• 1
  • 63
  • 64
  • 65(current)
  • 66
  • 67
  • 168
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)
(2015-08-31, 18:14)popcornmix Wrote:
(2015-08-31, 17:44)polo_joe Wrote: I tried both settings together and each alone.
No change.

mmal half works.

Do you know anyone else with a pi2?
I would be interested if your sdcard, power supply, TV and peripherals doesn't have these artefacts with a different pi.

unfortunately I don't know anyone else.
(2015-08-31, 17:44)polo_joe Wrote: I tried both settings together and each alone.
No change.

mmal half works.

mmal advanced at HD triggers a boost to 300MHz (of v3d clock).
The half version doesn't need the boost, nor does automatic (which uses bob for hd).

So my guess is that 300MHz v3d is not working on your Pi.
However with default config.txt settings (i.e. no v3d_freq or gpu_freq specified) and with disable_auto_turbo=1 you should still be running at 250MHz.

When running with default config.txt settings + disable_auto_turbo=1 and when getting the artefacts can you run:
Code:
vcgencmd measure_clock v3d
If it is zero then run a few times until you see a non-zero value and report what it is.
(2015-08-31, 18:42)popcornmix Wrote:
(2015-08-31, 17:44)polo_joe Wrote: I tried both settings together and each alone.
No change.

mmal half works.

mmal advanced at HD triggers a boost to 300MHz (of v3d clock).
The half version doesn't need the boost, nor does automatic (which uses bob for hd).

So my guess is that 300MHz v3d is not working on your Pi.
However with default config.txt settings (i.e. no v3d_freq or gpu_freq specified) and with disable_auto_turbo=1 you should still be running at 250MHz.

When running with default config.txt settings + disable_auto_turbo=1 and when getting the artefacts can you run:
Code:
vcgencmd measure_clock v3d
If it is zero then run a few times until you see a non-zero value and report what it is.

frequency(43)=250000000

config.txt http://sprunge.us/ifBC
(2015-08-31, 19:46)polo_joe Wrote: frequency(43)=250000000

config.txt http://sprunge.us/ifBC

And with this configuration you get artefacts with "MMAL Advanced" and no artefacts with "MMAL Advanced (half)", "Bob" or "Automatic"?
(2015-08-31, 19:52)popcornmix Wrote:
(2015-08-31, 19:46)polo_joe Wrote: frequency(43)=250000000

config.txt http://sprunge.us/ifBC

And with this configuration you get artefacts with "MMAL Advanced" and no artefacts with "MMAL Advanced (half)", "Bob" or "Automatic"?

Yes artefacts with MMAL advanced, no artefacts with mmal half, automatic or bob on 1080i50 livetv channel.
(2015-08-31, 20:04)polo_joe Wrote: Yes artefacts with MMAL advanced, no artefacts with mmal half, automatic or bob on 1080i50 livetv channel.

Same config settings as before, but add "force_turbo=1". Now I'm interested if "MMAL advanced (half)" has artefacts.
(2015-08-31, 20:06)popcornmix Wrote:
(2015-08-31, 20:04)polo_joe Wrote: Yes artefacts with MMAL advanced, no artefacts with mmal half, automatic or bob on 1080i50 livetv channel.

Same config settings as before, but add "force_turbo=1". Now I'm interested if "MMAL advanced (half)" has artefacts.

mmal half = no artefacts
mmal advanced = artefacts

config.txt http://sprunge.us/dQNP
(2015-08-31, 20:14)polo_joe Wrote: mmal half = no artefacts
mmal advanced = artefacts

config.txt http://sprunge.us/dQNP

Okay, default settings plus:
Code:
force_turbo=0
disable_auto_turbo=1
arm_freq=600
sdram_freq=400
(2015-08-31, 20:17)popcornmix Wrote: Okay, default settings plus:
Code:
force_turbo=0
disable_auto_turbo=1
arm_freq=600
sdram_freq=400

mmal half = artefacts
mmal advanced = artefacts

config.txt http://sprunge.us/WZhP
(2015-08-31, 20:25)polo_joe Wrote: mmal half = artefacts
mmal advanced = artefacts

config.txt http://sprunge.us/WZhP

I was expecting no artefacts with those settings. It does feel to me like a faulty Pi.
Can you remove non-default config.txt settings and try the shadertoy addon:
http://forum.kodi.tv/showthread.php?tid=...pid2017239

Enable the shadertoy screensaver and try a number of them.

Also can you try a HEVC file. E.g. http://www.elecard.com/assets/files/othe..._720p_c.ts

If either of these fail, then it would suggest a problem with the 3D hardware on the Pi (which is used by both of these tests).
(2015-08-31, 20:39)popcornmix Wrote: If either of these fail, then it would suggest a problem with the 3D hardware on the Pi (which is used by both of these tests).

screensaver and video are working perfectly fine.

config.txt http://sprunge.us/FRBR
I'm having problems seeking / with 4:3 to 16:9 transitions on RPi with the latest build (New OpenELEC Jarvis build #0830: RPi) using both the OMXPlayer and MMAL acceleration options. This seems to affect channels when they use different formats/bitrates for ads and shows.

I've uploaded a short extract here ( http://www.thomasinfletcher.com/temp/xaa.mpg ) which exhibits the problem.
The file is a .ts stream recorded using a WinFast DTV1000-T and mythtv from the 'Yesterday' channel in the UK.

It starts with adverts in 16:9 format, then switches to a program in 4:3 format.

Symptoms are as follows:
  1. If you play the file through without seeking, OMXPlayer manages the transition fine (shape changes, continues playing).
  2. If you play the file through without seeking using MMAL, the screen shape doesn't change (stays at 16:9).
  3. If you seek across the 16:9 to 4:3 transition with OMXPlayer, the shape changes correctly, but the image freezes. Selecting Pause/Play doesn't restart playback. When you press 'stop' it seems to start playing for a fraction of a second.
  4. If you seek across the 16:9 to 4:3 transition with MMAL, the file continues to play, but the shape stays at the old one (i.e. image is squashed/stretched)
  5. If you 'resume' from after the transition (not easily possible with this shortened clip), it plays correctly using both OMXPlayer and MAML in the correct shape.
  6. with acceleration disabled, playing across the transition is handled correctly, as is skipping across the transition (although the RPi can't render every frame and audio is out of sync).
  7. mplayer on my x86 gentoo box seems to handle the 16:9 4:3 shape transition/skipping across it correctly.

Similar issues happen later in the full recording at the start of the next add break when a 4:3 to 16:9 transition occurs.
New OpenELEC Jarvis build #0831: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.1.6 #1 Mon Aug 31 21:04:18 BST 2015 armv6l GNU/Linux

# vcgencmd version
Aug 23 2015 19:27:51
Copyright (c) 2012 Broadcom
version 41b1f861989dc4d1bbe19a52361a248ae57ee82b (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20150831210327-#0831-g494f51d [Build #0831]

# vcdbg log msg 2>&1 | grep DTOK
001573.433: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (494f51df, changelog) and tip of XBMC master (23bc0c1a, changelog) with the following modifications: Build Highlights:
  1. Several PVR fixes
  2. Revert "[PVR] API 3.0.0 Version Bump"
Build Details:
  1. XBMC:
    • [PVR] Timer settings dialog: Start/End display fix (PR:7875, 2 commits, 1 file changed)
    • [PVR][fix] Don't show 'will record' in EPG for inactive schedule (PR:7923, 1 commit, 1 file changed)
    • [PVR] Search context menu crash fix + 'play recording' context option (PR:7929, 2 commits, 2 files changed)
    • [guilib] fix button label2 not updated after emptying (PR:7927, 1 commit, 1 file changed)
    • fixed: Enable retry without range set for e.g. broken HDHomerun servers (fixes #14204) (PR:7924, 1 commit, 1 file changed)
    • [PVR] Fix: Change parameter to CPVRChannelPtr (PR:7933, 1 commit, 2 files changed)
  2. pvr.argustv:
    • Fix: Skip channel re-tuning if UID of new channel equal to the UID of already streaming channel (PR:30, 1 commit, 1 file changed)
  3. newclock4:
    • New commits in this build:
      • Revert "[PVR] API 3.0.0 Version Bump" (13e203f7)
      • [peripheral] Move initialisation after language init to fix notifcations (b4c36a38)
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.
Hey, I'm a few days behind, this is a crashlog from #0829 http://sprunge.us/GOAN
Came across a crash caused by scp'ing a 2GB file to /storage/videos. It starts bootlooping after awhile, although the scp continues to resume after each loop for awhile and the transfer continues and completes successfully.

Afterwards I had to force a reboot to get out of the bootloop, even after the transfer was complete.

I assume from the log this probably wasnt fixed in 830,831
Not seen that before, I'll upload a debug-enabled build in case you're able to reproduce as unfortunately the crash-log isn't providing much information other than a suggestion that something went wrong while servicing a webserver request.
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.
  • 1
  • 63
  • 64
  • 65(current)
  • 66
  • 67
  • 168

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