Posts: 89
Joined: Jan 2010
Reputation:
1
Adam
Any thoughts on supporting interlaced native resolutions like 576i and 480i? These would require awkward pixel doubling (horizontal only) due to the clock pixel lower limit of HDMI.
Posts: 16,946
Joined: Feb 2011
Reputation:
256
Thats way out of scope of the complexity the current approach has. (meaning what you suggests is a complete different thing which would need deep core changes...)
AppleTV4/iPhone/iPod/iPad: HowTo find debug logs and everything else which the devs like so much:
click here
HowTo setup NFS for Kodi:
NFS (wiki)
HowTo configure avahi (zeroconf):
Avahi_Zeroconf (wiki)
READ THE IOS FAQ!:
iOS FAQ (wiki)
Posts: 89
Joined: Jan 2010
Reputation:
1
Forgive me I don't have a receiver so I might be missing something here. Are you saying that your receiver will detect interlaced video delivered over a progressive frame stream? No television I know of can do that. Normally XBMC will weave together interlaced field pairs into a progressive frame, each frame then repeated.
I have never heard of such a feature, but it's a good one if it really exists.
What does 1080i video (with proper 50Hz 'fluid' motion content, not 2:2 telecined frame content) look like after going through your receiver (with XBMC deinterlacing off)? Do you get correct field sync every time?
Posts: 4
Joined: Jun 2012
Reputation:
0
Awesome discussion!!!
For the past 3 years, I have been using a WDTV Live and when I decided to take the leap and buy a D525/ION2 HTPC (Zotac ID41) I was taken aback with the change in the image quality of my SD (and 480p) videos playing on XBMC.
I was wondering, if the “Native Resolution” route is so complicated (Because of the menus and OSD), how about trying to implement a better upscaling similar to the techniques that receivers or TV have (Maybe using different algorithms)?
I hope this discussion leads somewhere; I think this is the last thing missing in a perfect solution for viewing media.
Thanks for all the work; I do appreciate the labor of love being done by the people who work on XBMC
Jack
Posts: 6,810
Joined: Jul 2010
Reputation:
198
I have thought we have good upscaling techniques and would like to know against what methods you compare the receiver. I can't judge because I have none but what I get out of XBMC is much better than the upscaling/de-interlacing my tv set does (which is in the rage of €1000).
Posts: 1,832
Joined: Nov 2010
Reputation:
58
2012-06-25, 20:17
(This post was last modified: 2012-06-25, 20:17 by joethefox.)
same here. XBMC deinterlacer + scaler is far better than the internal panasonic 30VT55'' for my eyes (nvidia vdpau)
Posts: 6,810
Joined: Jul 2010
Reputation:
198
I would rather see that we fix and optimize the existing features before we implement new ones.
- How do XBMC's advanced upscaling methods like spline or lanczos compare to the receiver.
- Add noise reduction filters
- crop distorted pixels at the borders (in particular needed for de-interlacing)
Posts: 6,810
Joined: Jul 2010
Reputation:
198
Yes, I do think that in general PC graphics does outperform those processors you named. It is actually designed to meet much higher requirements e.g. gaming and in the area of HTPCs we are dealing with by-products which in most cases are not maxed out.
If it turns out that XBMC can't compete I do support those features. E.g. some video card vendors don't provide proper de-interlacing methods on Linux. It makes sense to provide a method to hand this task off to a receiver or tv.
Posts: 3,571
Joined: Sep 2009
Reputation:
110
TV scaling is almost invariably crap (even on relatively expensive TVs), but processor scaling is often very good and measurably better looking that anything I have seen come out of any XBMC (Amd/nvidia, win or linux).
Even humble popcorn and WD machines have *way* better SD quality than XBMC, for whatever reason...it almsot put me off XBMC initially...
Addons I wrote &/or maintain:
OzWeather (Australian BOM weather) | Check Previous Episode | Playback Resumer | Unpause Jumpback | XSqueezeDisplay | (Legacy - XSqueeze & XZen)
Sorry, no help w/out a
*full debug log*.