[pr0xZen]

Ok, I installed the test-build, replaced userdata and gave it a go. I played the same file from start to end, as in the previous test.

First go: Playback off local storage (SSD): http://xbmclogs.com/show.php?id=185629

Second go: Off wired NAS storage: http://xbmclogs.com/show.php?id=185631

I did stop the auto-update library to trim down the logs. On the second go, I didn't wait long enough - so some stuff is going on in the background at the very beginning). Both tests seem to exibit the same errors as the previous test above (atleast the DXVA error).

I really want to do more here, but I'm not quite sure what. Hoping you might have some good suggestions.

[FernetMenta]

Either your hardware or drivers causing the issue or the files you play. Does it happen at the same positions or randomly? If you provide a sample I can try here and we can compare the results.

[pr0xZen]

With this file the errors seem to happen at the same spots - could be the file. I'll do some more playback tonight and see how it works with other media.
Clip details on PM as soon as its up (ADSL, low outbound bandwidth). I don't have any decent input controls, so editing is difficult - up'ing the whole thing. Timestamping in the logs should give a good indication on where the errors occur.

Edit: Sent.

[pr0xZen]

New run, different video, nas storage. Playback from ~line 1220, noticed artifacts here and there throughout the video - but not very dominant.
http://xbmclogs.com/show.php?id=185860

[FernetMenta]

I have played this sample on various testsystems with different hardware. The short pixelation at 10:30 is present on all systems, even when using Windows Media Player. Must be an issue with the material itself.

[pr0xZen]

Thank you very much for testing this. Guess it was a poor choice of file for testing. Saturday morning now - I'll get some more various playback done, and report back.

[pr0xZen]

On some media it seems to be a very decent improvement, on other it seems "just as bad". Storage is distributed according to contents "category". so I think I'm going to have to check on that - try altering storage for new media. If that DXVA error is as you're saying, I'm a bit confused. A few bits of media (that exibit these issues) has gone through bittorrent transfer. If there were any data corruption, then shouldn't the hash checks pick up on that?

I've had a suspicion of this being caused by improper upscaling as I mostly see this on SD contents - but that very same DXVA error (2 of...) seem present pretty much everywhere corresponding to the "larger" video corruptions like the one you saw too - - with only a few minor other artifacts here and there on rather hastily done encodes (wich I definitely attribute to the media quality).