GTX 1050TI VDPAU deinterlacers.

We are able to reproduce this issue. Our engineering team looking to it.

hi sandipt
is this issue solved ?
i ask bcause i am planning to buy a gtx1050 .
thank U

We are still investigating this issue.

I ordered a Quadro P400 specifically for it’s video decoding capabilities. Do you know if it is also affected and will there be a fix? If there is no fix, I will have to return the card when it arrives.

Please answer, I’m picking up the Quadro today so it would be nice to know if I can open the package or not.

Hi Sandip,

The problem/issue has changed with latest driver 381.09.
The temporal combing/smearing has gone but there is still a problem with 2xAdv DI, still not useable (IMO).

The issue is with stray pixels detached from moving edges.
This problem could have been masked/covered up by the gross temporal combing i.e. could have always been there.

Have attached new photos & video clip of ‘Cheses Slicies’ 1080i50 de-interlacing test track using:-

  • GTX 1050 TI
  • driver 381.09 from ubuntu prop. graphics driver ppa
  • VDPAU 2xAdv DI
  • MythTV playback.
  • custom 50Hz modeline
  • 4K LCD (video card scaling 1080 → 4K native res)

Sorry for the ugly video container (& low res) but that’s only option on the camera.

Thanks.
DSCF3227.AVI (10 MB)
DSCF3225.AVI (19.9 MB)







Sandip,

The default MythTV video playback paused state is to render one decoded field of interlaced video.
This is always been annoying/ugly.

So to get better photos, I hacked MythTV source to make playback pause display the full VDPAU de-interlaced frame.
Attached some static (paused) photos of advanced x2 de-interlaced rendered frame.




Hi Bipeda, We also noted that with 381 some corruption is gone but some still left. We are trying to resolve the issue.

Hi tecfreak,
Do you have reproduce steps for this issue VDPAU HQ Scaling issue? Can you share video clip you are using to test it. What desktop env you are running KDE, GNOME, Unity ? Please provide nvidia bug report. Please share video or photo showing this issue?

Got mail from tecfreak

Hello Sandip,

sorry for my late response. The last two weeks I had no access to my rig.
I was not able to reproduce the hq scaling issue, neither with 381.09 nor wit the previous 378.13 driver. I suppose it was some kind of software issue on my side. Sorry for the false alarm.

But I can confirm the deinterlacer issues Bipedal recently reported. I could reproduce it using mplayer and 381.09. Let me know if you need a bugreport and/or additional data.

Best Regards,
tecfreak

Hello,

with 381.22 I still have the same artifacts bipedal reported, so deinterlacing is still broken.
Is it a hardware issue or why can’t you solve this problem? It has been 4 months since this problem has been reported.

Best Regards,
tecfreak

It’s being investigated, but I don’t have an ETA for a fix.

Hello,

does this problem affect the new released GT 1030 card (GP108/Pascal) also?

Thanks and regards,
ciax

Hi,

the GT 1030 seems to be very interresting card for Homecinema and TV (DVBT2/Hevec265). I think that is a card a lot of people waiting for.

But with broken deinterlacing like GTX1050 the card is useless…

Does the problem affect the new GT 1030 too ?

Thanks and regards,

Perlbo

… just ordered such a GT 1030 card. Would be really annoying, if the iinterlacing problem exists with this graphics board too. I think there would be a lot of linux htpc’s using this card, cause you can pimp an older PC with it.

@Aaron: Any news for us? :o)

Thanks and regards!

Hi,

i have buy a GT1030 and this card is also affected. Is there any solution?

Some corruption gone with r381 but still some corruption is visible. We are investigating it.

Hello,

is there any hope to expect an improvement in the next driver version concerning this topic?

Thanks,
ciax

Thanks for your patience. We have fixed this issue and its will be available in upcoming driver release.

Hello,
have you any release date for the new driver.

Thanks,
motorsense