Just a question, do the new version have a function to change opacity on subtitles? Vincent on hdtvtet suggested that for watching hdr.
You can change colors, so you could select light grey, grey or even dark grey to change too less white over there.
I’m getting sad face errors playing “Gemini Man” today. Logs: https://paste.osmc.tv/jabehehopa
Admittedly the file is a slightly odd one: it’s in a .TS container, and contains both tracks of dual-layer Dolby Vision. But it starts to play the HDR10 layer correctly, and then crashes after a minute or two of playback.
I’ll get back to you later today.
1 Like
So odd error here… when there is an update sometimes the skin files get overwritten and sometimes they don’t. I am using a weather addon that overwrites a couple of the skin files with customisation. What it does is it copies the relevant files and adds .original to the end of the file name and it uses that to determine if it needs to update the skin files again. According to the Kodi devs, updating Kodi should wipe out all files in the skin folder, (Estuary in this case) but that doesn’t seem to be happening as OSMC update is rewriting the xml files but leaving the backup files in situ so the patcher doesn’t work.
Is this expected behaviour?
In ISO container which have 2d and 3d movie: when watching 2d, subtitles goes to the right,not stays in the center as normal. And it is not possible to read them, because it is seen just first 2 or 3 words of sentence. In 3d movie subtitle is ok in the middle bottom as normal
On Nvidia Shield kodi 19.1 the same movie 2d plays subtitles normal in the middle
To understand better problem I take a short movie about 100MB. I can upload it ,just say where?, if it is needed of course
In movie Ant-Man and the Wasp this problem can be seen very easy.
It is iso container: inside have two " .m2ts" files
Both files media info:
https://paste.osmc.tv/vokurelefe.coffee
and
https://paste.osmc.tv/laroxikeba.vhdl
And here is my log from vero 4K:
https://paste.osmc.tv/pitimafiqu
Do anybody have the same issue with such ISO containers there is 2d and 3d file?
Thanks for help
This is dependent on whether or not a skin update has been pushed. We only update what has changed.
It will – if Kodi is updated.
The last update and a previous one overwrote skin files but left the backup files untouched. I don’t think there was any change to the skin files so I don’t understand why the OSMC update pushed a skin change…
When will the “official” Kodi 19 be released on the Vero 4k+?
When it’s ready – this has been discussed previously.
Ok so it’s problematic if the skin is updated but OSMC is leaving old files in the skin folder when it updates
OSMC won’t delete any files that you create.
We only replace files that we ship in our packages.
Add-ons and scripts aren’t suppose to be making changes to things outside their own userdata folder (unless it’s available as a JSON or Python call). If a script or add-on has been hacking the skin the way you described, and it doesn’t work anymore, I’m not sure how that’s an issue for the OSMC (or Kodi) developers. For it to continue working, the hack has to be updated to account for the way things work now.
5 Likes
So, I know you guys have been tinkering a bit in this area, and you probably aren’t finished yet. But just for completeness: there’s currently some weird stuff going down with 1080i/50 material.
Here’s a link to some of my deinterlacing test clips: Dropbox - Deinterlacing test patterns - Simplify your life
If you whitelist both 1080p/25 and 1080p/50 and then play the Sherlock clip, it begins playing with output set to 1080p/25Hz and looks like it’s playing in slow-motion; then it switches to 1080p/50Hz output, and plays a few seconds of video speeded up; then the speed settles down again, but playback is now jerky. Same thing happens on the Spearhead from Space clip. (Both of those are 1080i/50 h.264). If memory serves this is a downgrade from 3.14/Leia - it always did the thing of starting at 1080p/25 and then switching to 1080p/50, but previously it didn’t do the playing at high speed for a few seconds after the switch, and also it used to play smoothly (albeit with some deinterlacing artefacts) after the switch, but it’s now jerky.
Contrast, if you wish, the Little Dorrit and Planet Earth clips - these are 1080i/50 VC-1. With default settings they play the same way as the h.264 clips, but if you pause them immediately after the start and change Deinterlacing from Auto-select to Off, they then play very nicely at 1080p/25 (and don’t attempt to switch refresh rate).
1 Like
Incidentally, there’s a (probably unrelated) issue that I think other people have already raised: there’s something something subtly off with VC-1 decoding. If you play the 1080i/60 clip with hardware acceleration disabled, and set Deinterlacing to Off, it plays perfectly. If you turn on hardware decoding (and, again, set Deinterlacing to Off) and watch the right-hand wedge, there are some faint diagonal bands that sweep from side to side (which shouldn’t be there).
1 Like
Updated to the latest build (15/07/2021), all is working well for me.
Question for v19 final release: Per the initial post the “buster-devel” repository was enabled during the installation of the April v19 test image. Will the sources.list be updated automatically to change this back to “buster”, or should we do this by manually editing the sources.list file once the v19 final is released and installed?
Thank you…
That file will not be changed by OSMC, you should change it back to buster by yourself
1 Like
Thanks for those. Can you confirm which Kodi version and kernel (see top of kodi.log) you see this behaviour on? There’s been some comings and goings with deinterlacing in the last few days.
I have to say, with those deinterlacing test patterns, I’ve no idea what I’m looking for. Made more difficult for me as I don’t have a native 1080 TV screen, let alone a 480 line screen. So I’m always going to see some Moire.
I just grabbed some basic logs here: https://paste.osmc.tv/jamoqaziru
(Those are just to get versions; if you need actual debug logs for the behaviour I was talking about, let me know; but hopefully it’ll be fairly easy to reproduce).