I've posted this:
https://forum.kodi.tv/showthread.php?tid...pid2835975
If a whole DirectShow chain of reproduction is unfeasible, maybe we could hope for a "simple" video renderer add-on, allowing use of madVR. I don't know if @
madshi reads these forums but I guess it's worth trying to investigate it.
I like that idea, but it would require a good developer interested primarily in madVR. You would still have the issue of getting the HDR metadata to madVR and likely slower rendering performance with Kodi's video decoders compared to the madVR integration possible with D3D11 Native in LAV Filters.
(2019-03-19, 23:20)Warner306 Wrote: [ -> ]I like that idea, but it would require a good developer interested primarily in madVR. You would still have the issue of getting the HDR metadata to madVR and likely slower rendering performance with Kodi's video decoders compared to the madVR integration possible with D3D11 Native in LAV Filters.
Hmmm... in a LAV-->madVR chain is LAV in charge of communicating the HDR metadata to madVR? Is that what happens?
I'm not sure that video decoders speed would be so sacrificed, unless I'm misunderstanding what you mean. Kodi's VideoPlayer decodes in hardware already both AVC and HEVC, right? Is there a point in measuring decoding speed once decoding happens in realtime?
D3D11 Native is integrated into madVR. Some tasks like highlight recovery strength for tone mapping are significantly faster with D3D11 Native. I don't know if there would be a difference with Kodi's video decoders, but certains tasks in madVR would have to be slower. Maybe the general performance could resemble DXVA2 copy-back, but I don't know. Copy-back matches D3D11 Native in most scenarios, but could become slower by comparison in the future.
The thing is, if we are talking about an addon to videoplayer instead of an update to Dsplayer, it would be just as well using 18.1 and using an external player with madVR, as if there are any performance hits, there is nothing to be gained really ?
To be honest I am in no rush to upgrade, I understand the Netflix side, but that plays fine with the Windows 10 app for me and the addon I tried a while back was nothing to get excited about, so personally I cannot see the need even if there was someone kind enough to step up, as it may be a work in vain due to any issues, and maybe being worse than we currently have.
(2019-03-20, 16:37)madjockxbmc Wrote: [ -> ]The thing is, if we are talking about an addon to videoplayer instead of an update to Dsplayer, it would be just as well using 18.1 and using an external player with madVR, as if there are any performance hits, there is nothing to be gained really ?
Warner306 describes in some details the drawbacks, in his guide to MPC-HC integration. For me the inability to save watched status (visible through GUI, I know you can have MPC-HC remember), the inability to use subtitles plugins are enough to discourage me from adopting an external player solution. On the other hand, while I subscribe to both Netflix and Prime, I prefer to watch tv episodes through Kodi and last night I was watching an episode that seriously suffered for the lack of debanding in VideoPlayer (dark scenes where compression artifacts were quite obvious).
In any case, $500 from me for a perfectly working 18.2 Kodi DSPlayer build.
If anybody wants to join the bounty, feel free. Since for the moment it's only from me, I consider it "honour bound", if anybody else wanted to join, we'd have to devise some method for it. Any dev interested present yourself here and state your intentions.
Maybe merging the changes from 17.6 to 18.2 is not *too* complicated, I don't know.
Edit: to avoid misunderstandings, the build would have to be publicly available for everybody to download.
I hope someone takes up the challenge. But I'm not sure who might be reading this thread.
(2019-03-20, 23:04)ashlar Wrote: [ -> ]In any case, $500 from me for a perfectly working 18.2 Kodi DSPlayer build.
If anybody wants to join the bounty, feel free. Since for the moment it's only from me, I consider it "honour bound", if anybody else wanted to join, we'd have to devise some method for it. Any dev interested present yourself here and state your intentions.
Maybe merging the changes from 17.6 to 18.2 is not *too* complicated, I don't know.
Edit: to avoid misunderstandings, the build would have to be publicly available for everybody to download.
I can't afford that much i can do $20
(2019-03-21, 14:42)Warner306 Wrote: [ -> ]I hope someone takes up the challenge. But I'm not sure who might be reading this thread.
I posted on doom9 about this but feel free to spread it wherever you deem necessary/useful. I might be wrong but I hope that having a "simple" merge of two existing codebases could be easier than starting from scratch.
The other option, as I mentioned on Doom9, could be using a madVR Envy after the HTPC output stage. With resolutions whitelist it should be now doable through Kodi 18+.
Maybe we could ask the kodi developers to integrate madvr/lav filters support into kodi 19? wouldn't that be possible for the windows version of kodi? I mean media portal and jriver got it too...
They've been asked, and they blatantly refuse implementing a feature only for Windows. It doesn't help that all of their devs are Linux.
I would put down 10 or 20 as well but I'll need some kind of cable TV integration that does timeshifting either via something like NPVR or the unofficial hd homerun pvr add-on. My current hold up in implementing Kodi ds has to do with live TV deficiencies and the current pvr options. I have at least one willing pvr author but no one to guide him.