[Meego-tv] Candidate specification for a generic video player - OpenMax-AL

Dominig ar Foll dominig.arfoll at fridu.net
Thu Mar 17 13:04:33 PDT 2011


Le 17/03/2011 14:11, Teemu Tuominen a écrit :
>
>
> Why OpenMAX AL isn't mentioned, discussed or considered in open - for
> Meego ? Even though I'm not the expert on it, it would look as
> promising as OpenGL has been for graphics and I bet it would make
> Meego more interesting from manufacturers (potential adopters) point
> of view as well.
Actually if you search the old mailing list you will notice that
Gstreamer / OpenMax IL was discussed a the beginning of the project. But
at the time, the advance in the number of format (codec and transport)
supported by GStreamer, its defacto acceptation in the Linux PC world
and its capability to accept OpenMax-IL components via gstopenmax bin,
seemed to have made the selection of Gstreamer a no brainer.
I honestly believe that the situation has not changed much and that a
similar discussion even today, we would likely end with a similar
selection. We need to remember that the goal of MeeGo is support all
type of embedded applications and not simply mobile phones.
I would be very interested to get OpenMax community alternative view on
that position.
>
> If the thing is that no-one has evolved OMX AL for Linux that far (to
> be generalized), I'm just thinking it could and further maybe even
> placed as UMMS. Would be nice to see more of manufacturers
> contributing into open Meego under the bearer of OpenMAX, even though
> the passing of Khronos compliancy requirements might not be near possible.
On that point you are perfectly correct. a year ago, while I worked for
Amino and was starting the Cubovision project for Telecom Italia (STB
based on MeeGo 1.0), I looked seriously at OpenMax-AL and was quite sad
not to be able to benefit from it.
Today, I just re-read the latest specifications of OpenMAX-AL (1.1 18
Jan 2011) and OpenMax-AL-Digital-TV (provisional spec 18 Jan 2011) and
there is obviously a lot of communality between the UMMS vision and
Kronos latest proposals.

My reading of Kronos specifications leaves me with a few serious doubt
that I hope other member of this list will help to clarify or even
better dissolve:

OpenMax requires to integrate the player in the application, what goes
against the Goal of the UMMS proposal which requires to
  a) isolate the licenses,
  b) isolate the application execution from the player execution (for
content protection model). While I agree that on a basic media play back
that has little influence, it does impose to qualify the application
with the entire system when a Pay TV protection is required for the content.

I know that many people of the Open Source community believe that CA and
DRM are something passé (I personnally asked the question to Linus
Torvald, last week in an open forum, and his response was, without
surprise, bluntly following that model), it remains that to ship any TV
product, it must still be implemented and that is not going to change in
a short or medium term. In clear what ever is our desired to be open, if
we want to to play in the TV domain with Linux, we need to cater for CA
and DRM support at least for the next foreseeable future.
>
> Meego architecture & compliancy is seen quite focused atm, due to
> gles2 for example and numerous other must-have support requirements. I
> would eventually like to see Meego in very lightweight OpenVG based
> devices as well and the Gstreamer "all-purpose" multimedia would not
> fit into that picture but OMX AL might?
>
While I have no real idea if OpenVG has a great future in front of it or
not, I am convinced that all new generation of SoC, used in mobile
platform and multi media domains, will support OpenGL (at least ES 2.0).
But I do agree with you that while Gstreamer is great all rounder, in
specialised domains, it's applicability is not optimal. You may have
noticed that for Blu-Ray stack for example, I do not expect to use
GStreamer but I see a bespoke stack (more than likely proprietary)
running in parallell with GStreamer under the control of UMMS, there is
no raison while we would not be able to use the same strategy to
overcome low-end platform limitation.

Dominig ar Foll
MeeGo TV
Open Source Technology Centre
Intel SSG

> _______________________________________________
> MeeGo-tv mailing list
> MeeGo-tv at lists.meego.com
> http://lists.meego.com/listinfo/meego-tv


More information about the MeeGo-tv mailing list