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

[Jbarteli]

The answer is the same. If you can wait, wait. If you can find a spyder 5 and need to calibrate, get one while you still can. There is no official news about spyder X supoort. It will come, but no idea when.

Thank you Manni, is it known if the X is superior to the 5?

[Manni]

Thank you Manni, is it known if the X is superior to the 5?

It might be a bit faster and maybe a bit more accurate especially in low light thanks to a larger lens, but I don’t expect huge differences. I plan to buy one as soon as JVC supports them, I’ll report then on any difference. I didn’t see much of a difference between the Spyder 4 and the Spyder 5, so probably similar.

Here is a review from a good website.

[Manni]

According to this other review, the difference in speed is significant, but whether it's the software or the meter, it's hard to say. Accuracy seems improved as well, but unlike the previous website I don't know how reliable this reviewer is. Anyway I'll test it when it's supported and then we'll know if there is any improvement with the JVC Autocal software.

[SirMaster]

@Manni, Do you always face your meters at your screen when calibrating your projector?

Like when you are profiling a colorimeter to a spectro, do you always point both the spectro and the colorimeter at the screen?

And then when you generate the 3DLUT, do you also point the meter at the screen?

I know the Spyder5 faces the lens for autocal, but I'm mainly wondering for everything else.

[Manni]

@Manni, Do you always face your meters at your screen when calibrating your projector?

Like when you are profiling a colorimeter to a spectro, do you always point both the spectro and the colorimeter at the screen?

And then when you generate the 3DLUT, do you also point the meter at the screen?

I know the Spyder5 faces the lens for autocal, but I'm mainly wondering for everything else.

It depends on the meters and the software.

Obviously with the JVC Autocal the i1pro2 faces the screen (color only) and the Spyder faces the lens (gamma, and color if no i1pro2).

With Calman (or Lightspace), I train the Discus facing the screen to the i1pro2 facing the screen, because the Discus can read black or near black easily even when facing the screen. The spectro always have to face the screen to take it into account. When I had a meter unable to read black or near black, such as my old i1display2, chroma 5 or even i1displaypro in some cases, I would usually train the tristim facing the lens to the i1pro facing the screen.

In any case, the meter should always point to whatever it was pointing at when it was profiled to the spectro. You can't move the meter (even slightly) once it's been profiled. If the meter moves (say you bump into the tripod, catch it before it falls if you're lucky, and put it back in place), you have to re-profile. This is true for any measurements, including when profiling the PJ for a 3D LUT. Once profiled to the spectro, the meter doesn't move.

To optimize profiling with the JVCs, I recommend the following:

• Set the iris to fully open to maximize brightness before profiling. You can then go back to your usual iris settings to take measurements.
• In some cases, you might have to reduce the picture size using the zoom (in extreme cases moving the PJ closer to the screen) to get repeatable readings from the spectro, for example on blue or red. Don't change the lamp mode or filter mode as this usually requires a new profile for max accuracy.
• If making only one profile, always make the profile with the widest gamut you want to measure.
• At the beginning of each session, warm-up of both the PJ and the meters for at least 45min before taking any measurements, especially before making a profile. That's assuming a room with A/C or constant temp. If the room temp keeps changing (a small room will heat-up quickly with the PJ on, especially in high lamp, and the temp will keep going up without A/C), there is no way to make accurate measurements as the color temp keeps changing (significantly). The PJ needs about 45 minutes to reach a stable state even in a room with a stable temp, so there is no point running a profile or taking measurements before that. 30min if you're in a hurry, 15min if you're just after ballpark readings. Less than 15min warm-up, you're reading junk.
• Every time you change lamp mode during a session, light to low or low to high, even for a few seconds, you have to wait at least 10-15 minutes before doing a profile or running measurements. That's how long it takes for readings to stabilize. That's one of the reasons why I use low lamp for HDR. It really speeds up calibration to use the same lamp mode for all calibrations if you can.
• Every time you change lamp mode during a session, light to low or low to high, even for a few seconds, you have to wait at least 10-15 minutes before doing a profile or running measurements. That's how long it takes for readings to stabilize. That's one of the reasons why I use low lamp for HDR. It really speeds up calibration to use the same lamp mode for all calibrations if you can.
• Ideally with projectors, if you want max accuracy, make a fresh profile at the beginning of each session. Even if you put the meter at the same spot/angle/distance, there is usually too much drift to get accurate results from one session to the other. If you just want to make quick measurements, then a new profile isn't really necessary.
• Ideally make a new profile not only at the beginning of each session, but also for each calibration if you change lamp mode or filter mode. I run a fresh profile (I simply update the existing ones in Calman) for SDR Rec-709 (low lamp, no filter, iris -12), SDR BT2020 (low lamp, filter, iris open), HDR10 (high lamp, filter, iris open) and 3D (high lamp, no filter, iris open) at the beginning of each calibration session.
• For max accuracy, make sure you match the field of view of the meters, especially if you don't use full size patterns. For example, my Discus is one foot away from the screen, the i1pro2 is two feet away, so that they have the same FOV.
• Make sure both meters read exactly the same area of the screen (center). For this, I select the smallest sized pattern (2%), I display a 100% white pattern and I use the LED in the Discus to aim it exactly at the center of this small area. I then maximize brightness readings to set the i1pro2 to read that area as well. Then I select full size patterns again for profiling/measurements. There is too much brightness and color variation on the PJ panels to get accurate results if you don't make sure that the meters read the same area of the screen as precisely as possible.
• With the JVCs, I highly recommend to use full size patterns, because that's what the Autocal uses. Using patterns of any size except full size increases the mismatch between the Autocal readings and other software reading.
• During profiling, set both meters to the most accurate mode. With the Discus, I use long integration/high average. With the i1pro2 (it also works with all meters with a low light mode), I move the threshold for the low light mode to 500nits and the repeatability to the max samples, so that all measurements are the slowest/most accurate.
• Always check the profile before taking measurements. I like to set the color temp to exactly D65 using the spectro before running the profile. That way, it's easier to check quickly if the profiled meter is close or not with a single read. Then of course do a WRGB check.
• Always run a gamut measurement with the spectro at the end of a calibration to make sure that profiling didn't go wrong at some stage.
• For 3D, use 3D patterns (the PJ has to be in 3D mode and the glasses have to be on and synced) and profile the spectro with the glasses to the colorimeter without the glasses. Make sure you max brightness with iris fully open (and usually high lamp with 3D). You might have to use the zoom during profiling (make the picture smaller) to get repeatable readings from the spectro when reading with the glasses.

You can get decent results without doing all the above, but if you want max accuracy and a profile that's good 99.99% of the time, following each of the above puts the odds in your favor.

Hope this brain dump of meter profiling tips helps!

[riddle]

Thanks Manni, your advice is worth considering, but I still have some minor irregularities. Gamma calibration via Spyder5 is OK but the color is somewhat delayed. I can't really fine-tune it with i1pro2. I compare the result with Calman and i think i make somewhere some mistake. Colors after calibration seem to me somehow scattered.

[Manni]

Thanks Manni, your advice is worth considering, but I still have some minor irregularities. Gamma calibration via Spyder5 is OK but the color is somewhat delayed. I can't really fine-tune it with i1pro2. I compare the result with Calman and i think i make somewhere some mistake. Colors after calibration seem to me somehow scattered.

I'm not sure what you mean with "color is somewhat delayed" or "colors after calibration seem somewhat scattered". Run gamma only with Spyder 5, run color only with i1pro2, post logs/measurements in the Autocal for both and measurements with the Spyder 5 trained to the i1pro2 in Calman (make sure the profile is good) using full-sized patterns for both as well, explaining in detail what you think the issue is. Then we can take a look at the possible issue(s).

[SirMaster]

Wow Manni, you wrote a lot!  That was not expected.

Yes, your information is extremely helpful and I appreciate you taking the time to write it all.

How do I measure my calibration result quality in calman?  Are there any special modes or settings I need to know?

After autocal, my high and low bulbs still look quote different and I would like to try and understand why they are different.  I think looking at the charts that calman can read for me would be a good place to start.

[Manni]

Wow Manni, you wrote a lot!  That was not expected.

Yes, your information is extremely helpful and I appreciate you taking the time to write it all.

How do I measure my calibration result quality in calman?  Are there any special modes or settings I need to know?

After autocal, my high and low bulbs still look quote different and I would like to try and understand why they are different.  I think looking at the charts that calman can read for me would be a good place to start.

I didn't expect it either, but once I got started, I thought why not try to cover all the bases?

I don't know what you mean by "my high and low bulbs still look quite different". Are you talking about the picture itself or the measurements? Do you use a different custom curve or different settings in madVR in each mode to take into account the difference in peak brightness? there are some differences, for example you can usually reach a larger color volume in high lamp than in low lamp, which is one of the advantages of using high lamp, but your statement is a bit vague.

It would really help if you could post measurements, pictures, screenshots so that we can see more clearly what you're talking about.

Yes Calman measurements can help to check a calibration but that's a bit off topic if you're not used to the software. 

I suggest you get used to Calman first (there are a few support threads around), then use the info above to create good profiles, and finally provide measurements taken in Calman showing the issues/differences between low and high lamp so we can keep the discussion on-topic. If anyone else wants to help re Calman, please PM SirMaster so that we don't clutter the thread.

[SirMaster]

What I mean is that my high bulb color looks less saturated and colors overall look lighter than the color in my low bulb.

This is just after doing autocal with the only difference being changing the lamp power.

I don't use any different settings anywhere else as I am really just seeing it by toggling the bulb between high and low in the JVC which is also the only thing that I change between autocals as well (lamp mode).

This was the case when I did color from the Spyder, and is still the case after doing the color with the i1Pro2.

It's not a huge deal.  In isolation I wouldn't really notice the difference, but toggling back and forth it seems that I like the color better in low bulb.

I just thought that the only way I can really understand what is fundamentally different about the color would be to measure it and look at the numbers.  I don't know which is more "correct".

Like you said, maybe if I changed my bulb and also changed the nits in madVR at the same time it wouldn't be as noticeable of a difference.  But I would have expected that to just change the brightness, not so much the color itself. 

I can look at it with SDR content as well and changing the bulb to high doesn't just make the picture brighter, it makes the colors look a different lighter shade.

Maybe it's just how color looks when brighter?  Again I don't know how describe it and thought that a measurement comparison would be good to tell what might actually be different about the color.

I can take pictures as well to try and show the difference I am seeing.

[Manni]

If you change the lamp mode (i.e peak brightness) without changing the settings either in madVR HDR tonemapping or in your custom curves to reflect the higher peak brightness, colors will likely be slightly desaturated. There is a strong link between brightness and saturation in HDR on a projector. The higher the brightness, the more saturation, provided the source/pj/whatever does the tonemapping knows that there is more brightness to play with. If you simply raise the brightness without telling the tonemapping, then you'll likely get a slightly desaturated picture in high lamp.

Anyway, calibrate and set-up properly both modes, and if you still have a difference, then take measurements with madVR. If anything, provided that the calibration and the tonemapping settings are correct, you get more color volume in high lamp, not less, but there is no point comparing the two modes if you just change the lamp mode (even calibrated).

You can try the same in SDR, you shouldn't have the issue (provided the calibrations are good of course) as there is nothing to tune re tonemapping.

[riddle]

I'm not sure what you mean with "color is somewhat delayed" or "colors after calibration seem somewhat scattered". Run gamma only with Spyder 5, run color only with i1pro2, post logs/measurements in the Autocal for both and measurements with the Spyder 5 trained to the i1pro2 in Calman (make sure the profile is good) using full-sized patterns for both as well, explaining in detail what you think the issue is. Then we can take a look at the possible issue(s).

It seems to me that if I release Autocal and calibrate colors with i1pro2, the resulting effect is worse than without it. Maybe this weekend i will have more free time for makes some photos with rezults.

Thank you

[pbiancardi]

Sounds good re the Oppo outputting HDR and the Lumagen doing the tonemapping to SDR BT2020. In that case I don't know what's wrong with your SDR BT2020 calibration. Maybe someone else will have an idea. I haven't used a Lumagen in a long time (since I got rid of my mini-3D) and have never used a Lumagen Pro, so can't help there.

If you change lamp mode or filter mode between 709 and bt2020, they are independant calibrations. So yes, you need to run one gamma (spyder) and one color (i1pro2) for each, assuming you don't want to be able to change the iris setting, otherwise you need to run a color calibration for each iris range that you want supported for each calibration (709 and 2020).

You might want to re-read the basic section, it's all in there.

Question regarding Calman, do I want to do 3d luts for both 709 and 2020, still trying to get comfortable with the software before I do anything beyond reads.  Also do I want to use both meters with Calman, Spyder for grayscale and i1 Pro 2 for 3d lut?  My first read shows that JVC autocal was basically spot on for 709 points....

[Set11]

Hi Manni,

Mind sharing your tips to get i1pro2 to read gamma at low IRE levels?
I keep getting a reading of zero on the first gamma step of Red. (iris is already set to max)

[Manni]

Question regarding Calman, do I want to do 3d luts for both 709 and 2020, still trying to get comfortable with the software before I do anything beyond reads.  Also do I want to use both meters with Calman, Spyder for grayscale and i1 Pro 2 for 3d lut?  My first read shows that JVC autocal was basically spot on for 709 points....

Yes you do need a different LUT for SDR BT709 and SDR BT2020 (it's not advised to use a static HDR 3D LUT because that means you can't use automatic tonemapping, whether static in the JVC/player or dynamic in madVR/Radiance Pro).

Unless you zoom the picture down to its minimal size (and even then, I'm not sure it will be accurate enough), you can't use the i1pro2 for a 3D LUT. It can't read the dimmest patches accurately, this will cause a lot of issues especially near black and with blue. You have to train the Spyder to the i1pro2 and use the profiled Spyder to generate the LUT in Calman. See  a few posts above for tips to profile a meter to a spectro.

A 3D LUT will NOT visibly improve things if you're already spot on after an Autocal. In fact, it can make the picture worse. So I would advise to only make a LUT if there are significant errors that the Autocal and manual fine-tuning can't resolve. If, after setting white to D65 using the i1pro2, you have gamma/gamut (including saturations) under 3dE accross the range, and a colorchecker SG with all points under 3-4, I would say don't bother with the 3D LUT. You're going to waste hours of good lamp time for little to no visible improvement, unless you can get significant improvements with a Lightning LUT (which works well after an Autocal). Remember, if you're picky, you have to redo the calibration every 200 hours or so, due to lamp drift.