AVS Forum banner

1421 - 1440 of 3759 Posts

·
Registered
Joined
·
17,811 Posts
Miki, Thanks for the review and your time. We need to remember this was a very high luminance LUT, 260 nits, so there was probably some drift due to the panel. I'm going to run another 10,000 pt at 100-150 nits to see what the difference is if any at a more "reference" luminance. I'm not saying that there is no problem with the LUT, but I do know from experience that the LG panel becomes more unstable displaying these high luminance patches for so long.

In discussions with Miki, I think he agrees, the PQ can look good even though the LUT has some problems. It's interesting to see any anomalies and test patterns that may show them. I hear there are to be some improvements in the next Betas so we will see what happens. :)
It appears I’ve missing the fireworks from earlier. Regarding this post...... The high light output does not account for 90% of what Miki pointed out. Yes the brighter the OLED, the more non linear the results are, but I’ve seen the same results from my tests whicre I set the light output to 130, 120 and 105 nits. I’ll let you all continue to test these betas. Ring me when R2 is available. Then and only then will I allocate equipment and time to test LUT creation via CalMAN.
 

·
Registered
Joined
·
9,197 Posts
TV native's mode is 10000K, for that reason you do 1D LUT first (to minimized the required changes for 3D LUT later). With LUT upload function you will be able to load 3D LUT only, it will be required a LUT upload for 1D LUT to work correctly so to have access to complete LUT features of the display.
I think I understand what you are saying here. You want a function to upload a 1D lut as well as a 3D lut. Also, we should perform the creation of the 1D lut first, so that the 3D lut will have less work to do.

What I don't understand is the start. "TV native's mode is 10000K".
 

·
Registered
Joined
·
8,913 Posts
I think I understand what you are saying here. You want a function to upload a 1D lut as well as a 3D lut. Also, we should perform the creation of the 1D lut first, so that the 3D lut will have less work to do.

What I don't understand is the start. "TV native's mode is 10000K".
Panels native color temp is 10000K. So when you reset the 1D LUT via CalMAN you get this color temp if you measure the grayscale.

Neil (the LG's architect and responsible for that LUT design) has posted about this here: https://www.avsforum.com/forum/40-oled-technology-flat-panels-general/2965862-lg-55-c8-4k-hdr-oled-tv-first-impressions-review-18.html#post56320392
 

·
Registered
Joined
·
111 Posts
Findings:

Target: 2.3, 110 nits. 3,500 pts. Technicolor.

Prior to any calibration, 54 seems to be my proper brightness setting (17 flashing). AutoCal will automatically reset back to 50 so the following impressions are based on 50.

1. Before any calibration, bar 21 is flashing with 20 barely visible. After GS autocal, 20 starts to barely flash.
2. After manually tuning 2.3% and 3%, 19 starts flashing but 17 is clipped and 18 is barely, barely visible. Just for testing, I had to increase the brightness to 57 at this point to get 17 to flash.
3. Set brightness back to 50 and after heavy manual tweaking, brightness pattern remains the same.
4. After 3DLUT, 18 barely flashes but 2.3% measures 0 nits. The other flashing bars also have a colored tint to them and are dimmer overall. 17 is still clipped. As you could see from the post-3DLUT grayscale, much has changed from the manual gs calibration.

So on my set, I don't think I was ever able to get 17 to flash, even with BT.1886 target. Manually calibrating at 54 brightness did not help as:

1. The 3DLUT will adjust the near black anyway (and will still clip 17, and 18 most of the time).
2. 17 won't flash even if 2.3% is adjusted properly at 54 brightness. Which is a bit odd.

Content looks great but some shadow detail is crushed, especially with overlaid dark clothing and hair. Ted's color ramps still have problems in magenta and blues but it didn't seem as bad as the 6,000 pt LUT I did a few days back (although the avg deltaE was larger this time).

I'll enjoy what I have now and will wait until the next beta before throwing the meter back on.

Btw, when I calibrated HDR10, I copied the GS values over via DDC to a Dolby Vision picture mode. From my understanding, this shouldn't have affected the DV picture mode itself but the picture seems to be better with familiar content (Lost in Space, Jessica Jones, etc). (shrug)
 

Attachments

·
Registered
Joined
·
1,081 Posts
My post worked like it should and I recived an answer from Tyler so just for tranceparency sake he said that the LUT upload will work with the public release of R2 we will see.
@BlackJoker ,

are u currently using the ee as an external LUT box on the C8 ? (somebody else mentioned that)

if so, and if u can, once the LUT upload works in CM, please compare the LUT processing of the LG vs. that of the ee...

so, if u currently have a known good cLUT - the one in the ee that cals the C8 just fine - convert that via LS to CM format, then upload to C8 and run validation...

or u could use the cLUT as active LUT in LS (and do a profile) and then compare to a profile when that same LUT is active in the LG...

if the internal LUT processing of the LG is not very clean then a lot of noise in here is pointless...

Thanks.
 

·
Registered
Joined
·
299 Posts
@BlackJoker ,

are u currently using the ee as an external LUT box on the C8 ? (somebody else mentioned that)

if so, and if u can, once the LUT upload works in CM, please compare the LUT processing of the LG vs. that of the ee...

so, if u currently have a known good cLUT - the one in the ee that cals the C8 just fine - convert that via LS to CM format, then upload to C8 and run validation...

or u could use the cLUT as active LUT in LS (and do a profile) and then compare to a profile when that same LUT is active in the LG...

if the internal LUT processing of the LG is not very clean then a lot of noise in here is pointless...

Thanks.
What about using the beta version where the upload was possible and do the upload with the ls lut.
 

·
Registered
Joined
·
8,913 Posts

·
Registered
Joined
·
299 Posts

·
Oppo Beta Group
Joined
·
11,496 Posts
So, please, stop thinking there should be a separated thread, this is not the CalMAN thread, this is LG C8 calibration thread, so THIS IS the "room" where confrontation MUST be kept. Because from confrontation we will improve, if you think you are already good... well... good for you, but only for you.

Competition is good, confrontation gets combative and that may not be good.
And posts such as this, are distractions from calibration.


- Rich
 

·
Registered
Joined
·
334 Posts
On my older ef9500 OLED rising the OLED light to 100 eliminated ABL. My question is does it work the same way on the c8?
 

·
Registered
Joined
·
8,913 Posts
I wrote the version where it worked. When it is this version then use this version.
I don't thing even worked. It will require one activation sacrifice to check.
 

·
Registered
Joined
·
299 Posts
I don't thing even worked. It will require one activation sacrifice to check.
I don’t understand what you want to say with the last part.
I know that Blackjocker uploaded an LS Lut with an oleder release.
 

·
Registered
Joined
·
8,913 Posts

·
Registered
Joined
·
588 Posts
@BlackJoker ,

are u currently using the ee as an external LUT box on the C8 ? (somebody else mentioned that)

if so, and if u can, once the LUT upload works in CM, please compare the LUT processing of the LG vs. that of the ee...

so, if u currently have a known good cLUT - the one in the ee that cals the C8 just fine - convert that via LS to CM format, then upload to C8 and run validation...

or u could use the cLUT as active LUT in LS (and do a profile) and then compare to a profile when that same LUT is active in the LG...

if the internal LUT processing of the LG is not very clean then a lot of noise in here is pointless...

Thanks.


Hi Mike yes I'm using eeColor with a LightSpace LUT and couldn't be happier once the LUT upload is working I will do that test of course but it wouldn't be comparable the LG can handle only 33^3 instead of 65^3 like the eeColor.

Even with the last beta the upload didn't worked after many tests @baller02 and I don't want to do further testa at the moment because it's a waste of time.

When someone else can proof that it's working I will give it a shot.
 

·
Registered
Joined
·
1,081 Posts
Hi Mike yes I'm using eeColor with a LightSpace LUT and couldn't be happier once the LUT upload is working I will do that test of course but it wouldn't be comparable the LG can handle only 33^3 instead of 65^3 like the eeColor.

Even with the last beta the upload didn't worked after many tests @baller02 and I don't want to do further testa at the moment because it's a waste of time.

When someone else can proof that it's working I will give it a shot.
trust me the cube size 33 vs. 65 is not a problem. Ur original profile did not contain enough data for either 33^3 nor 65^3, hence all that data is interpolated.

As long as u use the original LS profile and create the 33^3 CM LUT from it, the test is good to go... I don't expect much distortion from LG's internal LUT, but one never knows... we caught quite a few LUT holders in the past that exhibited it in action, Eizo comes to mind...

easiest/best way is to do LS active LUT vs. LG enabled LUT...

Thanks !
 

·
Registered
Joined
·
588 Posts
trust me the cube size 33 vs. 65 is not a problem. Ur original profile did not contain enough data for either 33^3 nor 65^3, hence all that data is interpolated.

As long as u use the original LS profile and create the 33^3 CM LUT from it, the test is good to go... I don't expect much distortion from LG's internal LUT, but one never knows... we caught quite a few LUT holders in the past that exhibited it in action, Eizo comes to mind...

easiest/best way is to do LS active LUT vs. LG enabled LUT...

Thanks !


I will definitely provide that data Mike ones the function is fully enabled.
 

·
Registered
Joined
·
976 Posts
Competition is good, confrontation gets combative and that may not be good.
And posts such as this, are distractions from calibration.


- Rich
I'm sorry Rich but I don't get the usefulness of your post. You've just extrapolated a piece of my post where I spoke about John's 3dl and you stamped it like "distraction"... Would you mind to explain that privately (pm)? Thanks.
 

·
Registered
Joined
·
334 Posts
Calibration with Autocal works. But need still some manual tweaking after it. So some willtekk then you can do it totally manual. If you Re really confidential and has alexoerience in calibration especially in OLED's then manually calibration is even faster with better results.
The best think is that u can set the brightness etc without TV menu. The lut creation is broken in this state as shown buy users who are passionate and got other software to prove the charts of cm.

So calibration works as before Autocal needs little manual tweaking after. The lut can't be recommended in this early state. But this was always the weak part of CM. But hey it isn't comparing a Porsche with an fiat 500 on the German Autobahn when wlyou will drive fast es hell you won't choose the fiat. So it is with lut creation. But for calibration purpose it is easier with Workflows to reach the goals.and the carts are easier to interpret. So you can calibrate you 2018 LG but not with the main deal bringer the lut.
Nice summary of the whole thread! I'm late to the game so I was reading through 40+ pages to catch up.
 
  • Like
Reactions: jrref

·
Registered
Joined
·
334 Posts
On the C8 has anyone notice any difference or drawbacks when using DDC instead of manual setting calibrations? When changing DDC settings, will it show up on the tv internal settings after calibrations? I do not currently have a meter as of right now therefore I am unable to test it out myself.
 
1421 - 1440 of 3759 Posts
Top