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

[Manni]

Hi Manni, can I ask for a little more info on this, I saw Tim said something similar and not at the same time haha. He says to raise values instead of reduce but then says raising may cause black lift (which coincides with your view).

When adjusting RGB Low values the rule of thumb is to only raise the necessary RGB values to attain Grey Scale Balance, and not reduce values. This is to prevent 0% blacks crushing. As with the RGB High rules of thumb, this is not always the best approach as lifting values may cause black to lift too much.

Is this mainly concerning a 2 point adjustment (not a 12pt?) and what is the correct route to take, always error on dropping offset values when adjusting the low end?

The main rules of RGB balance are:
- Never change more than two values out of three
- Never raise values, only lower, especially for offsets
- You might need to adjust brightness after setting the offsets, and vice versa.
- Never try to resolve 17 if that comes at the cost of raising the black floor. It's fine to crush or only half resolve 17. Raising black is a no-no (if you care about contrast).

Frankly, with the Autocal, there is little use for gamma controls. You just run a gamma autocal (preferably with the Spyder) then as many color ones as needed (preferably with the i1pro2) and then you correct 100% white if needed.

[grendelrt]

12-point is going to be a gamma adjustment as far as I know which is different.  And a gamma adjustment shouldn't be able to affect black.

Usually the first point on 12-point is 5%.

With offset/bias control it seems to affect a gigantic dark range including for some reason black at 0%.

The other control in the 2-point is the gain which can affect 100% white (but also the 12-point gamma can't affect white and the 12th point is usually 95%).

So it seems by my experience that gain/offset(bias) or "2-point" is fundamentally different than the multi-points which have always seemed to be essentially RGB gamma adjustment controls.

Right, I probably should not have mentioned 12pt in this case, I haven't calibrated since March and things get jumbled lol. So if calibrating the gray scale with gain and offset , on the low end you only drop values and try to never raise them?

[SirMaster]

I was considering using the 12-point gamma curve editor in autocal to fix my calibration after autocal, but I tihnk maybe I just need to try running it all again.

For some reason on my latest autocal I did not get very good results when I went to measure them after.

Stock:
https://nicko88.com/misc/nx5cal/stock.pdf

Autocal:
https://nicko88.com/misc/nx5cal/autocal.pdf

3DLUT:
https://nicko88.com/misc/nx5cal/3dlut.pdf

Autocal was run with Spyder5 and i1Pro2.

Measurements were done with i1D3 trained to the i1Pro2 (in the same configuration as used for the autocal).  3DLUT was also done with the i1D3 (via DisplayCAL) in the same configuration as used for the measurements.


At this point though I am thinking I will just wait for firmware 3.10 before I go redo autocal.

I am hoping that it also fixes my issue of my DCI color profile not accepting an autocal.  Because I would prefer to generate my 3DLUT against DCI color profile rather than HDR color profile (BT.2020).

I wonder though, do you think if I create a custom color profile set to DCI color space that maybe that would accept my autocal?  Any downsides to using a custom DCI color profile over the stock one?

[SirMaster]

Right, I probably should not have mentioned 12pt in this case, I haven't calibrated since March and things get jumbled lol. So if calibrating the gray scale with gain and offset , on the low end you only drop values and try to never raise them?

Personally if doing a manual calibration I prefer olny using the gain control to set 100% white and then much prefer using the multi-point RGB gamma controls to set the white balance rather than try to do it all with gain and offset/bias.

In HCFR for instance you can enable to show the greyscale with gamma, so you are calibrating both the gamma and white balance at the same time bringing all 3 colors to 100%.

[grendelrt]

Going to try autocal for the first time tonight. I read through the instructions (thank you!) and had one quick question. I see where you set your brightness, I have mine all set to -2 currently on a RS2000 to get the right black floor. When running autocal, do you use the brightness you have set or do you set it back to 0 prior? 

[Manni]

Going to try autocal for the first time tonight. I read through the instructions (thank you!) and had one quick question. I see where you set your brightness, I have mine all set to -2 currently on a RS2000 to get the right black floor. When running autocal, do you use the brightness you have set or do you set it back to 0 prior?

Keep the brightness as it's set. Most user settings are ignored during autocal. For example, your RGB offsets and gains are ignored (color temp settings), so you can leave them as they are in theory, but it's recommended to reset them so that 1) the post-cal log in the autocal is correct and 2) you don't forget to check and correct the gains afterwards, instead of having the "old" gains or offsets applied, as this will most of the time cause wrong results.

It's also recommended to use a 2.2 gamma target so you can see if your calibration is correct in the Autocal post cal log, as the "flat" line is always 2.2. Any other target, it's going to be difficult to know from the Autocal log if your gamma calibration went well or not. As all gamma curves are calibrated at the same time, you can simply select the best gamma preset after calibration.

[Killroy™]

I had already done one color calibration for BT.2020 with my i1Pro2 and it looked great but I had noticed some highlight crush and blooming. Could have been my calibration, could have been my settings with MadVR but the latter looked great on my other set up so I did not think that was it.

I took madshi's suggestion to try a DCI-P3 calibration so I re-ran autocal on a separate user setting that I had already done a gamma and I can honestly say that I prefer the DCI calibration better than BT2020. The only way I can describe it is that it calmed everything down. The saturations looked far more natural with UHDs and even with BDs they looked awesome. I think this will be my new standard setting.

BTW, the JVC AutoCal is just freaking awesome. SO easy and quick. Can't wait to see if the new tone mapping firmware brings anything as close to MadVR for my other sources.

[Manni]

I had already done one color calibration for BT.2020 with my i1Pro2 and it looked great but I had noticed some highlight crush and blooming. Could have been my calibration, could have been my settings with MadVR but the latter looked great on my other set up so I did not think that was it.

I took madshi's suggestion to try a DCI-P3 calibration so I re-ran autocal on a separate user setting that I had already done a gamma and I can honestly say that I prefer the DCI calibration better than BT2020. The only way I can describe it is that it calmed everything down. The saturations looked far more natural with UHDs and even with BDs they looked awesome. I think this will be my new standard setting.

BTW, the JVC AutoCal is just freaking awesome. SO easy and quick. Can't wait to see if the new tone mapping firmware brings anything as close to MadVR for my other sources.

If you use madVR (or the Radiance Pro), DCI-P3 is recommended indeed, especially if you use a 3D LUT after the Autocal, as long as you tell madVR that your display is calibrated to P3 or you use a P3 3D LUT.

BT2020 3D LUTs can cause a lot of posterization and don't give any advantage on our models as they don't reach beyond P3.

For consumer sources, BT2020 saturations is the only option, as they can't output DCI-P3.

Highlights crush and blooming in HDR is usually due to the DI, try disabling it, especially if the manual iris is fully open. The more open the manual iris, the more artifacts (at least on rs1000/2000, the rs3000 seems to be exempt).

There shouldn't be (m)any artifacts with the manual iris set to -10 or below. hopefully the upcoming f/w will improve this. I heard it might improve the yellowing, which would be great if true.

[Sittler27]

Did a JVC gamma+color autocal tonight on all my used modes (low/high lamp; with/without filter; iris positions) and then proceeded to use HCFR and my i1DPro to fine tune off screen my SDR BT2020 mode.
I also set my screen type.
Brightness @ -3; contrast @ 1

My gamma is set at 2.4.

Required a -8 deduction in green gain and couple more less major adjustments in blue/red offsets.  
Does that sound about right or too much?

When viewing a dark HDR movie (madvr is outputting SDR to this mode), I find I'm losing some details in dark scenes somehow - they seem too dark/losing shadow detail?
I'm wondering, should I be either setting gamma at 2.3; adjusting one of the gamma controls up; or, just increasing brightness to -2 instead of -3?
I can barely see 16 flash on Brightness -2 from my seating position, and can actually see 17 better @ -2.

[Sittler27]

I setup 4 user modes as per Manni's setup guide posts,  I'd like to use Natural for TV & Custom1 for film, as you have indicated, but if I calibrate gains/offsets on 6500 color profile for the SDR BT2020 mode (after autocal with HCFR+i1DPro), which other color profiles (7500? Custom?) should I use to adjust offsets/gains for the Natural and SDR Rec709 modes?  Does it matter?

I tried measuring and adjusting SDR Rec709 mode with Custom color profile and it seemed I needed quite a bit of boost in blue offset to get it even close to true 6500K.  Does that sound normal? 

[Manni]

Did a JVC gamma+color autocal tonight on all my used modes (low/high lamp; with/without filter; iris positions) and then proceeded to use HCFR and my i1DPro to fine tune off screen my SDR BT2020 mode.
I also set my screen type.
Brightness @ -3; contrast @ 1

My gamma is set at 2.4.

Required a -8 deduction in green gain and couple more less major adjustments in blue/red offsets. 
Does that sound about right or too much?

When viewing a dark HDR movie (madvr is outputting SDR to this mode), I find I'm losing some details in dark scenes somehow - they seem too dark/losing shadow detail?
I'm wondering, should I be either setting gamma at 2.3; adjusting one of the gamma controls up; or, just increasing brightness to -2 instead of -3?
I can barely see 16 flash on Brightness -2 from my seating position, and can actually see 17 better @ -2.

If your SDR BT2020 is properly calibrated, it doesn't matter which gamma it uses as long as you tell it to madVR in the calibration tab. If you use a 3D LUT with madVR, it should have a target of 2.2 (it doesn't matter what the display is calibrated to, you can leave it to 2.4 if you want it to work with other sources). You shouldn't have that problem with dark scenes, as long as you use dynamic tonemapping and the latest test build, with the correct settings. Please let's not discuss madVR here, it will derail the thread very quickly . All questions about the test builds in the madVR thread where they are posted, as they are not meant for the end user and settings change daily.

Back to topic. You shouldn't see 16 flash. 16 is black. It should NEVER flash. So use whicheve brightness setting that has 16 not flash and either 17 or 18 flash. Do not make 17 flash if it raises 16 and makes it flash too. That will kill your contrast.

-8 on green sounds a bit too much if you did the color calibration with an i1pro, but you shouldn't set your screen type when using the i1pro, because it reads off the screen so the screen is already taken into account, that's the most probable reason. It should be explained in the basic section, if it isn't please let me know and I'll correct it.

[Manni]

I setup 4 user modes as per Manni's setup guide posts,  I'd like to use Natural for TV & Custom1 for film, as you have indicated, but if I calibrate gains/offsets on 6500 color profile for the SDR BT2020 mode (after autocal with HCFR+i1DPro), which other color profiles (7500? Custom?) should I use to adjust offsets/gains for the Natural and SDR Rec709 modes?  Does it matter?

I tried measuring and adjusting SDR Rec709 mode with Custom color profile and it seemed I needed quite a bit of boost in blue offset to get it even close to true 6500K.  Does that sound normal?

I use 6500K (factory preset) in my SDR wide gamut calibration, and then I use a different custom color temp with a 6500K base for each separate calibration. Natural and rec-709 can use the same one as they should be sharing the same lamp/filter settings.

Make sure you don't set the screen type as you use an i1pro off the screen for color autocal, and see if it improves things. There shouldn't be any significant changes to color temp after an autocal with an i1pro.

[dlinsley]

I completely re-did my autocal, following the steps (almost) religiously. I think I went wrong previously by missing the need to reset offset/gain before autocal, and set them back to all zeros and also switched my gamma to 2.2 so that I see autocal results as it sees it.

I started with color only using the i1pro2 (EFI-ES2000), and then switched to gamma with the Spyder 5. The pictures are from Calman with the Discus trained to the i1 using the same low light 500 nits settings and long integration etc - note I switched back to 2.4 gamma that I use. In the After Autocal result below you can see the RGB balance has divergent green and blue, red almost perfect, but after setting gains to -3/-9/0 (offsets all zero) the result is almost perfect. Contrast is a little down from last time, but this was at -6 compared to -7 before.

I'm not quite hitting green or cyan in BT709, and these results were with the iris at -6, so I may try my SDR setting with the filter in place as I have some light to spare.

[Sittler27]

If your SDR BT2020 is properly calibrated, it doesn't matter which gamma it uses as long as you tell it to madVR in the calibration tab. If you use a 3D LUT with madVR, it should have a target of 2.2 (it doesn't matter what the display is calibrated to, you can leave it to 2.4 if you want it to work with other sources). You shouldn't have that problem with dark scenes, as long as you use dynamic tonemapping and the latest test build, with the correct settings. Please let's not discuss madVR here, it will derail the thread very quickly . All questions about the test builds in the madVR thread where they are posted, as they are not meant for the end user and settings change daily.

Back to topic. You shouldn't see 16 flash. 16 is black. It should NEVER flash. So use whicheve brightness setting that has 16 not flash and either 17 or 18 flash. Do not make 17 flash if it raises 16 and makes it flash too. That will kill your contrast.

-8 on green sounds a bit too much if you did the color calibration with an i1pro, but you shouldn't set your screen type when using the i1pro, because it reads off the screen so the screen is already taken into account, that's the most probable reason. It should be explained in the basic section, if it isn't please let me know and I'll correct it.

When ensuring 16 doesn't flash and 17 does, is this as per what I see from my viewing position - or when I'm up close to the screen?
It's literally -2 Brightness from my viewing position (knowing up close 16 will flash if I walked up to screen to view).  If I go to -3 from my viewing position I can barely see 17 flash.

Regarding the color calibration - I did i1dpro off screen - I did JVC autocal with Spyder5Pro facing lens before.  I set my screen type because I used the Spyder5Pro.  Does the fact I used the i1dpro+HCFR to fine tune after mean I should NOT be setting the screen type?

[Sittler27]

I use 6500K (factory preset) in my SDR wide gamut calibration, and then I use a different custom color temp with a 6500K base for each separate calibration. Natural and rec-709 can use the same one as they should be sharing the same lamp/filter settings.

Make sure you don't set the screen type as you use an i1pro off the screen for color autocal, and see if it improves things. There shouldn't be any significant changes to color temp after an autocal with an i1pro.

Ok, that's what I'll do too.

Just to be clear:  I use a Spyder5Pro facing projector for JVC autocal of gamma+color, then I fine tune with HCFR+i1DPro off screen.  I had thought that the i1DPro doesn't work with JVC autocal (otherwise, I'd use that for it and then the fine-tuning off screen).

Finally - all of this calibration is a bit premature as I only have 73 hrs on the bulb (just getting practiced up for doing it a bit later once bulb burns in).  On that last point, how many bulb hrs exactly should I be doing the real calibration after?