Originally Posted by suarsg
Anyone know why this test pattern by R.Masciola:
doesn't correctly clip at 1000nits on my and all of my friends' B7/C7? The 2000 and 4000nits patterns are clipping correctly but the 1000nits one goes even farther than the 4000nits pattern. I've tested it on FOUR different players, including the TV's internal player and it's always the same result. This is on the technicolor-preset (everything default, Dynamic Contrast off).
Why is this TV treating 1000nits data so weird
We've also tried it on a C8 and C9 and there it clips correctly at the 1000nits mark.
I can't imagine it's just one of those things LG never bothered to fix or implement correctly in the first place. Seems like a pretty big deal if it indeed doesn't tonemap 1000nits movies correctly.
Originally Posted by suarsg
11. HDR Clipping Grayscale 1,000 nit_HDR10.mp4 (under 06. Miscellaneous Patterns > 15. HDR Clipping Test Patterns):
clips at 1000 on C8 and C9 / clips at 4000 on C7
12. HDR Clipping Grayscale 2,000 nit_HDR10.mp4
clips at 2000 on C7, C8, C9
14. HDR Clipping Grayscale 4,000 nit_HDR10.mp4
clips at 4000 on C7, C8, C9
11. HDR Clipping Grayscale 1,000 nit_HDR10.mp4 (reencoded to 800nits as Max Mastering Display Luminance and MaxCLL)
clips at 800 on C7, C8, C9
11. HDR Clipping Grayscale 1,000 nit_HDR10.mp4 (reencoded to 1200nits)
clips at 1200 on C7, C8, C9
11. HDR Clipping Grayscale 1,000 nit_HDR10.mp4 (reencoded to 801nits)
clips at 800 on C8 and C9 / clips at 4000 on C7
11. HDR Clipping Grayscale 1,000 nit_HDR10.mp4 (reencoded to 1199nits)
clips at 1200 on C8 and C9 / clips at 4000 on C7
C7 on the latest firmware (05.80.35). Default technicolor profile, Dynamic Contrast Off. Internal USB player as well as HDMI-fed sources.
I did some further testing with 10%-sized window patterns with different mastering metadata. Default Cinema-preset without any calibration, Dynamic Contrast off
pattern, with Y code value 240, should output about 2.47cd/m²
with 800cd/m² mastering (MaxMDL) I measured 2.0cd/m²
pattern, with Y code value 504, should output about 94cd/m²
pattern, with Y code value 632, should output about 384cd/m²
I don't understand wtf the TV is doing, at all.
With 1000 mastering, it's way too bright. With 800 and 1200 it seems much closer to what it should be.
(4000 is too dark as expected on these TVs, I guess. The dark 20% pattern is not bright enough for any of the mastering data, but this seems to be expected on these LG TVs as well.)
1) Why is 800 and 1200 so close to PQ EOTF but the 1000 one is so bright?
2) Seeing how the 1000 clipping pattern behaved like the 4000 clipping pattern in the previous tests, I expected the 1000 above to be actually darker and similar to the measured values of the 4000nits mastered window pattern. How is it possible the clipping test and the window measurement test are so contradicting?