To preface, I’m not trying to be presumptuous with anyone’s time, I understand and appreciate the enormous volunteer effort that goes in.
In the past few days, the testing for VS10 has reached quite a mature stage. The creator @cpm has related that he will be ceasing his individual development effort on VS10 as bugs have been ironed out, and the integration with mainline CE is quite well tested and laid out (UI).
Summary of the VS10 engine / VSVDB injection / HDR10+ enhancements:
Upscaling SDR/HDR into DV
Downscaling DV into SDR/HDR
Playing SDR content using VS10 processing
Playing FEL LLDV on non-DV panels like Samsung
Fixes the ATEME HDR bug
Allows conversion of HDR10+ into P8 DV
Projector users can customize EDID to unit specs
Maybe some more that I’ve missed
These features are nothing short of groundbreaking as there are no implementations of VS10 that are this mature. Some people will point to Zidoo, but they still lack real TV-led DV and still miss many of the above functionality. The downscaling of HDR/DV into SDR using VS10 is far superior to the standard Kodi method and the VSVDB injection allows functionality only seen in expensive HDFury devices. On HDR10+ to P8, this is completely novel, no other device enables this functionality.
I am kindly asking to see if there is any interest or plan in merging this effort into the mainline CE build. Many users at this point utilize the “cpm” build to gain access to VS10, and sharing this with the wider community would allow people to enjoy their content even more.
If there is anything the community can do to smooth this addition into mainline, I am sure many people would be happy to help/test/validate the functionality.
I think if enough in the community are interested in new features making it into ne, no etc. then could plan that work with the CE team and get a hired hand to build if there is no dev willing to take it on for free.
That way everyone in the community that wants a new feature can work with the CE team to outline how that would work and what it would look like in terms of options etc. up front and then need a willing dev or hire a dev (I think plenty of options out there) to implement the agreed upon changes, so everyone can have a say and it all would be understood upfront and once done the specific feature can be PR’d one at a time - that way I think all parties would be happy.
Only down side is likely some expense, but if spread across the community then maybe it is not high per contributor.
Anyway just a thought, I will continue to have a background interest in my changes on ng and attempt to address any repeatable issues that come up.
From what I understand the SoC in AM6B+ is not suitable for ne, but could be ok for no - just need to dovi.ko for that as a baseline, so if that becomes a reality may also look to make it work with my changes.
There are simpler elements that could be pulled out and PR’d yes, such as the ATEME workaround.
But the majority would take effort to break out as separate PR as much is shared and changed and cleaned up - this is time which I don’t have and not something I would like doing.
To restack the changes into a nice neat set of PR which build on each other is just not in my future , but does not mean others could not help do this.
For example I could see the HDR10+ > P8.1 actually can just go up into Kodi, if there was a Kodi dev wanting to take it on, it is just C++. Then come back down to CE, Android etc.
If anything were to come from this I would recommend first all agree and are clear on what you want to achieve as a group - the UI brings this all together so I suggest that is used to determine available functionality and what should be worked on, as everyone can relate.
This is why I postulated the approach in this above post:
Everyone can determine what is desirable and what the CE team would accept in their product and then have a resource work on that - can refer the repo I have if that is helpful to that resource, then community can test inline with CE wishes and eventually the final product emerges with everyone happy.
Bottom line.
For me I built out the elements / tried to share with everyone / willing to explain etc. - but fundamentally I have what I want out of it - I think it will now age just fine and go hand in hand with my AM6B+ for my remux playing needs. I also provided features I would likely never use because of the challenge and to help others.
If the community / CE want to add similar feature sets to the mainline then it is there for reference and I can help with understanding, but not to the level of a set of neat PR.
Apologies if this is not the outcome everyone was wishing for at this point.
Hope that draws a line under this from my perspective.
I think a PR for the ATEME fix would be greatly appreciated by the wider community.
For me, my AM6b+ performs amazingly well with the latest T6A2 version. The HDR → DV upscaling is just too good. I watch all HDR-only content in VS10-DV now and it really is a huge quality of life upgrade. I think I will stop here in terms of updates on this box. After all even Nvidia goes several years without releasing updates
No promise I won’t bother you for a new build though, maybe on the next NG major release.
@Portisch Do you think it would be possible for you to look through these patches and merge them into the main CE branch or do the work necessary to bring them to CE. CPM and the community has spent months developing these features and they’re highly valued. It would be unfortunate if there was basically two branches of CE: CPM’s and the main one.
The changes must be split in smaller parts to be able to check it separately.
Also we are still look forward for NE/NO, not NG anymore.
The problem is it’s such a big change in code and we would require much more man power to manage this. It’s not only about merging the code. Also Wiki need to be maintained and multiple users will need help and support to be able to manage such advanced change.
We have currently no idea how to be able to handle this in “mainstream” CE.
Like as a example the 3D implementation took several weeks, about 8h/day and this does almost equal these feature request.
Maybe a fork like EmuELEC will be currently the better choose.
We are also partner with Shanti, the author of EmuELEC.
Time by time we can transfer items upstream to CE until all is stable and merged.
This is now a PR for 4.9, but you will need to make 2 more.
For 5.4 and 5.15 to keep all up to date and at the same level.
This is what I am talk about, the work is not just merge it in NG - done. It’s about upstream to every active branch and test all together on every SoC and with all test media we currently support…
So all minimum * 3 in man power. Please remind CE is not a commercial product, it’s a hobby.
Yep so hopefully now everyone sees the complexity involved.
I would have hoped AML had fixed this particular issue in the later kernels so no need for it, but would require someone with the right equipment, build and testing to be sure in the other kernels, that rules me out.
I can only go back to my previous suggestion for the community to hire in a resource to work with you to build out the features as wanted on the versions you want them on and formalize this all a bit more so you have a pipeline can easily work from - can use my repo as a resource for that if appropriate, it will have some cost but looks the way forward to tick all the boxes and have all the features within the limits of everyone’s time.
Thanks for you and the CE team for providing the open source foundations to try things on.
It does not have to be me to do update builds, it has a relatively low barrier to entry.
If you are responsibly competent around computers, can setup a VM, understand some command line and have an aptitude towards programming concepts, then I think with about 2-3 weeks of study on github and git commands and the CE structures and repo you will have the basis to be able to fork the repo, pull over the changes and do builds when something underlying changes.
No need for deep knowledge on programming etc. to get that done, not saying it wouldn’t be a challenge at first for sure, but we should all challenge ourselves, and you will have some satisfaction in building something.
I have everything already setup and can do updated builds once in a while (definitely not daily) as I’m doing it for myself anyway, but only if there’s anything ‘meaningful’ changed/updated.
I think it’s known to all of our users that our dev team is stretched thin. And when I say thin I mean really really thin. So, over the years, we’ve tried to “pull” people with knowledge to be in the team.
This is one of these times and we’re hoping that these changes can be brought on from the Team, officially. For that we need people to step in if they can or want. We can’t force anyone to be part of the team, we can just ask.
, but does not mean others could not help do this.