• 1
  • 66
  • 67
  • 68(current)
  • 69
  • 70
  • 89
OpenELEC Testbuilds for RaspberryPi (Kodi 15.0) Part 2
(2015-06-03, 16:34)Forage Wrote: Now that it's an add-on, is there a way to add and remove items from the list instead of having to replace existing ones? I didn't find a config file which includes the list with the names, but just in case I overlooked something. Not that it's a big deal, it's just for playing, I was just wondering.

You should subscribe to this thread where that functionality is already being discussed.
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.
I notice that the input from my TV remote with HDMI CEC is a lot slower than using an IR receiver on the Pi or Yatse.
Are there options to increase the speed of HDMI CEC?
(2015-06-03, 18:55)tdn135 Wrote: Are there options to increase the speed of HDMI CEC?

Yes, in the CEC Settings - you can increase/reduce the delay between button presses and repeats.
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 OpenELEC Isengard build #0603: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.0.4 #1 Wed Jun 3 21:02:34 BST 2015 armv6l GNU/Linux

# vcgencmd version
May 31 2015 14:24:30
Copyright (c) 2012 Broadcom
version b127f62bd972cebb74a816b01b90ecfacc226baa (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150603210146-#0603-g4681f97 [Build #0603]

# vcdbg log msg 2>&1 | grep DTOK
001574.862: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (4681f977, changelog) and tip of XBMC master (2363c66d, changelog) with the following modifications: Build Highlights:
  1. HEVC Update:
    popcornmix Wrote:Another commit that fixes the hevc cache issue (horizontal lines in Sintel 720p) and a commit: "Option to parallelise coefficient decode and inter prediction and deblock for each frame"
    which may improve performance.

    I believe heap corruption [GPU Mem free] was fixed in previous update.

    The Sintel 1080p artifacts we used to have were dangerous - incorrect scaling meant we were writing beyond end of image buffers, and that caused relocatable heap corruption.

    The Sintel 720p lines were actually harmless (although ugly). We were just using stale pixels from the cache, rather then the newly computed (by GPU) ones.

    Peter finally found the missing place where a cache flush was required which has fixed the lines.
    There is a known issue that causes some files to appear thus:
    Image
    The screenshot is from Big Buck Bunny 720p (bbb_720p_c.ts). Hopefully this issue will be fixed in tomorrow's build.
Build Details:
  1. OpenELEC:
    • project imx6: add sgtl5000 analog output alsa config file (PR:4177, 1 commit, 1 file changed)
  2. XBMC:
    • [WIN32][VFS][CURL]Possible fix for deadlock when stopping playback (PR:7167, 1 commit, 1 file changed)
    • [addon] Fix for trac 16016 (PR:7223, 1 commit, 1 file changed)
    • [binary addons] bump audio encoders (PR:7221, 1 commit, 4 files changed)
  3. newclock4:
    • New commits in this build:
      • [ffmpeg] squash: Option to parallelise coefficient decode and inter prediction and deblock for each frame and fix for cache flushing (9df94f24)
  4. kernel 4.0.y:
    • New commits in this build:
      • ARM: bcm2835: Set Serial number and Revision (fcf95f09)
      • bcm2835: Match BCM270X Device Trees (55106b1b)
      • dts: Enable overlays on ARCH_BCM2835 (83d89c2a)
      • bcm2835: Make camera and sound drivers available (7012d9ea)
      • bcm2835: Merge bcm2835_defconfig with bcmrpi_defconfig (d0f392c3)
      • Merge pull request #1001 from notro/closer (a9941291)
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.
(2015-06-03, 12:31)sunstealer Wrote: Firstly - thank you to Milhouse, popcornmix and Team Kodi very much for all your hard work on this. I have been installing the test builds for a while now with great success so far.

There is one problem I have encountered with these builds that I have not been able to figure out - it has been present for all the Kodi 15 builds that I have tried.

Setup: RPi model B -> HDMI -> Sony AV Amp -> Samsung TV with Logitech Harmony remote.

Problem: switching off TV & Amp at night with remote and leaving RPi / Kodi running. In the morning, switch on - RPi still responds to ssh and is 'alive' but TV says "no signal". Screensaver has been disabled. Kodi remote on ios device has active connection but pressing buttons does not wake up HDMI signal. Starting a movie from remote does wake up HDMI signal and then returning to GUI does work.

It seems to me that something is switching off the HDMI out and it is not waking up properly when using the remote - is there a power saving feature now?

I have an almost identical setup - the only difference is I have an RPi2.

If you select "Suspend" in System | Power saving then these builds will turn off the HDMI when you invoke the power off (e.g. power toggle on your remote). For me the HDMI is reenabled when I press remote buttons - I use the Stop button. I'm using a USB MCE R6 remote, so this might be dependent on the remote you have.

So in my Logitech setup I've configured the power options for my MCE remote to send:

Power on: Stop
Power off: Stop, Power toggle

This means that when powering off, the Stop will either deactivate the screensaver, stop whatever is playing, or nothing. And then the Power Toggle disables HDMI. Actually, I should probably have 2 Stops ...

For powering on, the Stop will re-enable the HDMI.

If Stop (or another button) doesn't re-enable HDMI, then just configure your device in the Harmony software so it's left on all the time.
I've noticed recently that when I skip forward through a video or TV recording, the OSD no longer opens. Before, the OSD would open so you had some visual feedback on how far you would skip.

Does anyone have any idea when this changed? I use the Titan skin, in case it is a skin issue.
(2015-06-04, 21:45)mcelliott Wrote: Does anyone have any idea when this changed? I use the Titan skin, in case it is a skin issue.

Test with Confluence.
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 OpenELEC Isengard build #0604: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.0.4 #1 Thu Jun 4 21:04:17 BST 2015 armv6l GNU/Linux

# vcgencmd version
May 31 2015 14:24:30
Copyright (c) 2012 Broadcom
version b127f62bd972cebb74a816b01b90ecfacc226baa (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150604210329-#0604-g3d5b644 [Build #0604]

# vcdbg log msg 2>&1 | grep DTOK
001571.666: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (3d5b6446, changelog) and tip of XBMC master (81a115bb, changelog) with the following modifications: Build Highlights:
  1. Fix for tiled hevc videos (Big Buck Bunny etc.)
Build Details:
  1. OpenELEC:
    • projects/imx6/patches/linux: update udoo poweroff driver (PR:4181, 1 commit, 3 files changed)
    • projects/WeTek_Play/linux: Disable CONFIG_FB_AMLOGIC_UMP in kernel config (PR:4178, 1 commit, 1 file changed)
  2. XBMC:
    • [gui] render dialog teletext always on top (PR:7207, 1 commit, 1 file changed)
    • [gui] Improve overloads in CGUIDialogYesNo (PR:7227, 2 commits, 27 files changed)
  3. libnfs:
    • READDIRPLUS: use lookup on entries where the server did not return any attr (d8ba04d9)
    • Add missing NULLing of data->continue_data for READDIR (68b02981)
  4. newclock4:
    • New commits in this build:
      • [rpi] Always add desktop resolution to supported list (b0fe3096)
      • [ffmpeg] squash: Support tiled decode (0d2c2fa9)
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.
Nice job guys with the recent HEVC optimisations and bug fixes.
All video Artifacts / horizontal streaks are now gone in my 720p HEVC testing with build #604.

I think a lot of RPi2 owners are going to be happy even with low bitrate 720p HEVC playback due to the improved video quality with this codec.

(2015-06-04, 21:56)Milhouse Wrote:
(2015-06-04, 21:45)mcelliott Wrote: Does anyone have any idea when this changed? I use the Titan skin, in case it is a skin issue.

Test with Confluence.

Should have done that first Wink Skin problem. My bad!
I second that. Too glad to see this.

(2015-06-05, 07:33)wrxtasy Wrote: Nice job guys with the recent HEVC optimisations and bug fixes.
All video Artifacts / horizontal streaks are now gone in my 720p HEVC testing with build #604.

I think a lot of RPi2 owners are going to be happy even with low bitrate 720p HEVC playback due to the improved video quality with this codec.
Hi,

I have problems with 3d videos since this commit from 15.03

https://github.com/popcornmix/xbmc/commi...c36a36df29

i normaly zoom videos to get rid of the black borders of widesceen videos. Since this commit the zooming doesn't work more as expected. The video is out of the center.

There is no problem when i watch the video unzoomed (with black borders)
(2015-06-05, 11:21)misterboo Wrote: I have problems with 3d videos since this commit from 15.03

https://github.com/popcornmix/xbmc/commi...c36a36df29

i normaly zoom videos to get rid of the black borders. Since this commit the zooming doesn't work more as expected. The video is out of the center.

There is no problem when i watch the video unzoomed (with black boarders)

Did this really work correctly before the commit? My understanding was any zooming affected the eye separation and spoilt the 3d effect.
Can you zoom arbitrarily far (with the older version) using the "zoom amount" in OSD and still keep the 3D effect?
(2015-06-05, 12:59)popcornmix Wrote: Did this really work correctly before the commit? My understanding was any zooming affected the eye separation and spoilt the 3d effect.
Can you zoom arbitrarily far (with the older version) using the "zoom amount" in OSD and still keep the 3D effect?

Yes zooming was always working perfectly, I normally use zoomfactor 1.3 for widescreen videos, but i could use any amount without affecting the 3d effect

It worked with the official Openelec 5.0.8 and all Testbuilds until the version from 15.03 ... the latest beta from openelec 5.95.1 either doesn't work anymore with zooming
I setup a couple of recordings last night. Before I did so, I checked that in the tvheadend browser interface that the recording folder was set to an existing folder, which it was (/var/media/USB/recordings). I thought this was a bit weird as I haven't got a USB stick connected at the moment but I checked with WinSCP and this folder did exist and when the first programme started recording, I saw the .ts file created in there.

However, when I checked later, the RPi had frozen on the EPG (debug overlay as well) and there were no recordings to be found anywhere and now when I check, there's no subfolders under /var/media. I've zipped up the log as it was rather big at 2MB https://drive.google.com/file/d/0B1fDI89...sp=sharing

I guess it's probably something I need to report to the tvheadend maintainers but thought I should report it here first as I'm using the testbuilds, just in case it was caused by something specific to those.

I've changed the recordings folder to /storage/recordings now but I'm not confident that this will prevent the problem reoccurring, as the previous folder did exist and I saw the recording file created there.
  • 1
  • 66
  • 67
  • 68(current)
  • 69
  • 70
  • 89

Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi (Kodi 15.0) Part 214