• 1
  • 131
  • 132
  • 133(current)
  • 134
  • 135
  • 168
OpenELEC Testbuilds for RaspberryPi (Kodi 16.0)
(2015-11-07, 15:25)Caramba69 Wrote:
(2015-11-07, 12:35)mule1 Wrote:
(2015-11-05, 21:24)mule1 Wrote: Is it usual that navigating (using skips) in a 3D movie using the latest testbuilds (started testing from build #1029) is yet nearly impossible?
In rare cases after a skip the movie continues to play. But most times Kodi only fills the buffer (shown in the GUI) but never continues to play the movie or the movie continues to play but only with 2-4 frames per second or if it continues there are heavy stutters. Stopping and trying to resume shows the same results. ISO seems to be even worse than MKV files. Navigating in non 3d movies is working as it should.

No one else having these issues? Is navigating (skipping forward, backward, fast forward, rewind) in 3d movies working flawless for you?

I will check this tomorrow as well. Perhaps also somebody else can check?

Skipping works for me.
(2015-11-07, 13:53)popcornmix Wrote:
(2015-11-04, 00:07)popcornmix Wrote: The Advanced deinterlace at 1080i60 is expensive. It generates a lot of sdram traffic.
DVB USB dongles generate a lot of data, and depending on the driver, it may need to read prompty to avoid losing data.

The suspicion is that the heavy sdram traffic delays the USB driver occasionally and it loses a packet, causing corruption in the video.

Now there may be ways of working around this. Forcing the deinterlace to spread out its sdram accesses the right amount may solve this.
E.g. currently deinterlace might take 13ms every 16ms. If I were to make the deinterlace pause for 10us every 100us, it would still likely get it's job done,
and hopefully the USB driver won't be held up during the pauses.



Unfortunately it's not something I can test (I have encrypted cable at home, and too weak a terrestrial signal at both work and home).
I may be able to produce a test build that attempts this (perhaps with a config setting to tweak how frequent/long the pauses are).

@bagofcrap24 @polo_joe

I've added a test setting to see if we can fix the artefacts with advanced 1080i deinterlave and locally attached DVB dongle.
Use latest Milhouse build and add:
Code:
hvs_priority=0xN0000000

Replace the N with a hex digit from 0 to f (i.e. 0,1,2,3,4,5,6,7,8,9,a,b,c,d,e,f). That number determines how long to pause for, with 0 meaning no change in behaviour, to f means lots of pausing.
Start with 8, and if you still have artefacts then increase it, and if no artefacts then decrease it. Try to find the minimum setting that doesn't have artefacts.
Too high a setting will cause lagging of the video, as the deinterlace will take longer. Hopefully there will be a value that fixes artefacts and doesn't have lagging.

Tried the settings from 8 to f but unfortunately no change.
(2015-11-07, 15:25)Caramba69 Wrote:
(2015-11-07, 12:35)mule1 Wrote:
(2015-11-05, 21:24)mule1 Wrote: Is it usual that navigating (using skips) in a 3D movie using the latest testbuilds (started testing from build #1029) is yet nearly impossible?
In rare cases after a skip the movie continues to play. But most times Kodi only fills the buffer (shown in the GUI) but never continues to play the movie or the movie continues to play but only with 2-4 frames per second or if it continues there are heavy stutters. Stopping and trying to resume shows the same results. ISO seems to be even worse than MKV files. Navigating in non 3d movies is working as it should.

No one else having these issues? Is navigating (skipping forward, backward, fast forward, rewind) in 3d movies working flawless for you?

I will check this tomorrow as well. Perhaps also somebody else can check?

I'm having this same issue. I don't remember it happening with the old 15 based builds... after updating to the 16 build I now have issues skipping... sometimes it works, sometimes doesn't, but there's always stuttering, and there's also flashing when the OSD is brought up.
In 3D MVC MKV files i even noticed strange behaviour after navigate in the file.

Yesterday Evening i have tried to play a 3D MKV Movie from a moment i have stopped it one day before. I ve got no sound and only ultra slow slowmotion in the video. After that i restarts Kodi, shutdown my avr and gave it a new try. But the same behaviour as without restarting the engines. Another try with settings audio and video to standard. Even this shows no effort on the movie, no sound and ultra slow slowmotion.

My workaround was to try openelec 6.0 final. And now the movie plays.

My hope ist Milhouse and other kodi developers find a solution for the coming Jarvis Beta Versions.

Kind Regards

Gerald
Oh so openelec 6.0 final now includes mvc support? I didn't know that, I'll have to try that, thanks.
(2015-11-09, 18:29)tate1293 Wrote: Oh so openelec 6.0 final now includes mvc support? I didn't know that, I'll have to try that, thanks.

Yes, mvc in mkv is supported in OE 6.0.
mvc in iso is still experimental so only in these builds.
(2015-11-09, 10:07)polo_joe Wrote: Tried the settings from 8 to f but unfortunately no change.

You can check the setting has applied from log or by running:
Code:
vcgencmd get_config hvs_priority

bagofcrap24 did see an effect from this setting (but not a solution).

Can you both try:
Code:
hvs_priority=0x0N000000

This will lower the priority of the 3d hardware doing the deinterlace.
Using 1 should be minimum priority, so I'd expect no artefacts, but you may have lagging.
The default is "e" (which you also get with zero) which is very high priority. I'd be interested if that setting has the expected effect.
If artefacts are fixed, then finding the highest value without artefacts would be useful.

If you have lagging, that can be improved by increasing v3d_clock. e.g. v3d_clock=400
(2015-11-08, 23:03)HiassofT Wrote: Currently some important patches to the bcm2835 DMA driver are missing in the 4.3 kernel tree so I2S soundcards won't work.

These patches are included in a rather huge rework PR, https://github.com/raspberrypi/linux/pull/1178 which is currently being tested. Could take a while until all issues are resolved, but after that PR gets merged I2S soundcards should start working again.

I've pulled the (hopefully) relevant commits from that large PR to rpi-4.3.y_rebase. Does it look plausible to you, or do we need any of the other commits?
I now tried
hvs_priority=0x01000000 and v3d_clock=400

no artefacts but glitches when watching hd channel servus tv.
(2015-11-09, 23:52)polo_joe Wrote: I now tried
hvs_priority=0x01000000 and v3d_clock=400

no artefacts but glitches when watching hd channel servus tv.

Can you wind the "N" up? E.g. try 0x08000000 and if still no glitches try 0x0C000000 etc.
(2015-11-09, 22:06)popcornmix Wrote: I've pulled the (hopefully) relevant commits from that large PR to rpi-4.3.y_rebase. Does it look plausible to you, or do we need any of the other commits?
Looks fine! Did a very quick test and it worked well. Thanks a lot!

so long,

Hias
(2015-11-10, 01:18)HiassofT Wrote: Looks fine! Did a very quick test and it worked well. Thanks a lot!

Thanks for checking. Hopefully Milhouse's build will have those changes in and if testing is positive I'll pull them into main 4.3 tree before the rest of the big PR.
New OpenELEC Jarvis build #1109: RPi / RPi2
(Supercedes previous build)

Code:
# uname -a
Linux rpi512 4.3.0 #1 Mon Nov 9 21:48:47 GMT 2015 armv6l GNU/Linux

# vcgencmd version
Nov  6 2015 21:55:08
Copyright (c) 2012 Broadcom
version 40eee6c6d44edb1db4302217927a66a8dffd9374 (clean) (release)

# lsb_release
OpenELEC (Milhouse) - Version: devel-20151109214758-#1109-g25c23fe [Build #1109]

# vcdbg log msg 2>&1 | grep DTOK
001691.791: Kernel trailer DTOK property says yes

# Kernel device tree status: Enabled

Based on tip of OpenELEC master (25c23fef, changelog) and tip of XBMC master (b44c403d, changelog) with the following modifications: Build Highlights:
  1. Add i2s patches, hopefully fix broken audio devices
  2. Add DMA_BCM2835 kernel config option (required by i2s changes)
  3. Add Bluetooth audio options to resume playback on connect, and pause/stop on disconnect
Build Details:
  1. XBMC:
    • [ios] - fix playback controls from ios control center when kodi is pl… (PR:8367, 1 commit, 1 file changed)
    • [PVR] Context menu cleanup, step 4 (PR:8360, 6 commits, 16 files changed)
    • [builddep] fix build Ubuntu 16.4 Xenial (PR:8350, 3 commits, 3 files changed)
  2. dcadec:
    • Return frame losslessness status to API user. (59335fca)
    • Remove clipped samples counter. (a9888b67)
  3. pvr.hts:
    • Removed unused vars from ChannelTuningPredictor:TongueredictNextChannelId (PR:132, 1 commit, 1 file changed)
  4. service.openelec.settings/bluetooth_pulseaudio (lrusak):
    • New commits in this build:
      • initial pulseaudio support (c486dbf4)
      • bluetooth: add option to play/pause kodi on device connect/disconnect (0b6912ee)
  5. newclock5:
    • New commits in this build:
      • ffmpeg: bump to 2.8-pre (38b7efd6)
      • squash: rbp: Allow interlaced resolutions into supported hdmi mode list 2 (bbc5f7e7)
    • Commits no longer in build:
      • gl: split output stage from yuv2rgb shader (98c673ee)
      • omxplayer: squash: ignore rectangles from right eye (309cd087)
      • fixup! filesystem: Make support of browsing into archives optional (2f21d902)
      • squash: Fixup prototype to decode (2adc5fd6)
  6. kernel 4.3.y_rebase:
    • New commits in this build:
      • bcm2708-i2s: Remove driver (39fee189)
      • dmaengine: bcm2708: Make legacy API available for bcm2835-dma (58c46cd9)
      • dmaengine: bcm2835: Add slave dma support (cca19d58)
      • dmaengine: bcm2835: set residue_granularity field (635d0dda)
      • dmaengine: bcm2835: Load driver early and support legacy API (e2fb222a)
      • dmaengine: bcm2835: Add debug option for setting wait states (3f90c1f0)
      • configs: Enable DMA_BCM2835 (e26d3431)
      • dmaengine: bcm2708: Change DT compatible string (b8313464)
      • dmaengine: bcm2708: Remove driver but keep legacy API (91aeeb42)
      • bcm2835-dma: Fix dreq not set for slave transfers (05e51918)
      • bcm2835-dma: Limit cyclic transfers on lite channels to 32k (180bbb9e)
  7. Additional commits/pull requests/changes not yet merged upstream:
    • Added: [env] patch: Enable DMA_BCM2835, needed by i2s audio cards in 4.3 kernel
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.
(2015-11-09, 09:09)Caramba69 Wrote:
(2015-11-07, 15:25)Caramba69 Wrote:
(2015-11-07, 12:35)mule1 Wrote: No one else having these issues? Is navigating (skipping forward, backward, fast forward, rewind) in 3d movies working flawless for you?

I will check this tomorrow as well. Perhaps also somebody else can check?

Skipping works for me.
Thanks for the info. Did you try it by using an NFS connection to a NAS or with a direct attached storage?
(2015-11-09, 18:28)newjoe1 Wrote: In 3D MVC MKV files i even noticed strange behaviour after navigate in the file.

Yesterday Evening i have tried to play a 3D MKV Movie from a moment i have stopped it one day before. I ve got no sound and only ultra slow slowmotion in the video. After that i restarts Kodi, shutdown my avr and gave it a new try. But the same behaviour as without restarting the engines. Another try with settings audio and video to standard. Even this shows no effort on the movie, no sound and ultra slow slowmotion.

My workaround was to try openelec 6.0 final. And now the movie plays.
I can confirm extactly the above described behaviour while navigatiing using the testbuilds and openelec 6.0 final. Therefore openelec 6.0 final gives me some stutters with non 3d material while the testbuils run smooth. Someone who can mix best of both versions? Wink
  • 1
  • 131
  • 132
  • 133(current)
  • 134
  • 135
  • 168

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