Intel NUC - Haswell (4th Generation CPU) - Printable Version +- Kodi Community Forum (https://forum.kodi.tv) +-- Forum: Discussions (https://forum.kodi.tv/forumdisplay.php?fid=222) +--- Forum: Hardware (https://forum.kodi.tv/forumdisplay.php?fid=112) +--- Thread: Intel NUC - Haswell (4th Generation CPU) (/showthread.php?tid=176718) 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
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
|
RE: Intel NUC - HTPC (Haswell Late 2013 edition) - fritsch - 2014-03-24 Once scaled from full to limited, it's non reversible. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - axbmcuser - 2014-03-24 (2014-03-06, 20:21)zag Wrote: For me r17780 is the last working release using livetv and hardware acceleration. I also get blocking/artefacts with all the latest nightlies with Live-TV on my D54250WYK(H). As you mentioned, videos are fine. I tried an old build around the one you mentioned (i used r17733) and with that one hardware accerlation + Live TV works fine. +1 from me regarding what you experience. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - zag - 2014-03-24 (2014-03-24, 15:59)axbmcuser Wrote:(2014-03-06, 20:21)zag Wrote: For me r17780 is the last working release using livetv and hardware acceleration. Post a debug log so we can see whats going on. I have a feeling it was the openelec commit that has caused this. Every new nightly has blocking on live tv now. EDIT: Going back to openelec nightly r17780 from 21st February, WORKING EDIT2: Going back to openelec nightly r17824 from 27th Feburary, BROKEN RE: Intel NUC - HTPC (Haswell Late 2013 edition) - noggin - 2014-03-24 (2014-03-24, 14:56)fritsch Wrote: Once scaled from full to limited, it's non reversible. Yes - but from Limited to Full it is (ignoring <16 BTB and >235 WTW clipping) isn't it? That was my point re: banding. So if you have a Limited source file, scale to Full (introducing banding), and then scale back to Limited, that can remove the banding can't it? Doesn't every Limited value scale to a unique Full value (with some Full values not being filled, creating the banding jumps?) thus allowing a lossless return to Limited (ignoring BTB and WTW) Won't this remove any banding that has been introduced by the Limited->Full scaling on the return Full->Limited scale. Isn't there a 1:1 mapping from Limited -> Full (which is thus reversible?) Obviously if you start with a native Full signal, then you will lose information, but if your Full signal is itself derived from a Limited signal, won't that have no content in the areas where the information will be lost from? I guess I'm imagining two 256 (or 1024 for 10 bit) value LUTS as that would be how I would have done it back in my hardware days! (256 bytes of ROM LUT - with the input video used to address the memory and the output video being the content of that memory location.) RE: Intel NUC - HTPC (Haswell Late 2013 edition) - fritsch - 2014-03-24 Check: Input: Full Range 0 .. 255 Scale to Limited: sl(255) = ? sl(235) = ? scale back to full: sf(sl(255)) = ? sf(sl(235) = ? Assumption linear weighting. Please solve the equations. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - noggin - 2014-03-24 (2014-03-20, 06:57)trsqr Wrote: August T210 should be supported in kernel 3.14. I've ordered one from China, but it takes weeks to arrive... I don't know any other USB T2 tuners that do work with Linux. I took a punt on an August T210, and already have a PCTV 290e. I installed a 3.14 kernel on a fresh Ubuntu install, and built TV Headend. The T210 is seen by the kernel and registered as a DVB T and DVB C frontend. However I can only get it to tune DVB-T Freeview muxes. When I try DVB-T2 Freeview HD muxes it fails to find any services. Hmm... (2014-03-24, 17:30)fritsch Wrote: Check: You're missing my point. Your example is the opposite of the situation that I was discussing. I'm assuming the input is Limited range - as almost all source video is (DVD, Blu-ray, DVB, ATSC etc.). I'm assuming the output is Limited range, as many of us want our HDMI outputs to be (to match other consumer gear) However there may be a Full range process (XBMC?) in the way. There was a statement made that this double scaling introduced banding on the output as well as clipping <16 and >235 content. Whilst I agree about the <16 and >235 content - isn't the Limited-Full-Limited processing potentially transparent? Input : Limited Range 16-235 Scale to Full Scale back to Limited for output I totally agree that in your case you will inevitably lose information as the scaling is from a wider range to a narrower range. However in the case I was talking about the scaling is from narrower range to wider range, which won't lose content, and back again, which won't lose content as there isn't any content at the levels that will be lost by the reverse scaling. In other words in the case of XBMC playing 16-235 video and outputting 16-235 video, my point was that the two scaling processes (Limited->Full followed by Full->Limited) don't inevitably create banding and can be lossless. I think we are talking at crossed purposes. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - lmyllari - 2014-03-24 (2014-03-24, 13:05)kouze Wrote:With current OE you can use VAAPI decoding, SW filtering and 16-235 xbmc setting. You won't even need my passthrough patch with a Panasonic - it assumes limited range. You just need to use xrandr to set full range output (to avoid GPU scaling your limited range signal again).(2013-12-05, 02:17)lmyllari Wrote: The banding comes from expanding luma from 16-235 to 0-255. This is done for typical PC monitors, which expect a full range signal. New Intel linux drivers by default give TVs a limited range signal, taking the expanded 0-255 and scaling it to 16-235. This does not eliminate the banding that was created by the original expansion, and original information below 16 and above 235 was already lost in that same step. (Why do they do it? It's the right thing according to the standards, and some TVs don't like full range -> crushed blacks.)Hi guys, The SW filtering downloads decoded frames and puts them through the same path as software decoded frames -> the 16-235 setting in xbmc works fine and you get a nice untouched signal. No idea about Windows, sorry. Usually I don't associate Microsoft with "good RGB signal". (<- unfair stab at the whole xbox360 fiasco) (2014-03-24, 18:19)noggin Wrote: There was a statement made that this double scaling introduced banding on the output as well as clipping <16 and >235 content. Whilst I agree about the <16 and >235 content - isn't the Limited-Full-Limited processing potentially transparent?I think that would be ok if you were just scaling luminance back and forth, but you are also doing YCbCr -> RGB in the first step. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - Roby77 - 2014-03-24 Mmhm has this bug from several month Why it is not fixed yet ? It's possible to use ycbr 709 ? RE: Intel NUC - HTPC (Haswell Late 2013 edition) - fritsch - 2014-03-24 Cause you did not fix it, quite easy. I await your PR even daily, but it does not come. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - Roby77 - 2014-03-24 Mmh it wasn't a critique just a curiosity about the origin of the bug Sorry for misanterpretation caused by my bad english RE: Intel NUC - HTPC (Haswell Late 2013 edition) - noggin - 2014-03-24 (2014-03-24, 20:12)lmyllari Wrote: I think that would be ok if you were just scaling luminance back and forth, but you are also doing YCbCr -> RGB in the first step.That's a very good point. YCrCb/RGB round-tripping in 8-bit truncated space isn't a good idea is it... I knocked up an Excel sheet to go from Limited to Full to Limited level space - with rounding to nearest integer - but hadn't considered the RGB matrix. Presumably XBMC copes with ITU 601 and 709 levelspace on input content (either by reading header information or by taking a punt that <=720 horizontally is 601 and >720 horizontally is 709. I assume you'd use horizontal resolution rather than vertical to cope with the letterbox cropping that happens on >16:9 aspect ration 720p and 1080p sourced content?) I assume there is some driver interaction on the output level space and RGB to YCrCb conversion - so no real control over whether output is 601 or 709 colour space if YCrCb? RE: Intel NUC - HTPC (Haswell Late 2013 edition) - kouze - 2014-03-24 (2014-03-24, 20:12)lmyllari Wrote: With current OE you can use VAAPI decoding, SW filtering and 16-235 xbmc setting. You won't even need my passthrough patch with a Panasonic - it assumes limited range. You just need to use xrandr to set full range output (to avoid GPU scaling your limited range signal again).Brilliant ! Thank you so much ! RE: Intel NUC - HTPC (Haswell Late 2013 edition) - deaerator - 2014-03-26 (2014-03-24, 20:12)lmyllari Wrote: With current OE you can use VAAPI decoding, SW filtering and 16-235 xbmc setting. You won't even need my passthrough patch with a Panasonic - it assumes limited range. You just need to use xrandr to set full range output (to avoid GPU scaling your limited range signal again). Can you explain in detail how to set this up or is there a sticky somewhere. RE: Intel NUC - HTPC (Haswell Late 2013 edition) - kouze - 2014-03-26 (2014-03-26, 12:30)deaerator Wrote:SSH to openelec box and type:(2014-03-24, 20:12)lmyllari Wrote: With current OE you can use VAAPI decoding, SW filtering and 16-235 xbmc setting. You won't even need my passthrough patch with a Panasonic - it assumes limited range. You just need to use xrandr to set full range output (to avoid GPU scaling your limited range signal again). cd /storage/.config touch autostart.sh chmod 0777 autostart.sh nano autostart.sh # RGB fix xrandr --output HDMI1 --set "Broadcast RGB" "Full" The other parameters are found in XBMC directly in expert mode RE: Intel NUC - HTPC (Haswell Late 2013 edition) - micoba - 2014-03-26 I got two external hdds connected via usb 3.0, but one just won't go into standby. Does anyone know how to force standby in win 8.1? It must be some driver issue, as the same hdd in the same external case didn't always cause this problem. Also, the copy speeds between win 7 and the nuc are too low - going down from 80 to under 40mb/s in about 2 minutes... |