LG 32LM63xx calibration (HCFR)

[Luminance Leon]

Hi all, and seasonal greetings!

So the LG TV we'd shortlisted came up in stock, so we bought one and I got to build the stand and install it Christmas morning, happy days.
Comparing the image to the old CRT was tough as IPS can't do great blacks and this TV doesn't have amazing levels of white - which isn't terrible as it can help reduce eye strain with a somewhat softer feel. Set to Cinema mode, dropped the backlight and tweaked a few settings to reduce out-of-the-box garishness, and I think it was an improvement in realism over the CRT, which may have been changing colour with age.

Anyway, enough waffle...

I grabbed the old Windows 7 netbook with the latest HCFR (v3.5.1.4) installed , the ColourMunki Display colourimeter and an HDMI cable and set about seeing if I can dial things in a somewhat organised fashion.

Reviews of some other IPS LG TVs suggested the following settings:

Spoiler: settings

Mode: "isf Dark Room"
Backlight: 46
Contrast: 80-90
Brightness: 50
Colour: 50
Sharpness: 10
Tint: 0
Colour Gamut: Auto
Gamma: 2.2
Mode: Warm 2
All of those dynamic and auto image adjusting enhancements should be disabled.

Using this as a base, Warm 2 was too yellow, and Warm 1 was a little cold. Brightness was far too milky, and a more reasonable 25 made for a more realistic image at some expense of shadow detail.

I fired up HCFR for the first time, I have used DisplayCAL before which just does it all for me and created a LUT at the end too. This time I knew I would have to be much more hands on with adjustments, getting feedback as I go. No where to load in a LUT here!

HDTV Rec709 and the 25%/75% version seemed to be the right choice. D6500 white point. Gamma 2.2 power.

I started with a grey scale measurement, and it showed that things weren't too bad, just that blue was lacking across the board - the reason why it was too yellow!
My first thought was to start messing with the IRE100 10 point adjustments, adding blue along the way. It was slow and tedious and I am not sure I was getting anywhere. One step forward seemed like a step back. It wasn't really clear which % measurement actually corresponded to what IRE % setting, I think there was shift/overlap.
Deciding to abandon that approach, I went to the equally setup "isf Bright Room" setting (looked identical with same base settings as "isf Dark Room", so I had a quick A/B comparison). Instead of the granular 10 step or even more so, 22 step, I went with the two point adjustment High and Low.
These settings were much quicker to play with and get instant results across the board.

Spoiler: settings

"ifs Dark Room" or "isf Bright Room"
Backlight: 45
Contrast: 80
Brightness: 25
Sharpness: 10
Colour: 50
Tint: 0
Expert
Gamut: Auto
Gamma: 2.2
White Balance: Warm2
2 Point :
Low Red 1, Green 0, Blue 3
High Red -4, Green 0, Blue 14

Apart from the very lowest dark, I have a measured dE around 1 for the 10% greyscale points. The gamma curve is followed very closely, with only the darkest 3 points being ever so slightly lighter than the curve.

I have work to do with the actual colours themselves, as the primary and secondary scans seemed to show some large dE shifts, and the triangle plot showed blue and red to lack saturation or extension.

For watching it's a pleasant image for sure.

 

[Luminance Leon]

I've just messed around again today.
With Backlight set to 47, the measured brightness is only ~70 cd/m2, so I upped it to 86 which gave 100 cd/m3 with 0-255 input levels, but ultimately felt too bright for us.

We prefer the less bright image, but I can understand the reason and appeal of a brighter white point.

I re-ran the white balance grey scale tests, adjusting low blue to 2 instead of 3 before.
Gamma showed a bit crushing in the lows, so I upped brightness to 30 to avoid milkyness, it's still a little low maybe gamma 2.4 at 20%.

Now, cyan was still a dodgy colour. Skin tones were fine though.
Some variant of "Finding Nemo" was on so I used all of those animated aquatic and sky tones to check out the blue-green area. Yuk! Green banding and not so blue skies. Something was off.

Some messing around in menus, I realised what the colour gamut option is. It needs to be set to extended, even if green-blue colours seemed a bit over vibrant and perhaps somewhat darker initially. Set to auto, it simply makes a mess of green-blue and cyan in particular in, what I think is, remapping of values.

I used a Gretag/Macbeth colour chart image with known good colour values (can't remember where it came from, but it's in my toolbox folder), along with a real colour chart leant up near the screen.
Whilst we don't have calibrated light in the room, using whatever comes through the window, we could see relative perceptual differences.

Using the CMS to manually tweak the RGB Saturation, Tint, and Light values I managed to make the colour chart image closer to the real one in front of me. Whilst I should have been able to just use the HCFR program with the calibrator to adjust colours for accuracy, I could never get the dE low enough on any of them except for Magenta, which was almost spot on after grey scale tweaks.

Spoiler: Settings:

Backlight: 47 (~70 cd/m2)
Contrast: 80
Brightness: 30
Sharpness: 10
Colour: 50
Tint: 0
Expert Controls . . .
Gamut: Extended
Gamma: 2.2
White Balance: Warm2
2 Point . . .
Low: Red 1, Green 0, Blue 2
High: Red -4, Green 0, Blue 14
Colour Management System . . .
Red: S -3, T 10, L 0
Green: S 0, T 0, L -20
Blue: S 3, T -3, L 20
Cyan: S -10, T 30, L -5
Magenta: S 0, T 0, L 0
Yellow: S 0, T 0, L 0

Whilst the picture is almost perfect, there is still a little banding on occasion in the cyan area, where green and blue transition. The included stock image Tahiti shows the band in the sky on the left side.

We'll live with it for now, the cyan banding issues weren't there pre calibration, however the colours weren't anywhere near accurate. It could just be a case that it's not too good around that neighbourhood of colours.