•   
  • 1
  • 210
  • 211
  • 212(current)
  • 213
  • 214
  • 355
  •   
  Thread Closed
v18 LibreELEC Testbuilds for x86_64 (Kodi 18.0)
(2018-04-19, 21:33)rudrood Wrote: Here's the debug log, after reboot playing the two sample files:
  1. H.264 1080p/23.976 (MKV)
  2. H.264 1080p/24.000 (MKV)
http://ix.io/18dR

@Milhouse: Yes "adjust display refresh rate" is on, and working cause it switches TVHeadend footage. 

You don't have a 24.0 Mode - therefore it will stay on 60 hz for this video.

But this:
Quote:21:28:19.532 T:140681716432256 DEBUG: display width: 1920 vs video width: 1920
21:28:19.532 T:140681716432256 DEBUG: display fps: 23.977369 vs video fps: 23.976025
21:28:19.532 T:140681716432256 DEBUG: display flags: 0 vs video flags: 0

is wrong. The code tries to computer the float directly and matches too many numbers behind the "." therefore it won't match. It should try to do something like:
if (int) (fps * 100) == (int) refreshrate * 100) - that way it would only match two positions after the "."

Though - what I wonder: Your refreshrate - is that again hand tuned? It's quite wrong :-)
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
No hand tuning, all "right out of the box".
I'm a noob about this but shouln't the pre whitelist methode give me and the system better results even for the not availible 24 fps (at least it switches to 24 fps on the TV and apperently take the closest available fps) instead of staying at 60 fps?
(2018-04-19, 07:33)debutanker Wrote: I mean, if i rapidly press any buttons several times on my harmony remotes (i have one hub in my living room and one hub in my sleeping room), my NUC reacts without any delays, more straighter, compared to my HTPC with my RC6 IR-Receiver.
Both Hub's/Harmony's are having the same setup.
Interesting. I currently have no real explanation where this could come from.

Can you check if remotedelay is set identically on both boxes? https://kodi.wiki/view/advancedsettings.xml#remotedelay Just a wild guess, not 100% sure if that could be responsible but it'd still be worth to check.

Another idea: could you plug the USB IR receiver into your NUC? Just cover the built-in IR receiver of the NUC with black tape or put a book in front of it so only the USB IR receiver gets the signals.
Quote:With new build #0418 everything works - many thanks!
Thanks a lot for reporting back!

so long,

Hias
(2018-04-20, 12:40)rudrood Wrote: No hand tuning, all "right out of the box".
I'm a noob about this but shouln't the pre whitelist methode give me and the system better results even for the not availible 24 fps (at least it switches to 24 fps on the TV and apperently take the closest available fps) instead of staying at 60 fps?
 No. 23.976 is too low for 24.0 fps, 25 would need audio resampling. 60 is (3*24+2*24) / 2. But as said the matching is not good enough, it should not compare with == operator.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
(2018-04-20, 12:24)fritsch Wrote: But this:
Quote:21:28:19.532 T:140681716432256 DEBUG: display width: 1920 vs video width: 1920
21:28:19.532 T:140681716432256 DEBUG: display fps: 23.977369 vs video fps: 23.976025
21:28:19.532 T:140681716432256 DEBUG: display flags: 0 vs video flags: 0

is wrong. The code tries to computer the float directly and matches too many numbers behind the "." therefore it won't match. It should try to do something like:
if (int) (fps * 100) == (int) refreshrate * 100) - that way it would only match two positions after the "."

I think that will be fixed tonight as @lrusak has added a new commit to improve matching with fractional refresh rates.

https://github.com/xbmc/xbmc/pull/13274/...2960295bf9
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Forgot to say I really appreciate @fritsch & @Milhouse for thinking with me!
My KODI install only gives me a black screen + cursor on boot.
Network share and SSH works fine.
Tried to delete Addons27 and guisettings, but no luck.

Is it an known issue?
(2018-04-20, 21:46)SirMacke Wrote: My KODI install only gives me a black screen + cursor on boot.
Network share and SSH works fine.
Tried to delete Addons27 and guisettings, but no luck.

Is it an known issue?

No, no such issue is known.

To have any hope of knowing what is happening we will need your crash log, and if you don't have a crash log then "journalctl -a | pastebinit".

Also some indication of what version last worked for you would be useful.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
New LibreELEC.tv Leia build #0420: Generic
(Supercedes previous build)

SHA256 Checksum: e30c6a58887be11f490835f8d9717d0719c8d7769ec858d76fd203404dafe489 (Generic)

# uname -a
Linux NUC 4.14.34 #1 SMP Fri Apr 20 21:09:00 BST 2018 x86_64 GNU/Linux

# lsb_release
LibreELEC (Milhouse): devel-20180420210258-#0420-g6ce3be6 [Build #0420]

# Kodi version
(18.0-ALPHA2 Git:af4a548). Platform: Linux x86 64-bit

Based on tip of LibreELEC.tv master (6ce3be6, changelog) and tip of XBMC master (51df226, changelog) with the following modifications: Build Highlights:
  1. Updated DVB drivers
  2. Fix missing "Search" in blade menu
  3. Fix fractional refresh rate matching in Whitelist
Build Details:
  1. XBMC:
    • ffmpeg: drop some deprecated functions (PR:13791, 1 commit, 1 file changed)
    • Fix duplicate ID (PR:13788, 1 commit, 1 file changed)
    • GLES: allow using limited color range (PR:12506, 3 commits, 14 files changed)
  2. inputstream.adaptive:
    • fix effective_url from 2.2.15 change (cd6d9c9)
    • version bump 2.2.16 (aedfa50)
  3. Additional commits/pull requests/changes not yet merged upstream:
    • Added: [env] PR:2572 (perma): update dvb addons
    • Updated: [pkg] PR:13274 (perma): Mode whitelist
    • Added: [pkg] PR:13794 (perma): [guiinfo] Fix Container.CanFilter (CONTAINER_CAN_FILTER).
    • Reverted: [pkg] PR:13791: ffmpeg: drop some deprecated functions
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2018-04-20, 16:46)rudrood Wrote: Forgot to say I really appreciate @fritsch & @Milhouse for thinking with me!

Please test #0420 and confirm if 23.976 is still a problem (upload a new debug log if it is) - thanks!
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Tested and seems to work oke now Smile Thanks!
Will test a second PC also with a (newer) NVidia card tomorrow and let you guys know about that one.

Edit:
Found that whitelist listing shows 1920x1080p 60Hz twice (see: https://imgur.com/a/r3JW5vL )
(2018-04-21, 01:18)rudrood Wrote: Tested and seems to work oke now
Thanks.
(2018-04-21, 01:18)rudrood Wrote: Found that whitelist listing shows 1920x1080p 60Hz twice
Yes that's a known issue that will need fixing in the X11 side of things, should be OK to ignore one of them for now.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Is it possible to get a testbuild with Kernel 4.16.3 for testing AMD Ryzen 5 with Radeon Vega Graphics?

Regards Karsten
(2018-04-21, 12:08)KarstenL680 Wrote: Is it possible to get a testbuild with Kernel 4.16.3 for testing AMD Ryzen 5 with Radeon Vega Graphics?

Regards Karsten

If you ask on the LibreELEC forum you might find someone with a build based on a more recent kernel, possibly escalade.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
(2018-04-20, 13:09)HiassofT Wrote: Interesting. I currently have no real explanation where this could come from.

Can you check if remotedelay is set identically on both boxes? https://kodi.wiki/view/advancedsettings.xml#remotedelay Just a wild guess, not 100% sure if that could be responsible but it'd still be worth to check.

Another idea: could you plug the USB IR receiver into your NUC? Just cover the built-in IR receiver of the NUC with black tape or put a book in front of it so only the USB IR receiver gets the signals.


so long,

Hias

Many thanks for your efforts!

Short tests between beer and BBQ...
Tested my USB-receiver with my NUC and the NUC behaves like with integrated receiver: more directly compared to my living room pc - strange?!
Also tested with several delay-values in the advancedsettings.xml (had to create it) - no difference.

But as this is only a "little cosmetic thing" and there's not much difference, nothing which bothers me!
Asus H270M-Plus (i3-7100),NUC7i3BNH,WeTekHub,Pi2,Win10Server.
  •   
  • 1
  • 210
  • 211
  • 212(current)
  • 213
  • 214
  • 355
  •   
  Thread Closed
 
Thread Rating:
  • 24 Vote(s) - 4.5 Average



Logout Mark Read Team Forum Stats Members Help
LibreELEC Testbuilds for x86_64 (Kodi 18.0)4.524