MythTV Developers Plan Xv, XvMC, OpenGL Changes

[phoronix]

Phoronix: MythTV Developers Plan Xv, XvMC, OpenGL Changes

MythTV 0.24 was released one month ago on their expedited release schedule, but the developers of this popular free software multimedia project are beginning to focus on their next release: MythTV 0.25. This release will drop support for XvMC (X-Video Motion Compensation) and libmpeg2 decoding and they also plan to drop Xv (X-Video) support in due time as well...

http://www.phoronix.com/vr.php?view=ODg2Mw

[Chewi]

I'm still using Xv for lack of anything better with the Radeon 4670. Looks like they're not leaving us out in the cold but this wasn't the approach I was expecting. Wouldn't it be better to help out with the recent Gallium3D efforts?

[mattst88]

Dropping XvMC... ugh. But dropping Xv? That's just stupid.

[Chewi]

Actually I have an old Toshiba Port?g? tablet hybrid that I run MythTV once in a blue moon. It has a Trident Cyberblade, which isn't much fun but the driver does at least have Xv support. It only just about manages with that but with no Mesa support, it wouldn't work at all with anything else. Oh well.

[Milan]

As said, dropping XvMC is fine, but Xv?! What, they want now their users to buy new hardware, some deal with hardware vendors or what...

[monraaf]

XvMC support these days on Linux is also rather limited to just the Intel graphics driver,

IIRC Catalyst also has XvMC support, or at least it used to have it. Don't know if they dropped it.

X-Video is supported by more open and closed-source Linux drivers drivers, but it is even of less benefit than X-Video Motion Compensation.

I don't know why you're always bashing Xv and referring to it as 'being useless' in your articles. It's not. For content up to 720p, Xv will do just fine even with a not so powerful CPU. Please more facts and less opinions in your articles.

[iamlindoro]

Dropping XvMC... ugh. But dropping Xv? That's just stupid.

No, it's really not. First off, note that we didn't announce that we are dropping Xv for our next release, just that it will likely be removed in the future.

With Xv, we cannot easily composite in an rich OSD at full screen resolution (it's limited to the resolution of the source material for performance reasons) so any OSD on SD content on an HD display looks like death. We cannot perform transformation and effects, animate it, shrink it down into a corner, composite the rest of the UI over it, run a mini video player, or do anything else that one has come to expect of a graphically rich modern media center UI.

To give you a comparison, XBMC uses OpenGL exclusively for display. They don't support XvMC, Xv, or other one-off implementations. Even VDPAU decode is displayed with OpenGL. Though we Myth devs prefer to use MythTV, there is no denying that their UI is flexible and offers themers the opportunity to produce an experience that is rich, beautiful, and full of eye candy. If we want to provide a comparable experience/opportunity for our themers, this means consolidating on a display mechanism that allows it-- and one that has existed for over a decade.

We're providing one release of notice before dropping XvMC and libmpeg2 support, and at least two in advance of dropping Xv. that means that XvMC users can stop upgrading, and Xv users get at least one more. Where do we draw the line? How long should we hamstring MythTV to support a 17 year old API? We're trying to provide adequate warning so that those who wish to join us as we move forward can begin to plan for it, and so that those who cannot or choose not to upgrade their hardware can be aware not to upgrade when the upcoming releases are available.

Hope this clarifies things.

Robert
MythTV

[Chewi]

Thanks for backing up your decision with some strong points, Robert, and thanks also for the advance warning. I may ditch that old machine by then.

[gbee]

As said, dropping XvMC is fine, but Xv?! What, they want now their users to buy new hardware, some deal with hardware vendors or what...

On the contrary, we welcome users using their existing hardware with their existing versions of MythTV. You will be able to use MythTV 0.24 for as long as you want.

Sadly you can't have it both ways, endlessly upgrading to the latest software for new features but never upgrading your hardware. Many new features inevitably come at the cost of increased hardware requirements and these new features tend to clash with older code.

Would you have us maintain support for older hardware indefinitely? How long before you complain of slow development progress and code bloat when we do?

Sincerely we have no intention of abandoning our users, we simply ask that they meet us half-way. If you want MythTV to be a modern media centre then we must be free to use modern protocols and hardware.

[gbee]

Sorry, I forgot to include my affiliations in that last post. I am a developer on MythTV, the news post announcing our plans was written by me.

I'd just like to reiterate what Robert said. We've set no firm date for dropping Xv support but we thought it was only fair to warn users as early as possible. By the time it finally happens good OpenGL support across a range of drivers and hardware will almost certainly be available.

Stuart
MythTV