•   
  • 1
  • 56
  • 57
  • 58(current)
  • 59
  • 60
  • 156
  •   
  Thread Closed
OpenELEC Testbuilds for RaspberryPi Part 3 (Kodi 14.0)
I have now testet #0803 and #0803b (newclock4).

And I see some problemes, when I playback MPEG2/DVD/ISO on each chapter shift.
When the chapter change, the video is moving fast forward 1-2 sec.

I have allso seen thin on older Helix and Gotham builds, but only 1 or 2 times, on each movie. But on #0803b, the issue is constant on each chapter change/shift.

I use omxplayer:
<video>
<defaultplayer>omxplayer</defaultplayer>
<defaultdvdplayer>omxplayer</defaultdvdplayer>
</video>
read one post ahead - advancedsettings are not working in newclock4. But if you haven't changed the settings under Video -> Acceleration you'll be using omxplayer
New OpenELEC Helix build: #0804
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 3.15.8 #1 PREEMPT Mon Aug 4 20:47:19 BST 2014 armv6l GNU/Linux

# vcgencmd version
Aug  1 2014 18:05:07
Copyright (c) 2012 Broadcom
version 158db03216ab9f49243655d4d325da3264567104 (tainted) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20140804204605-r18886-gd6b472f

Based on tip of OpenELEC master (d6b472f, changelog) and tip of XBMC master (201b890, changelog) with the following modifications:
  • Includes newclock3 patches
  • Excludes the OpenELEC fernetmenta patches due to conflicts with newclock3
  • Excludes the OpenELEC linux-01-RPi_support patch in favour of sourcing these and possibly more recent patches directly from kernel branch rpi-3.15.y
  • Excludes the OpenELEC xbmc-001-newclock3 patch in favour of sourcing these and possibly more recent patches directly from newclock3 branch
  • Default setting for "Show RSS Feed" changed to disabled
  • Disabled "Total Duration" in Confluence (see build #0221 for details)
  • Includes latest libnfs master
  • Includes latest libcec master
  • Includes libcec double-key suppression.
  • Includes libcec CEC Standby Fix.
  • Increase scan interval of PeripBusCEC from 5000 to 60000, reducing CPU loading by about 2% (1GHz Pi) every 5 seconds (even when CEC is "disabled")
  • Includes CONFIG_COREDUMP=y to allow creation of coredumps
  • Includes PR4990: Allow larger font size
Build Highlights:
  1. newclock3-based build.
  2. I'll most likely create newclock4 builds every other day for the time being - so no newclock4 build for today!
Build Details:
  1. OpenELEC:
    • bluez: dont compile libbluetooth.so
    • swig:host: build without pcre
    • systemd: simplify userconfig.service
    • parted: update to parted-3.2
  2. XBMC:
    • videolibrary: by default sort movies inside a set by year and ascending (PR:5141, 1 commit, 1 file changed)
    • [AE/osxsink] - support multistream devices (PR:4768, 10 commits, 8 files changed)
    • Store subtitle settings only on user action - replacing #4981 (PR:5003, 6 commits, 7 files changed)
    • [builtins] don't activate an already active window (PR:5081, 1 commit, 1 file changed)
    • skin confluence: change channel group by clicking on channels or guide (PR:5076, 1 commit, 1 file changed)
    • Remove HAL (PR:5077, 1 commit, 14 files changed)
    • binary addons: add DEPENDS_TO_BUILD parameter to addon depends buildsystem (PR:5134, 2 commits, 4 files changed)
    • Use reentrant localtime if available (PR:5110, 1 commit, 5 files changed)
    • Store audio stream setting only on user action (PR:5131, 1 commit, 4 files changed)
    • fixed: broken GetKernelBitness()
  3. kernel 3.15.y:
    • Revert "fix soc-core's inverse range and let IQaudIO DAC use this fixed SOC_DOUBLE_R_RANGE_TLV support"
    • Revert "Fix volsw_range functions so SOC_DOUBLE_R_RANGE_TLV works."
    • Revert "Add IQaudIO Sound Card support for Raspberry Pi"
    • soc-core: Fix volsw_range funcs so SOC_DOUBLE_R_RANGE_TLV works.
    • pcm512x: Use a range macro for Volume and rename to PCM.
    • Merge pull request #654 from hmbedded/rpi-3.15.y
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.
@popcornmix - I just tried a 3D sample with #0803b (newclock4). While the video conversion (i.e. 3D -> mono) is working fine, it doesn't seem to do the resolution switch anymore and bring the TV into 3D mode. Is this patch simply missing or is it supposed to work but broken?
(2014-08-04, 22:21)da-anda Wrote: @popcornmix - I just tried a 3D sample with #0803b (newclock4). While the video conversion (i.e. 3D -> mono) is working fine, it doesn't seem to do the resolution switch anymore and bring the TV into 3D mode. Is this patch simply missing or is it supposed to work but broken?

Yes, that patch is missing. So far I've not tested 3D. I know the SBS->TAB modes won't work yet (needs a firmware update).
I'll do some 3D testing when I get a chance.
Loving newclock4, though three problems with it so far.

To be clear, all of these are with OMXPlayer disabled (I've always despised its habit of seeming to apply volume amplification, when I have a TV which does this far better), and all with live/recorded tv which is all I've had time to test it with so far.

1. Skipping forward and backwards is far far quicker with newclock4, but both result (Wmc.pvr/Mpeg2) in a grey screen for a moment. Not an issue, as the faster fastforward/rewind is greatly appreciated!

2. Subtitles appear broken (UK Freeview DBT). They never worked anyway on live tv, but now no longer work on recorded tv either. Subtitles are now always recognised as "Unknown language", and don't actually display.

3. Not an issue at all, because browsing the library now works perfectly and, from quickly reading previous replies this seems to be expected with DVDPlayer, but when playing media there's now a slight pause when bringing up the OSD - and, more often than not - a brief glitch on the video stream. Very liveable with, though.

Overall, the improvements in newclock4, even when causing the grey screen, are such that this is a great improvement over the already suberb previous, and thankyou to both popcornmix and Milhouse for a great upgrade to the system.
(2014-08-04, 21:59)da-anda Wrote: read one post ahead - advancedsettings are not working in newclock4. But if you haven't changed the settings under Video -> Acceleration you'll be using omxplayer

Yes, I use omxplayer Acceleration, when I see the Fast-forward behavior on each chapter change, on playback of MPEG2/DVD/ISO's.
This behavior are only sporadic on the other Milehouse builds.
(2014-08-04, 23:38)Unfledged Wrote: 2. Subtitles appear broken (UK Freeview DBT). They never worked anyway on live tv, but now no longer work on recorded tv either. Subtitles are now always recognised as "Unknown language", and don't actually display.

I always had subtitle issues on some livetv channels with omxplayer. Try toggle deinterlacing, on, off, auto it will bring subs back for me on livetv
Ok, these are my results with newclock4 build:

1 - with acceleration disabled, the cpu usage is close to omxplayer;

2 - I can play big_buck_bunny 1080p movie through Nfs, but only if '<buffermode>' has been set to '1'. Otherwise it does stuttering. Anyhow at startup there are 14-15 dropped frames.

3 - Some videos don't play correctly.

4 - I've noticed some trouble with omxplayer enabled: during playback the screen turns black for a few seconds, then returns to normal but with some graphical artifact. This seems to occur randomly.

Some infos about my settings:

Code:
# cat /proc/cmdline
root=/dev/ram0 rdinit=/init BOOT_IMAGE=/kernel.img dma.dmachans=0x7f35 bcm2708_fb.fbwidth=1824 bcm2708_fb.fbheight=984 bcm2708.boardrev=0x100000e bcm2708.serial=0xabbce687 smsc95xx.macaddr=*:*:*:BC:E6:87 sdhci-bcm2708.emmc_clock_freq=250000000 vc_mem.mem_base=0x1ec00000 vc_mem.mem_size=0x20000000  boot=LABEL=SYSTEM disk=LABEL=Storage quiet ssh smsc95xx.turbo_mode=N

Code:
# vcgencmd get_config int
arm_freq=850
core_freq=375
over_voltage=2
hdmi_force_hotplug=1
program_serial_random=1
config_hdmi_boost=4
emmc_pll_core=1
hdmi_ignore_cec_init=1
hdmi_force_cec_address=65535
temp_limit=70
force_pwm_open=1
pause_burst_frames=1
avoid_fix_ts=1

Code:
gpu_mem_512=160

P.s.
As in my previous posts, do not hate me for my English ^^
Got some weird boot-issues with the last handful of releases. I've set it up once, in my setup, boot from SD and storage on an usb-thumb. so far so good, freshly installed and configured - working. watched some series the other evening and did shut down the device in the night. next day, booting up the rpi does not work (led stays simply red). switched the hardware to another pi of mine - same thing. re-created the sdcard (deleted all partitions, copied the relevant files ot it) - WORKING! ... eh?! ok then.. back working. watched some tv-series and did shut down the device at night. next day - SAME THING! pi not booting up. i then cloned the SD to another one (with minitool partiton master) - and? WORKING! ... eh?!?!?!?

any1 got a clue?

got the model b with 512mb, a 2A powerplug, cpu @900, core @400, [email protected], class 10 uhs1 transcend 8gb sd with an usb 3.0 transcend 8gb thumb drive.
You posted in the Raspbmc thread too, are you getting this with both OE and Raspbmc?

With OE (maybe Raspbmc too) you don't need to reformat the entire SD card if the Pi isn't booting, just try restoring the boot files and only reformat the FAT partition if the filesystem is hosed (which happens hardly ever, but is possible).

Sometimes it can be a physical connection issue due to a broken SD card. Check the card hasn't cracked and is now flexing, resulting in intermittent connection.

If the card appears to be physically OK, try using a different card and if the problem continues try to identify the build where the problem starts.
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.
Thanks for your reply - I really think that it could be the sd-card by now. The sd-card i far away from being corrupted, all files are ok, non of em are missing.
So it could really be a connection issue. I've got the same problem with bot, OE/RaspBMC, yes. I initially used OE and then switched to RaspBMC, because I thought it would be an OE issue.

But one more question then, regardings both systems - I noticed, that RaspBMC in its current version seems noticably faster with the same cpu/gpu/core/ram OC settings and the same skin, etc. How could that be?
(2014-08-05, 12:34)FattyMcDirty Wrote: But one more question then, regardings both systems - I noticed, that RaspBMC in its current version seems noticably faster with the same cpu/gpu/core/ram OC settings and the same skin, etc. How could that be?

Pass... it would be useful if you could quantify that somehow, such as "Raspbmc is 2 seconds faster than OE when doing XYZ" (where XYZ might be opening a library of 200 movies). Presumably this is testing with the same addons, library, filesystems and network mounting settings (NFS rsize/wsize etc.). Maybe monitoring CPU behaviour during various activities would be useful to ensure the CPU frequency is scaling correctly (bcmstat.sh will confirm this).
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.
just a minor thing that I noticed. I think the the amplitude of the waveform viszalization is way to high/extreme on the PI compared to my Windows setup. Looks like it's something like 200% (curve exceeds canvas massively instead of having it's peak somewhere near the screen borders)
(2014-08-05, 13:53)da-anda Wrote: just a minor thing that I noticed. I think the the amplitude of the waveform viszalization is way to high/extreme on the PI compared to my Windows setup. Looks like it's something like 200% (curve exceeds canvas massively instead of having it's peak somewhere near the screen borders)

Can you compare it to another GL ES platform (e.g. ATV2 / Android?)
I doubt Pi behaves differently to them.

It would be good to unify the different platforms though (especially if one looks better).
  •   
  • 1
  • 56
  • 57
  • 58(current)
  • 59
  • 60
  • 156
  •   
  Thread Closed
 
Thread Rating:
  • 8 Vote(s) - 4.88 Average



Logout Mark Read Team Forum Stats Members Help
OpenELEC Testbuilds for RaspberryPi Part 3 (Kodi 14.0)4.888