Linux VAAPI: Nuc, Chromebox, HSW, IVB, Baytrail with Ubuntu 14.04 - Printable Version +- Kodi Community Forum (https://forum.kodi.tv) +-- Forum: Support (https://forum.kodi.tv/forumdisplay.php?fid=33) +--- Forum: General Support (https://forum.kodi.tv/forumdisplay.php?fid=111) +---- Forum: Linux (https://forum.kodi.tv/forumdisplay.php?fid=52) +---- Thread: Linux VAAPI: Nuc, Chromebox, HSW, IVB, Baytrail with Ubuntu 14.04 (/showthread.php?tid=165707) Pages:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
|
RE: vaapi-sse4: Deinterlacing Testing - puithove - 2014-08-20 Really excited to try this out because I'm hoping that it can deinterlace the strange content from my cable provider whereas Yadif was having trouble. Question though - there was the driver bug that was apparently causing tearing & hanging so that you had to disable "Prefer VAAPI Render method". Is this still an issue? If you disable that, doesn't it disable the hardware assisted filtering including the deinterlace methods? RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-20 Just try it out. I added all the fixes that have been produced into the intel driver since then, including the joe_sample fix. RE: vaapi-sse4: Deinterlacing Testing - puithove - 2014-08-20 Sounds good, will try it as soon as I get a chance. RE: vaapi-sse4: Deinterlacing Testing - puithove - 2014-08-21 Soooo.... turns out MADI & MACI both have the same issue as the software Yadif with my cable content. RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-21 Not really sure if this issue is an issue with the deinterlacer. They are quite "dump" and only react on the way they are filled :-) RE: vaapi-sse4: Deinterlacing Testing - puithove - 2014-08-21 Yea, I agree, just providing the info. Not really surprised that it's the same result, was hoping that one might handle it better than the other though. Oh well. RE: vaapi-sse4: Deinterlacing Testing - robo989 - 2014-08-21 Dumbo question incoming, Which de-interlacer should give better outpute the new "motion-adaptive" or "motion-compensated". I have poor eyesight but others in the family don't but they lack the incling to properly look and I'm a fussy fool. Thanks RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-21 lol :-) RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-21 If you have some segfaults, e.g. xbmc is restarting itself without anything relevant in the logfiles. This problem known - it's a bug in the driver which segfaults. I have reported the relevant logfiles to the intel developer. Also remember that this driver which makes mcdi and madi usable is _a_prerelease_ driver. We are very thankful that gwenole of intel gave us this driver to get our code in shape. We hope for a 1.4.0 intel release which will have all the quality we will need. RE: vaapi-sse4: Deinterlacing Testing - noggin - 2014-08-24 Think I've just installed the new driver and XBMC/Kodi build from the PPAs. I get the new VAAPI options for de-interlacing. However on native 1080/50i content on a Haswell 2955U they are unusable. The de-interlaced frames are presented out of order, so you get very juddery, jerky motion as the frames jump forwards and backwards in time. Tried with both VAAPI preferred render options. VAAPI Bob, Bob and De-interlace are all fine. Log here : http://paste.ubuntu.com/8130473/ Am running XBMC 14.0-ALPHA3 Git : 93e742b (Compiled : Aug 24 2014) under Ubuntu 14.04 LTS, with XFCE as the XBMC window manager (I think) According to apt-cache I'm running : libva-intel-vaapi-driver: Installed: 1.3.3~pre3-1~trusty Have I done something wrong (have I installed the right Intel driver package?) - or are Haswell de-interlacing modes still borked? (I'm running on a Chromebox with a single 2GB RAM SODIMM and have no control over the amount of RAM allocated to the GPU - as the Chromebox doesn't have a user-editable BIOS as other x86 platforms do) RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-24 No. All tested hsw versions were fine until now. RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-24 Reboot and post vainfo Plesse. RE: vaapi-sse4: Deinterlacing Testing - noggin - 2014-08-24 Sure this is something I've done wrong... Here's the output of vainfo straight after a reboot : http://paste.ubuntu.com/8130618/ RE: vaapi-sse4: Deinterlacing Testing - fritsch - 2014-08-24 Wrong driver - it even tells version 1.3.0 please post the complete output - it's missing which drv actually is loaded. Has to look like that: Code: fritsch@t440s:~$ vainfo RE: vaapi-sse4: Deinterlacing Testing - noggin - 2014-08-24 Yes - thought it was the wrong driver - how do I update? I did an apt-get install libva-intel-vaapi-driver after adding the ppa for wsnipex's vaapi - is there something else I should do? I followed the vainfo guide in the original post which suggests : Code: DISPLAY=:0 vainfo | pastebinit I get this returned before the URL: Code: libva info: VA-API version 0.35.0 If I don't prefix with DISPLAY:=0 I get : Code: error: can't connect to X server! |