AVS Forum banner

1301 - 1320 of 1453 Posts

·
Registered
Joined
·
5,351 Posts
What it does show though is that it is really beneficial to profile your meter to a Spectro so that you have some sort of reference.
SIGH....yea, that's what I'm afraid of.
 

·
Registered
Joined
·
79 Posts
I really wish xrite added the capability for people with high-end spectros to actually upload CMF offsets, it would probably create a market where corrected i1d3 could be upsold for more $$$ but they probably have their reasons for doing it this way.

On another note your i1d3 whitepoint done with the older probe matches almost exactly what my eyes see when I did a photo using my expensive DSLR camera in my room illuminated by CRI 95+ D65 lights when viewed on the OLED, so I'm starting to wonder if the match done to the laptop isn't so accurate after all.

In all honesty I really don't think it makes too much of a difference as the two that were close together were outliers. They all read differently but within that range. I could post each individual measurement, but it would be useless as there is no way to know if any other particular meter would match it. It would be just as useful to take a stab in the dark and pick a random value within the given range of x 0.310-0.314 and y 0.330-0.334 (with the WOLED EDR active).

What it does show though is that it is really beneficial to profile your meter to a Spectro so that you have some sort of reference.
Hmm x 0.310-0.314 and y 0.330-0.334 doesn't seem to be that big of a range really especially with only needing to touch only G/B channels, it's quite possible our eyes will adapt to any whitepoint inside that range.
 

·
Registered
Joined
·
1,081 Posts
I was doing a real case. The final reading is there, with a reference spectro, for assessing its quality, not for measuring (I should have run the usual 10 cube I do for that). But what I see as tendence is likely true.

as I mentioned before it can be hard to read your posts, possibly b/c English is not your native language - hence I'm trying make sense what you write and/or correct so that we can all be on the same page.

but yes, I do "real cases" and test scenarios :):):) - either way what I wrote applies 1000% true. The validation sample size is much more important than the profile size, and many folks cut on the validation b/c they think they see a small grid as an indication of what is going on... the numbers can shift (sometimes quite a bit) when you do proper validation - in either direction.

I'll give u an example:
we have high-end, super linear screens here used for professional color grading. Some of them are so linear that u can calibrate them very accurately with a simple LS greyscale profile. i.e. 21 GS points + a single R|G|B read. The LS cLUT calibration result from that is better or at least as good until u start profiling w/ 4K points or higher. (read the last sentence again)

How do we know all of this ?

b/c we properly validate w/ very large patch sets to find irregularities causes by either the color engine (!) used to create the cLUT or the screen's response. if using a smaller val patch set, many things would never pop up. in addition, not only for screens used in color grading application but also all end consumer screens, there's certain part of the gamut that always has a higher priority, for example memory colors - #1 priority w/ greyscale, although the latter can be slightly contaminated w/ less perceptual issues. so when u evaluate probe offset approaches, u obviously wanna make sure that ur priority colors are doing very well in whatever probe offset approach u pick for a given display tech...

For comparing volumetric you have to do two measurements, one with data not correct (that then you can correct to TMC and FCMM) and one with volumetric. The cannot be identical, despites all cautions you take. That was my point ....
if (?) I understand what you're writing here correctly, then no, u don't have to validate twice.

btw, volumetric does not refer to what LI is doing, we've been discussing this as an approach 8 years ago - it was just unnecessary as FCMM proved very usable and sufficient on the now "older" display technologies. Whether LI's implementation of volumetric is proprietary (meaning: their own scheme) or not, u can ask them - I'm glad the option is finally in LS/CS.

... but, u'd be surprised and (often) horrified what u find once u start properly validating anything and everything....
 

·
Registered
Joined
·
1,081 Posts
I really wish xrite added the capability for people with high-end spectros to actually upload CMF offsets, it would probably create a market where corrected i1d3 could be upsold for more $$$ but they probably have their reasons for doing it this way.
if u're looking for ref probe offsets, argyll/displaycal has had this functionality for years, but your problem is you don't know...

(1) who measured the screen (--> user error), incorrectly positioned, not properly warmed up etc etc etc
(2) if the spectro was actually within acceptable tolerance (--> as in: is it serviced yearly ?)
(3) the exact model of the screen (not just display tech, e..g Plasma)
(4) and the age of the screen and the settings on the screen as that will also possibly affect probe readings
(5) was readout area of screen the same for both probes

so these offsets from other folks could do more harm than good...
 

·
Registered
Joined
·
1,081 Posts
@Iron Mike could you send me your best match (bdps and csv used for it) and verification, please?
I don't use bpds. what is the csv for ? u mean from Klein ?

re CRI vs. Jeti probes, one thing I wanted to mention and I'm sure it's been mentioned before:
most of H'wood uses CRI probes to cal their screens - NOT Jeti. Jeti uses a different reference lamp for probe calibration than PR/CRI does, hence readings for the exact same Illuminant can result in slightly different spectrum result (and therefore XYZ sum, and possibly xy, depending on rounding precision - although if rounded to only 4 values only the 4th digit may differ - if u round to 5 or 6 u'll def see deviation).

So, if ur intend is to perceptually match H'wood content than CRI is ur best bet - that goes especially for content from the big studios. Now for European content, it's the other way around and Jeti may be "better".


... just thought I throw it out there, u know where THE FORCE is............................................................................................... and yes, Disney uses CRI.... :rolleyes::cool::D
 

·
Registered
Joined
·
17,806 Posts
I don't use bpds. what is the csv for ? u mean from Klein ?

re CRI vs. Jeti probes, one thing I wanted to mention and I'm sure it's been mentioned before:
most of H'wood uses CRI probes to cal their screens - NOT Jeti. Jeti uses a different reference lamp for probe calibration than PR/CRI does, hence readings for the exact same Illuminant can result in slightly different spectrum result (and therefore XYZ sum, and possibly xy, depending on rounding precision - although if rounded to only 4 values only the 4th digit may differ - if u round to 5 or 6 u'll def see deviation).

So, if ur intend is to perceptually match H'wood content than CRI is ur best bet - that goes especially for content from the big studios. Now for European content, it's the other way around and Jeti may be "better".


... just thought I throw it out there, u know where THE FORCE is............................................................................................... and yes, Disney uses CRI.... :rolleyes::cool::D
Every now and then I get the itch to get a CR300. That itch always goes away as soon as I use my tried and true PR-670.
 

·
Registered
Joined
·
4,896 Posts
I really wish xrite added the capability for people with high-end spectros to actually upload CMF offsets, it would probably create a market where corrected i1d3 could be upsold for more $$$ but they probably have their reasons for doing it this way.
I'd bet the onboard EEPROM is too small to add anything but the little that's already written to it during production. Even X-Rite includes their EDRs in software, not in the hardware.
 

·
Registered
Joined
·
1,081 Posts
Every now and then I get the itch to get a CR300. That itch always goes away as soon as I use my tried and true PR-670.

hah ! my itch went away once I upgraded to the CR300 :rolleyes::D:D:D

for narrow band light sources, it will make a difference in the resolved SPD obviously and then from there in x, y deviation up to the 3rd digit, most often "just" the 4th or lower...

since u know what's coming display tech wise, u'll have no other choice but to join me in this sh%t, D-Nice !!!! 5nm is for b%tches !!! :cool:
 

·
Registered
Joined
·
17,806 Posts
hah ! my itch went away once I upgraded to the CR300 :rolleyes::D:D:D

for narrow band light sources, it will make a difference in the resolved SPD obviously and then from there in x, y deviation up to the 3rd digit, most often "just" the 4th or lower...

since u know what's coming display tech wise, u'll have no other choice but to join me in this sh%t, D-Nice !!!! 5nm is for b%tches !!! :cool:
LMAO. You’re right. I’ve also been looking at the PR-740. Far pricier than the CR300.
 

·
Registered
Joined
·
1,081 Posts
LMAO. You’re right. I’ve also been looking at the PR-740. Far pricier than the CR300.
so what additional advantage would you personally get from the PR-740 over the CR300 ?

it's 2nm in highest optical bandwidth config as well. I like that it has built-in battery, bluetooth, and a color display for out in the field measurements - none of these matter for display cal IMO - but nice nonetheless as I use the probes often out in the field for measurements of Illuminants and/or light sources. The bluetooth can be nice for cal'ing PJ's and being further away from the probe but I don't trust the lag in a profile session - been through this a million times with tethered probes but measuring screen over network - the lag can mess up the profile often in a way that u won't directly notice...

But, the CR300 is smaller (at least I'm assuming that, looking at the freaking nuke box that 740 is :eek:), hence more portable, can be powered via USB from a phone, and the free CRI android/ios app can take measurements anywhere. I've asked them to implement a feature for me that I can save the full SPD to a txt file on the phone, and they'll implement it - this is like the SD card feature of the PR740.
 

·
Registered
Joined
·
970 Posts
I don't use bpds. what is the csv for ? u mean from Klein ?

re CRI vs. Jeti probes, one thing I wanted to mention and I'm sure it's been mentioned before:
most of H'wood uses CRI probes to cal their screens - NOT Jeti. Jeti uses a different reference lamp for probe calibration than PR/CRI does, hence readings for the exact same Illuminant can result in slightly different spectrum result (and therefore XYZ sum, and possibly xy, depending on rounding precision - although if rounded to only 4 values only the 4th digit may differ - if u round to 5 or 6 u'll def see deviation).

So, if ur intend is to perceptually match H'wood content than CRI is ur best bet - that goes especially for content from the big studios. Now for European content, it's the other way around and Jeti may be "better".


... just thought I throw it out there, u know where THE FORCE is............................................................................................... and yes, Disney uses CRI.... :rolleyes::cool::D
I thought (erroneously?) you verified a volumetric match with 1.2k patches, that's why I asked for bpds and the patchset (CSV) been involved. Nevermind.

Does Disney uses LightSpace or CalMAN?
 

·
Registered
Joined
·
1,081 Posts
I thought (erroneously?) you verified a volumetric match with 1.2k patches, that's why I asked for bpds and the patchset (CSV) been involved. Nevermind.

yes, wrong.

no bpds. volumetric probe offsets. validation was custom patch set. I'm assuming u're talking about a LS profile (?), that would be .bcs file. did not use LS.
 

·
Registered
Joined
·
17,806 Posts
Wow, I wake up and see I probably caused this thread to derail with my CR300 comments last night. Iron Mike, we can take that discussion offline if you like.

For those who do not know who Iron Mike is, I do and can say he is a good guy, knows his stuff and has my respect. I agree with him 100% regarding validation and the need to use more than 1000 points when doing 17^3 and 21^3 LUTs (It would also be beneficial with FCCM vs. TCM vs. MPVL ;)) A few 100 just wont cut it.

I also understand why he will not/cannot provide the files for you Miki. It’s nothing personal, it’s business. However Miki, I will help you where possible.
 

·
Registered
Joined
·
1,081 Posts
I also understand why he will not/cannot provide the files for you Miki. It’s nothing personal, it’s business. However Miki, I will help you where possible.

no, it's not LS files - that part was all custom to cut time. I'll re-run the offsets at one point w/ LS, I can share that.
 

·
Super Moderator
Joined
·
17,795 Posts
Off-topic, bickering and profanity removed. Infractions issued.

Discuss the topic and not each other. Please read the topic title and be sure that's what your post is all about.

Questions? PM me. Don't post here.

Doc
 

·
Registered
Joined
·
472 Posts
@Iron Mike, only for clarification, as I don't ahve disagrements about what you have written:
as I mentioned before it can be hard to read your posts, possibly b/c English is not your native language - hence I'm trying make sense what you write and/or correct so that we can all be on the same page.

but yes, I do "real cases" and test scenarios :):):) - either way what I wrote applies 1000% true. The validation sample size is much more important than the profile size, and many folks cut on the validation b/c they think they see a small grid as an indication of what is going on... the numbers can shift (sometimes quite a bit) when you do proper validation - in either direction.

I'll give u an example: ...
First of all mny thanks for the good and interesting example.
When I have written "real case" I meant a "real calibration workflow". There was no reference to you cases.

Yes, I share your view about importance of validation samples, but from my limited verification (I have called a quality assessment), I see avg TMC dE slightly better than FCMM one and several FCMM points better than TMC ones. The first, off course will deviate with a valid sample, the second not.


PS: I have added to my excel file also FCMM and TMC dY, dx dy and dxy to references.


... but, u'd be surprised and (often) horrified what u find once u start properly validating anything and everything....
and unluckily there are stuff you cannot fix, as my panel ....
 

·
Registered
Joined
·
1,081 Posts
When I have written "real case" I meant a "real calibration workflow".
well, they are all real cases and/or usable in a real calibration workflow IMO - maybe I'm not understanding what u mean.

The first, off course will deviate with a valid sample, the second not.
what do u mean by "valid sample" ?

So one other thing I wanted to mention, while it's good that u ran this specific test, it's overall result (while def interesting) is of less importance b/c u used the spectro as the validation probe. For the "best possible" result u have to use the same probe used as for the main profile (w/ same probe characteristics), used in the same place reading the same pixels on the screen. Spectro, reading different pixels now, will have less or different repeatability than the colorimeter w/ different sensitivity in the shadows.

Only reason I'd measure some patches w/ spectro afterwards on the cLUT (for this particular workflow) is to ensure that the cLUT is actually somewhat accurate, which u'd only do if u're doubt of the math that was employed to calculate probe offsets or the math used by the color engine - this would just be a last sanity check. As long as dE is lower than what it was before cLUT u know the math was not completely off (but this doesn't tell u if better math could have resulted in an even better cal...).

but for direct comparison, u need to eliminate as many variables as possible that distort the final numbers, not introduce more.

so what panel was your test done on and what is the rest to your last line regarding that your panel is defective ?
 

·
Registered
Joined
·
970 Posts
Only reason I'd measure some patches w/ spectro afterwards on the cLUT (for this particular workflow) is to ensure that the cLUT is actually somewhat accurate, which u'd only do if u're doubt of the math that was employed to calculate probe offsets or the math used by the color engine - this would just be a last sanity check. As long as dE is lower than what it was before cLUT u know the math was not completely off (but this doesn't tell u if better math could have resulted in an even better cal...)
That is exactly why he did verify with his spectro. He probably used a small cube to preserve his spectro integrity as he had a bad experience in the past.
 

·
Registered
Joined
·
1,081 Posts
That is exactly why he did verify with his spectro. He probably used a small cube to preserve his spectro integrity as he had a bad experience in the past.
right, but as I explained, that is a failsafe verification. These dE numbers are less meaningful as the ones he would have gotten w/ the colorimeter still in place.

or do u run validation of a cLUT w/ a spectro ? no, u use the colorimeter w/ the same probe offsets as used for main profile. u keep the same conditions as in the main profile.

in any case, looking at the dE numbers clearly indicates the math worked (to some degree), would be good to know what the dE numbers would have been the proper way, b/c the whole point of this exercise is that the former dE reports (using FCMM on WRGB) were misleading... now - w/ a proper validation profile and a proper probe offset method - u'd have a better idea how the calibration performs.
 

·
Registered
Joined
·
970 Posts
Maybe we are talking of different things, ebr9999 is not verifying his panel calibration but what is the best probe matching method for his panel, which means what method is closer to spectro measurements. Of course, that is the theory that has to be correctly applied, and on this last subject you are completely right: meters have to read the same exact pixels every time in the same ambient/panel conditions.
 
  • Like
Reactions: mrtickleuk
1301 - 1320 of 1453 Posts
Top