + (s64)(m_dSlope*m_uiByteRead);
> 
> i.e. parenthessis need to be added around m_dSlope*m_uiByteRead and only
> the result changed to s64.
> 
>   After the above change the bigger blocks are only transmitted as a result
> of extrapolation. Using interpolation as suggested by Jean-Paul 
> will solve it perfectly but I don't know how to find the next PCR in the 
> buffer. Could someone please help?
> 
Try the patch I send you that does intrapolation if it can and falls 
back to extrapolation if there is no other choice. It already seeks the 
next PCR.


>   Now also the input from MPEG2 encoder board works fine, but usePCR flag
> needs to be changed into true in videoinput.cpp
> 
>   In any case, high-resolution timers are desperately needed.
> 
> 


greetings,
Jean-Paul Saman.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: vls_patch_for_marian_durkovic.patch
Url: http://mailman.videolan.org/pipermail/vls-devel/attachments/20030305/ccc46be0/attachment.txt