• 1
  • 84
  • 85
  • 86(current)
  • 87
  • 88
  • 138
Android nVidia Shield TV (2015 & 2017 Models) - UPDATED: May 25, 2018
(2018-06-12, 21:20)m.r.f. Wrote: Plasmas and OLEDs TVs no need HDR.
HDR is good for LCD (led) TVs to disguise the blackest. 
Wrong.
Reply
(2018-06-13, 16:19)Nazgulled Wrote:
(2018-06-12, 21:20)m.r.f. Wrote: Plasmas and OLEDs TVs no need HDR.
HDR is good for LCD (led) TVs to disguise the blackest. 
Wrong.
HDR is nothing more than brightness increases in color. which is ideal for disguising the blacks of the LCDs.
Reply
No, it's also range of colors / color depth & visibility of additional details from the expanded range.
Reply
(2018-06-14, 00:49)m.r.f. Wrote: HDR is nothing more than brightness increases in color. which is ideal for disguising the blacks of the LCDs. 
I am just not sure how much you know about HDR. HDR isn't just about max. luminance. It is all about the dynamic range. Human eye has a dynamic range of ~20 stops. Conventional gamma curve used in SDR has a dynamic range of 6 stops. PQ transfer used in HDR video has ~14 stops. It is a different debate whether existing display technology can achieve this or not.
Reply
Not to mention those with tetrachromacy
Reply
(2018-06-14, 02:09)davilla Wrote: Not to mention those with tetrachromacy
Sounds like a Kodi Hardware love story ! Wink

Reply
(2018-06-14, 01:17)wesk05 Wrote:
(2018-06-14, 00:49)m.r.f. Wrote: HDR is nothing more than brightness increases in color. which is ideal for disguising the blacks of the LCDs. 
I am just not sure how much you know about HDR. HDR isn't just about max. luminance. It is all about the dynamic range. Human eye has a dynamic range of ~20 stops. Conventional gamma curve used in SDR has a dynamic range of 6 stops. PQ transfer used in HDR video has ~14 stops. It is a different debate whether existing display technology can achieve this or not. 
 This is what is announced.
When they implemented hdr on tvs, they certainly would not say it was due to the problem of lcds with blacks, because it is not possible due to the limitation of the technology.
It was never implemented in plasmas or in the first oleds because it was not necessary.
Of course, today's oleds have why it's what the market dictates.
The abitual consumer when going to buy a tv, does not know the difference of a LED TV, OLED or QLED, for most everything is equal because everything ends in "LED", just want to know if the TV is 4K (even if it is seen to 10 meters) and HDR the rest does not matter.

I apologize this is not a subject here.
It's offtop.
Reply
(2018-06-09, 10:46)Tinwarble Wrote: A little surprise for music lovers in Oreo.

Image

It works for Pandora also, but I'm going to go out on limb and say it may work for most other music apps.
I can confirm for Spotify, Deezer and TIDAL.

Image

Image

Image
Nvidia Shield TV 2015, Sony BRAVIA XBR55HX929, Yamaha Aventage RX-A3050, Dolby Atmos/DTS:X (5.1.4) Front Mains: Jamo C109, Surrounds: Jamo C103, FH/RH: Jamo C93, Center: Jamo C10 CEN and Subwoofer: Jamo J112 SUB
Reply
Now that nVIDIA has provided the recovery image download for 7.0.1 update, I updated one of my Shields that has an open bootloader. I did a quick testing of the new gamut mapping API in Android Oreo. It is different from SE 6.3 update especially at 100% saturation levels. The mapped values all seem to lie within Rec. 709 boundaries (In SE 6.3, the 100% points were close to BT.2020 boundaries). With that said, I got some mixed results. While the theoretically calculated CIE 1931 points were very close to the reference Rec.709 points, the results were different when checked on a CIE scope. CIE scope display showed a desaturated image with 7.0.1 update. The desaturation is very obvious in the frame grab. 

CIE 1931 Chromaticity Chart

Image

Pink diamond: reference Rec.709, Filled circle: Shield Experience 6.3 update, Filled square: Shield Experience 7.0.1 update. Black triangles represent the boundaries for BT.709 and BT.2020 gamuts (smaller one is BT.709).

The tests were done on two different Shield Android TVs, one with 6.3 update and the other with the 7.0.1 update. The test patterns were classic color checker and 100% saturation/stimulus test pattern of primaries and secondaries. The test patterns were played in the latest Kodi Leia alpha build. Shield's output was set to 4K 23.976Hz 12-bit Rec.2020. The code values of the output signal were read of a HDMI analyzer. XYZ and xy calculations were based on ITU-R BT.2020 and BT.709 recommendations.

EDIT 6/17/18
I have removed the CIE scope captures that I posted earlier because I suspect something weird going on with the waveform monitor when the primaries are set to BT.2020.

At this time, I am just going to go ahead and say that the new colorspace API does a good job rendering BT.709 gamut to BT.2020 gamut. It is doing an "absolute" or "saturation" rendering. This is as good as it can get! 

By making this bold statement, I hope I won't have to eat my hat.  Wink  
Reply
(2018-06-18, 02:27)wesk05 Wrote: Shield's output was set to 4K 23.976Hz 12-bit Rec.2020.

Quote:At this time, I am just going to go ahead and say that the new colorspace API does a good job rendering BT.709 gamut to BT.2020 gamut. It is doing an "absolute" or "saturation" rendering. This is as good as it can get! 
Lots of details thx.

Ok but what about:
a) refresh switching and colorspace (and bit depth) output after refresh switching happens, is that not the whole point of Kodi Leia's "Whitelist" ?
b) and then what happens to non 4K content in regard to upscaled picture quality if you permanently set a 4K 23.976Hz 12-bit Rec.2020 ?

Reply
(2018-06-18, 08:46)wrxtasy Wrote: Ok but what about:
a) refresh switching and colorspace (and bit depth) output after refresh switching happens, is that not the whole point of Kodi Leia's "Whitelist" ?
b) and then what happens to non 4K content in regard to upscaled picture quality if you permanently set a 4K 23.976Hz 12-bit Rec.2020 ? 

a) There is no change to refresh rate switching. The color space (gamut) doesn't change. It remains at what ever you have selected in Shield's settings. If you have selected YUV 12-bit Rec.2020, that's what it will be even after refresh rate switches. The new API does a great job in mapping BT.709 to BT.2020. I haven't checked HDR to SDR or BT.2020 to BT.709 conversions.
b) I haven't specifically checked upscaling quality. I don't think that would have changed from the previous update. It is not required to set it to 4K 23.976Hz, I just set it to that to make things easier for testing.
Reply
So, in my case, with Optoma UHD51A projector, the "recommended" display setting I get is [email protected] / 4:2:0 / 10-bit. Whether I stick w/this or force Shield to "4K 23.976Hz 12-bit Rec.2020", all my 4K HDR 23.976 content plays at 8-bit 4:2:2 BT2020, at least according to my PJ. Believe @wesk05 you mentioned previously that my PJ may not be seeing this as 10-bit as it is. 4K@60 HDR like the Sony 'Camp' clip do play/show at 10-bit 4:2:0.

Problem w/forcing 4K 23.976Hz 12-bit Rec.2020 w/my PJ is that I get no picture w/4K@60 HDR like the Sony 'Camp' clip. On the plus side, Amazon Prime Video 4K HDR play really well at 24p vs. leaving at "recommended" [email protected] / 4:2:0 / 10-bit and everything rendering at 60fps.
Reply
(2018-06-19, 01:40)hdmkv Wrote: So, in my case, with Optoma UHD51A projector, the "recommended" display setting I get is [email protected] / 4:2:0 / 10-bit. Whether I stick w/this or force Shield to "4K 23.976Hz 12-bit Rec.2020", all my 4K HDR 23.976 content plays at 8-bit 4:2:2 BT2020, at least according to my PJ. Believe @wesk05 you mentioned previously that my PJ may not be seeing this as 10-bit as it is. 4K@60 HDR like the Sony 'Camp' clip do play/show at 10-bit 4:2:0.

Problem w/forcing 4K 23.976Hz 12-bit Rec.2020 w/my PJ is that I get no picture w/4K@60 HDR like the Sony 'Camp' clip. On the plus side, Amazon Prime Video 4K HDR play really well at 24p vs. leaving at "recommended" [email protected] / 4:2:0 / 10-bit and everything rendering at 60fps.
Yes, you can ignore the 8-bit 4:2:2 on your projector's info page.  As for the issue with forcing 4K 23.976Hz 12-bit Rec.2020 and no picture at 4K 60, is this with the Oreo update or did you have it before that? I noticed that when you set Shield to 4K 23.976 12-bit 4:2:2, 4K 50 and 60 output are also 12-bit 4:2:2. Maybe your projector doesn't support those modes (they are within HDMI 2.0a/b specs.).

Another thing that I noticed is, when I set Shield's output to 4K 59.94/60Hz 10-bit 4:2:0, 4K 23.976 is not listed in the whitelist of Kodi Leia, because of that when you play 4K 23.976fps content, the output is switched to 1080p23.976. I'm not sure whether this is a Kodi problem, Shield problem or a problem specific to my Samsung TV's EDID. Will have to look into this.
Reply
I have exactly the same problem with my Sony VPL-VW760 projector!
Reply
It's not possible to set the shield TV to 4K 12 bit 4:2:2 ( I don't have my shield at work lol )? and then it will be possible to play 50/60Hz and 23.976 4K files !
Does banding issue available when Shield TV is set to 12 bit REC 2020 and we play 4K 10 bit HDR files ?
Reply
  • 1
  • 84
  • 85
  • 86(current)
  • 87
  • 88
  • 138

Logout Mark Read Team Forum Stats Members Help
nVidia Shield TV (2015 & 2017 Models) - UPDATED: May 25, 20188