• 1
  • 52
  • 53
  • 54(current)
  • 55
  • 56
  • 168
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)
New OpenELEC Jarvis build #0820: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.1.6 #1 Thu Aug 20 21:03:18 BST 2015 armv6l GNU/Linux

# vcgencmd version
Aug 20 2015 17:56:26
Copyright (c) 2012 Broadcom
version 3f87086448257aa66881de858a0b4b0e6daa8719 (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150820210229-#0820-g41b0ef9 [Build #0820]

# vcdbg log msg 2>&1 | grep DTOK
001575.730: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (41b0ef92, changelog) and tip of XBMC master (37a742a1, changelog) with the following modifications: Build Highlights:
  1. New firmware - deinterlace has been optimised a little bit more which may (or may not) help with deinterlacing and audio
  2. PVR API 3.0.0 bump - all clients bumped
  3. Re-include Python distutils
  4. Switch to in-kernel MediaTek mt7601u WiFi driver
  5. Eliminate 60 second delay at startup when network not available
Build Details:
  1. Firmware (Aug 20):
    • firmware: hdmi: disable dhcp completely to avoid registered i2c user See: link
    • firmware: threadx: Avoid calling a NULL interrupt handler. See: link
    • firmware: arm_loader: Load standard touchscreen overlay
    • firmware: di_adv: Remove dma and copy non interlaced lines from shader
    • firmware: di_adv: Add vector code for copying top/bottom lines
    • firmware: di_adv: Do not deinterlace first frame and have one less frame of latency
  2. OpenELEC:
    • Use a temporary filename when creating new crashlog (PR:4276, 1 commit, 1 file changed)
    • Revert "Python: dont install distutils" (c9f693af)
    • linux: add backport of mediatek mt7601u driver from kernel 4.2, stolen from https://github.com/raspberrypi/linux , thanks much (ce9105f7)
    • remove package 'mt7601u', added as kernel patch (41b0ef92)
  3. XBMC:
    • [pvr][gui] reduce amount of dialogs for PVR (PR:7744, 2 commits, 5 files changed)
    • CLangInfo: properly set the time format when setting/changing regions (PR:7842, 1 commit, 1 file changed)
    • [guilib] fix double memory free in panel container (PR:7849, 1 commit, 1 file changed)
    • [crossguid] all: also check for installed location (PR:7854, 2 commits, 3 files changed)
    • [addonbrowser] dont set content to "addons" for category listings and other menus (PR:7817, 1 commit, 2 files changed)
    • [lang] fix inconsistent naming of norwegian languages (PR:7858, 1 commit, 1 file changed)
    • [lang] sort special values for preferred subtitle/audio language on top (PR:7856, 1 commit, 1 file changed)
    • [PVR] API 3.0.0 Changes (PR:7626, 11 commits, 14 files changed)
    • [Confluence] remove dead code (37a742a1)
  4. pvr.argustv:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:25, 1 commit, 2 files changed)
  5. pvr.demo:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:19, 1 commit, 1 file changed)
  6. pvr.dvblink:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:25, 1 commit, 2 files changed)
  7. pvr.dvbviewer:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:16, 1 commit, 2 files changed)
  8. pvr.filmon:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:20, 1 commit, 2 files changed)
  9. pvr.hts:
    • Minimal support for PVR API 3.0.0 (PR:83, 2 commits, 4 files changed)
  10. pvr.iptvsimple:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:35, 1 commit, 2 files changed)
  11. pvr.mediaportal.tvserver:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:21, 1 commit, 2 files changed)
  12. pvr.mythtv:
    • Minimal support for PVR API 3.0.0 (PR:26, 2 commits, 3 files changed)
  13. pvr.nextpvr:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:22, 1 commit, 2 files changed)
  14. pvr.njoy:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:15, 1 commit, 2 files changed)
  15. pvr.pctv:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:19, 1 commit, 2 files changed)
  16. pvr.stalker:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:18, 1 commit, 2 files changed)
  17. pvr.vbox:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:94, 1 commit, 2 files changed)
  18. pvr.vdr.vnsi:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:24, 1 commit, 1 file changed)
  19. pvr.vuplus:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:17, 1 commit, 2 files changed)
  20. pvr.wmc:
    • Bump add-on for PVR API 3.0.0 (Compatibility mode) (PR:19, 1 commit, 3 files changed)
  21. Additional commits/pull requests/changes not yet merged upstream:
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 have more core temperature on RPi2 after #0816 test builds. It is normal?
#0815 build is 52 celcious but after builds are 56 celcious.

It's not problem, but make curiosity Huh
Nothing obvious in #0816, could it be add-on related? Do you still see higher temperatures when testing a clean .kodi install? Is your CPU load similar in the two builds?
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.
In this sample file subtitles are shown on wrong side of the screen (left in the middle and not at the bottom) (you have to enable it if they are not shown). I know that in older versions it was playing fine but can not say when.
Using omx player.
File (size 13mb): http://bite-in.com/debug/24Kitchen-HD.ts
(2015-08-21, 02:30)Milhouse Wrote: Nothing obvious in #0816, could it be add-on related? Do you still see higher temperatures when testing a clean .kodi install? Is your CPU load similar in the two builds?

Yep, it may be addon related.
And new #0820 test build stalled online streaming video(RT news) after 4+ hours. I 'm back to #0815 again.
(2015-08-20, 21:47)Milhouse Wrote:
Code:
<FullscreenVideo>
    <remote>
      <select>AspectRatio</select>
    </remote>
  </FullscreenVideo>

would configure the Select (OK) button to toggle through the different aspect ratios.

Thanks, that will make life easier.

Quote:Also, have you tried configuring a value for Video > Playback > "Display 4:3 videos as" - try something like "Stretch 16:9" and all 4:3 videos should automatically use the selected aspect ratio.

That seems to need to be on Normal. With both Normal or Stretch 16:9 the Zoom is 1:00 but for the latter it changes the Pixel Ratio to 0.92.

I compared two programmes, The Simpsons on Ch.4 and Alias Smith and Jones on ITV4 on my TV (both SD) which are both broadcast in 4:3 (so my TV auto-switches to 4:3 mode and displays grey bars either side).

On the RPi, The Simpsons on Ch.4 sets the View Mode to Normal (as per the Display 4:3 videos setting) but Alias Smith and Jones on ITV4 sets it to Stretch 16:9 (which is the saved default view). I have no idea why it doesn't detect the latter as 4:3.. Anyway, I still need to change the TV mode from Auto (which sets it to Just with the RPi) to 4:3 for 4:3 programmes and back to Auto for 16:9 programmes.

With my TV in 4:3 mode and the RPi in Normal view mode, when the 4:3 programme switches to 16:9 adverts, not only do I have the grey bars either side but also black space above and below the adverts, so they're displayed in a fairly narrow strip in the middle. Switching the RPi view mode to Stretch 16:9 expands the advert vertically to remove the black space and then I have to switch the TV mode to Auto to expand it horizontally to remove the grey bars.

Hey, it's only adverts though, so not worth worrying about. It's just a shame I have to manually switch the TV mode and in some cases the RPi view mode, depending on whether the actual programme is 16:9 or 4:3. I think the only way round this would be if the RPi could automatically detect 4:3 programmes and create a 16:9 frame with black or grey bars either side of the 4:3 picture, which I guess is what Freeview STB do.
(2015-08-21, 13:58)doveman2 Wrote: With my TV in 4:3 mode and the RPi in Normal view mode, when the 4:3 programme switches to 16:9 adverts, not only do I have the grey bars either side but also black space above and below the adverts, so they're displayed in a fairly narrow strip in the middle. Switching the RPi view mode to Stretch 16:9 expands the advert vertically to remove the black space and then I have to switch the TV mode to Auto to expand it horizontally to remove the grey bars.

Hey, it's only adverts though, so not worth worrying about. It's just a shame I have to manually switch the TV mode and in some cases the RPi view mode, depending on whether the actual programme is 16:9 or 4:3. I think the only way round this would be if the RPi could automatically detect 4:3 programmes and create a 16:9 frame with black or grey bars either side of the 4:3 picture, which I guess is what Freeview STB do.

Can you test the behaviour with Kodi on a PC and see if it's the same (I imagine it is). If it is, then it's best to make your suggestion in a general or feature request forum where the right devs may be able to comment.
(2015-08-21, 14:01)popcornmix Wrote: Can you test the behaviour with Kodi on a PC and see if it's the same (I imagine it is). If it is, then it's best to make your suggestion in a general or feature request forum where the right devs may be able to comment.

I think I can connect my USB tuner to the PC and presumably run tvheadend, so I should be able to. Obviously it will be connected to the TV via HDMI not composite though, so I'm not sure it will be a meaningful comparison, as Kodi may well be able to make the TV change mode via HDMI. In fact, I should test the RPi via HDMI first.
(2015-08-21, 14:05)doveman2 Wrote:
(2015-08-21, 14:01)popcornmix Wrote: Can you test the behaviour with Kodi on a PC and see if it's the same (I imagine it is). If it is, then it's best to make your suggestion in a general or feature request forum where the right devs may be able to comment.

I think I can connect my USB tuner to the PC and presumably run tvheadend, so I should be able to. Obviously it will be connected to the TV via HDMI not composite though, so I'm not sure it will be a meaningful comparison, as Kodi may well be able to make the TV change mode via HDMI. In fact, I should test the RPi via HDMI first.

CompositHuh I'm not aware that there is any signaling of 16/9 on a composit output, so HDMI will be the solution to go for the right aspect and a clear picture :-). I have never seen any issues on HDMI with wrong AR if it is implemented correctly on the broadcasters side.

EDIT: there is WSS signaling in PAL signals, the question is if it is implemented in the Firmware: https://en.wikipedia.org/wiki/Widescreen_signaling
With both #819 and #820, when playing a TV recording (MPEG2), even with deinterlace disabled, it locks up when skipping backwards (I think it did skipping forwards when testing with #819 yesterday as well) and I have to stop playback. I've only quickly tested with MMAL with #820 but with #819 I tried with OMXPlayer enabled and that wasn't any better, the only thing that fixed it was disabling both OMX and MMAL so that it wasn't using acceleration at all.

http://pastebin.com/dGGimucC
(2015-08-21, 14:22)doveman2 Wrote: With both #819 and #820, when playing a TV recording (MPEG2), even with deinterlace disabled, it locks up when skipping backwards (I think it did skipping forwards when testing with #819 yesterday as well) and I have to stop playback. I've only quickly tested with MMAL with #820 but with #819 I tried with OMXPlayer enabled and that wasn't any better, the only thing that fixed it was disabling both OMX and MMAL so that it wasn't using acceleration at all.

So the recorded .ts shows this behaviour when played through file interface (not PVR interface)?
If so can you provide a sample file (e.g. through google drive or dropbox).
(2015-08-21, 14:18)L-S-D Wrote: CompositHuh I'm not aware that there is any signaling of 16/9 on a composit output, so HDMI will be the solution to go for the right aspect and a clear picture :-). I have never seen any issues on HDMI with wrong AR if it is implemented correctly on the broadcasters side.

EDIT: there is WSS signaling in PAL signals, the question is if it is implemented in the Firmware: https://en.wikipedia.org/wiki/Widescreen_signaling

I'm only setting this up with composite as that's the only option my brother has on his TV. I use HDMI myself normally.

The fact that OE correctly detects one 4:3 programme and sets the view mode to Normal but leaves another on Stretch 16:9 is a separate issue of course, regardless of composite/HDMI, although to be fair the distortion with it set to a pixel ratio of 0.92 (with Stretch 16:9) instead of 1.00 (with Normal) probably isn't that obvious to most people.
(2015-08-21, 14:24)popcornmix Wrote: So the recorded .ts shows this behaviour when played through file interface (not PVR interface)?
If so can you provide a sample file (e.g. through google drive or dropbox).

Yep, just playing via Video files. I wasn't even aware there was a PVR interface for playing back recordings.

What's the best way to cut a sample out of a .ts file?
(2015-08-21, 14:28)doveman2 Wrote:
(2015-08-21, 14:18)L-S-D Wrote: CompositHuh I'm not aware that there is any signaling of 16/9 on a composit output, so HDMI will be the solution to go for the right aspect and a clear picture :-). I have never seen any issues on HDMI with wrong AR if it is implemented correctly on the broadcasters side.

EDIT: there is WSS signaling in PAL signals, the question is if it is implemented in the Firmware: https://en.wikipedia.org/wiki/Widescreen_signaling

I'm only setting this up with composite as that's the only option my brother has on his TV. I use HDMI myself normally.

The fact that OE correctly detects one 4:3 programme and sets the view mode to Normal but leaves another on Stretch 16:9 is a separate issue of course, regardless of composite/HDMI, although to be fair the distortion with it set to a pixel ratio of 0.92 (with Stretch 16:9) instead of 1.00 (with Normal) probably isn't that obvious to most people.

well I'm quite sure I would see it. I'm not having any issues with tvh as backend and aspect ratio so far. Maybe you could provide 30sec of the whole ts where you are having this problem and open a ticket on tvh issue tracker?
(2015-08-21, 14:30)doveman2 Wrote: What's the best way to cut a sample out of a .ts file?

Simplest is to record a new shorter file and then check that it has the same issue.
Otherwise ffmpeg can be used for cutting down a file.
  • 1
  • 52
  • 53
  • 54(current)
  • 55
  • 56
  • 168

Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)10