[multicat-devel] [PATCH] Allow seamless reads and DTS/PTS manipulation

Dr. Net! - Eugen Rieck eugen at drnet.at
Thu Dec 22 09:25:14 CET 2016


Hi again,


Am 2016-12-21 um 22:48 schrieb Christophe Massiot:
> I have just checked in a similar patch, without the while loop, and always enabled (it should be the default behaviour).

Perfect, this removes the need for the first part of my patch.

>
>> The less important second part adds the -o (for DTS) and -O (for PTS) flags for usage with restamping (-P). They allow offsets for DTS/PTS to acommodate e.g. slow DVB heads.
> However, I don’t understand why you need this. First of all, you should apply the same offset to DTS and PTS, because most elementary streams do not have a DTS so only the video would be impacted. But also, the resulting streams would be totally non-compliant with the TS buffering model.

I made that not clear enough: Yes it would. This (of course) doesn't stop some of the equipment out there from needing it.

In our special case, we use multicat (and other OSS tools) for scheduled playout of TS files to a multitude of DVB heads (different and competing cable providers in Europe) and came across one, where the transcoding heads would introduce visible stutter if they became CPU-bound. Manipulating the DTS (a few ms earlier) fixed the issue by tricking the heads into decoding the image a tad earlier and thus be able to process it a bit earlier.

This is a highly unusual and compliance-breaking edge case, that has no use at all if every part of the chain behaves as expected. If they all do.

So I have no problem at all with this second part of the patch being rejected - my rationale for submitting it was, that maybe others come across similar ill-behaving third-party equipment and can use it to work around such issues.

All the best and have nice holidays,

Eugen


-- 

:-) Eugen Rieck (-:
  eugen at drnet.at

The greatest test of an engineer is not his technical ingenuity 
but his ability to persuade those in power who do not want to be persuaded
and convince those for whom the evidence of their own eyes is anything but convincing.



More information about the multicat-devel mailing list