JVC Autocal Software V12 calibration for 2019 RS1000/RS2000/RS3000/NX5/NX7/NX9

[Manni]

Reserved [Under Construction]

[Manni]

OOTB and basic Rec-709 calibration measurements at 30 hours:

I attach a ZIP with measurements of my rec-709 mode (User1, Rec-709 color profile, 6500K, low lamp, no filter, iris -12). These were taken at about 30 hours on the lamp.

I get 60nits with the iris at -12, and this leaves some headroom for calibration as I like to end up with 50-55nits calibrated.

All measurements taken with my Discus trained to my i1pro2. The Discus used under the right conditions reads very close to my Minolta T10, so I trust it even for black/contrast measurements. For these, I used the i1pro2 on its own for all gamut and color volume measurements, and the Discus trained to it for gresycale and contrast. Don't pay attention to the meter on the screenshots, I don't take the screenshots at the same time as the measurements, so either meter could be selected then.

The only thing I did before taking these is set 100% white to D65. I then selected a gamma 2.4 preset and adjusted picture tone gamma to +3 and dark gamma to +3 to get a rough BT1886 curve (specific to my unit, so unlikely to work the same on yours). I never use the CMS.

These pre-cal measurements are taken before any Autocal is run. When I use Autocal I usually don't bother with picture tone, I just set dark gamma afterwards, but I wanted to show what you could get in 5mn without using the Autocal.

Of course gamma and RGB balance could be made better adjusting offsets and/or using the 12-point gamma controls in the Autocal software if you don't like the idea of running the Autocal itself.

I include gamut saturations, greyscale (before and after dark/picture tone gamma adjustments), contrast at -12 with iris set to manual and auto2, and color volume measurements. I also add for comparison a measurement of the Natural user preset, which is very close to my user1 and tracks very well but for some reason is a bit less saturated (96.84% of rec-709 gamut cover vs close to 97.98%). I'll add a colorchecker in the next set.

I think they are very close to perfect for an OOTB state. Certainly good enough for me to watch some content before running an Autocal and a 3D LUT. Kudos to JVC for delivering such a good calibration OOTB. 100% of rec-709 gamut volume, with near perfect saturations and very decent greyscale, I'm happy with that for now.

I'll post measurements after the Autocal once I've had a chance to play with it.

[Manni]

No detailed measurements yet but I thought I'd post gamut and brightness/contrast measurements for the BT2020 Color Preset (BT2020 with the filter), so you can compare gamut cover increase and brightness loss compared to the HDR preset (without the filter).

As expected, the BT2020 profile tracks BT2020 saturations, as BT2020 is the container for consumer sources/content.

I measured P3 within BT2020 to give a better idea of how it tracks P3 content, which is the most common at the moment.

I did these measurements in high lamp because historically that's what gives the largest gamut cover (and that's usually what people use for HDR content).

Using the filter cuts around 10% of brightness (164nits vs 183nits without the filter) in order to get close to full P3 cover (99% with the filter, 85% without). The contrast is also higher in BT2020 than in HDR (around 30,000:1 vs 24,000:1). I posted all the new contrast and brightness measurements in the contrast section.

So interestingly, it looks like the filter cuts less light in high lamp (10% cut / 15% gain) than in low lamp (20% cut / 25% gain), and increases the contrast compared to HDR.

I would think that my unit is at the low end regarding gamut cover, I would expect a bit above 100% with the filter and around 90% without.

[Manni]

No detailed measurements yet but I thought I'd post gamut and brightness/contrast measurements for the HDR Color Preset (BT2020  without the filter), so you can compare gamut cover decrease and brightness gain compared to the BT2020 preset (with the filter).

As expected, the HDR profile tracks BT2020 saturations, as BT2020 is the container for consumer sources/content.

I measured P3 within BT2020 to give a better idea of how it tracks P3 content, which is the most common at the moment.

Losing the filter bring back around 15% of brightness (183nits vs 164nits) at the cost of 15% of saturation (99% of P3 with the filter, 85% without).

I would think that my unit is at the low end, I would expect a bit above 100% with the filter and around 90% without.

[Manni]

OOTB and basic DCI-P3 calibration measurements at 30 hours:

I attach a ZIP with measurements of my DCI-P3 mode (User3, DCI-P3 color profile, 6500K, low lamp, with filter, iris open). These were taken at about 30 hours on the lamp.

I get 120 nits with the iris open and the filter, which is more than enough for HDR content tonemapped by MadVR, as this is what I use this calibration for. It's relevant even for non-MadVR users because it shows what the color volume is for P3 content when using a BT2020 or HDR colour profile. I'd get around 150nits (25% increase) without the filter, but I only get 85% of P3 that way (vs 97%) so I prefer using the filter. The filter cuts around 20% of brightness on my unit.

As usual, all measurements taken with my Discus trained to my i1pro2. The Discus used under the right conditions reads very close to my Minolta T10, so I trust it even for black/contrast measurements. For these, I used the i1pro2 on its own for all gamut and color volume measurements, and the Discus trained to it for gresycale and contrast. Don't pay attention to the meter on the screenshots, I don't take the screenshots at the same time as the measurements, so either meter could be selected then.

The only thing I did before taking these is set 100% white to D65. I then selected a gamma 2.2 preset (as this will be the target of my MadVR 3D LUT). I didn't adjust anything except the gains in the 6500K preset. I never use the CMS. It took 3 minutes of adjustments (if that) to get these results.

These pre-cal measurements are taken before any Autocal is run.

Of course gamma and RGB balance could be made better adjusting offsets and/or using the 12-point gamma controls in the Autocal software if you don't like the idea of running the Autocal itself.

I include gamut saturations, greyscale, contrast with the iris set to manual and auto2, and color volume measurements. I'll add a colorchecker in the next set.

I'm happy with the gamut saturations and the greyscale, I think it's very good for an OOTB after just a white point adjustment. I would have liked red to go a bit further, but it seems to be a common occurrence on the new models to have red (instead of green and cyan on older models) falling a bit short.

Still, 97% of P3 gamut cover and 99% of P3 color volume, with saturations tracking very well, I'm happy with this.

I'll post measurements after the Autocal once I've had a chance to play with it.


Measurements after Calman Lightning LUT

I was hoping that I could live with my quick and dirty basic calibration for a while but the greyscale errors in the low end were too visible. I remeasured today and they were above 6dE.

As I don't have the time at the moment to dive into the Autocal, I decided to run a Lightning LUT with Calman (101 points, it takes only around 10 minutes) as the gamut is very linear. In the past, that gave me very good results when the baseline is linear. So when my rs500 was under 1000 hours, that's what I used. After that, I had to use progressively larger LUT as it became less linear as I put 3000 hours on the clock.

Measurements are much better, as expected, especially for greyscale. I ran a colorchecker SG (with/without the MadVR Lightning LUT), and I add updated before/after for today's session. I try not to run a colorchecker SG if I try to avoid a proper calibration to not influence me, but this time it was quite obvious a bit more was needed.

I haven't had a chance to check for visual issues (especially posterization near black or near the edge of the gamut), but I'll report on that later if I find any.

[Manni]

Reserved [Under Construction]

For now see this post.

[Manni]

Reserved [Under Construction]

[Manni]

Reserved [Under Construction]

[Manni]

Reserved [Under Construction]

For now see Sirmaster's post here.

[Manni]

Reserved [Under Construction]

[Manni]

Reserved [Under Construction]

[Manni]

Here are my results if I apply a madVR 3D LUT on top of the calibration obtained after the JVC Autocal (see baseline here). You don't need to run the autocal beforehand, but I like running the autocal and applying a LUT on top because it allows me to use the autocaled calibration with my non-HTPC sources, and apply the madVR LUT on top for HTPC content.

The results are near reference (well, within the error margin of my i1pro2, which hopefully and visually should be below 3dEs).

At the moment, because the baseline is so linear, I can create my 3D LUT with a very quick profile (Lightning LUT, 101 points in Calman Ultimate).

After a while, I might have to run a larger LUT. When I get there, I'll compare results between Lightspace and Calman. I had issues with both with my slightly undersaturated rs500, so I can't recommend either. But with a nearly fully saturated baseline such as this, I would expect to get good results from either software. I would also include DisplayCAL as an option, because I got very good results with it. Unfortunately, it doesn't support the Discus, otherwise I would most certainly be using it more regularly.

I'll update this post when/if I make an attempt with larger LUTs, but given the current results my motivation to burn hours of bulb time running larger profiles is very low.

Again, the before measurements are available in the basic - calibration section. You would get the same results with a LUT run from the before calibration state, because it's already very linear., at least for now at 165 hours.

[Manni]

Reserved

[Manni]

Reserved

[Manni]

Reserved