Color HCFR Calibration Discussion (Post your calibration files here) - Page 69

My opinion is that you're probably generally within the accuracy of the meter. If you were to switch the colorimeter for another of the exact same model, odds are that it would measure more off than what you're talking about trying to correct for here. So one colorimeter will require a certain set of settings for its best readings and another of the same model will need a slightly different set for its best readings. Due to this inherent inaccuracy of relatively inexpensive colorimeters, I just consider it a bit silly to try to get by the numbers perfect measurements.

pbc, are you referring to the small drop in gamma from 20% to 40% (2.37 to 2.3)? I think you have to remember that these are consumer displays and when you throw in the fact that you are using a lower end instrument to measure, you're not going to get a perfect flat gamma graph across the stimulus range. Your gamma response actually looks quite good to me and so does the grayscale color balance. Like I said earlier the only thing I'd do with grayscale is try to tick blue down in the low end just a little while balancing it with ticking blue up in the high end if necessary. But I wouldn't spend too much time trying to get it perfect in the low end. A trend toward blue in the low end of the grayscale seems to be common (my Samsung LCD does the exact same thing) and is less noticeable (than say red or green) to your eyes anyhow.

And alluring is right - while we all love to tinker and experiment , eventually you'll get within the variance of the meter in question (not to mention whatever variance the display has and the granularity of it's controls) so don't worry too much if the numbers don't end up perfect.

cheers,


--tom

Is the eye one or the so called spyder 1 or 2 the best one?

Do some searces but eye one LT is best bang for buck.

Lots of discussion on the Eye-One vs. Spyder2 and the pros will tell you to stay far away from the Spyder2. Some are good but on average 2/3 of them are not accurate enough to be taken seriously.

See Appensix A: Eye-One vs Spyder2 comparison which is part of my Greyscale & Color Calibration for Dummies guide.

Kal

I have done two runs using the different modes on my PJ. One is the 'video' mode and the other is the 'cinema'. In terms of gamma, I am guessing the cinema is better because it is more linear, even though it is a little lower in the bottom end. Please comment.

 

Cinema gamma.zip 1.10546875k . file

 

video gammachc.zip 1.220703125k . file

I think that Bill from Calman has commented on prefering a high top-end gamma, so I would say this might be one area where it really comes down to user preference. What sticks out to me is that the video setting measures darker for the entire range except 100%. A darker 0% is probably good, if that really exists and isn't related to how the measurements were taken, but some people are likely to prefer the brighter mid-range of the cinema setting.

Quick question regarding brightness. Is the 0.65% of white setting for any gamma, or ideal for a 2.2 gamma? I.e., with a higher gamma (say 2.3/2.4), does that brightness "rule" still apply? Only ask as I find when I use the AVS709 pattern by eye, it seems to call for an extra couple clicks of brightness vs. the measured 0.65% of white? Maybe they have nothing to do with eachother, but thought I'd ask!

pbc,

Gamma_____10% white (in proportion to 100% white)
2.2________ .0065
2.3_________.0050
2.4________.0040

pbc,

That rule is for a 2.2 gamma. If your gamma response is 2.2 at 10% white
then 10% white's Y value will be .63% of 100% white's Y value.

The reason that you are seeing differences in the brightness setting between the two methods is that your display does not have a consistent gamma response at the low end. Using the AVS 709 pattern is the more accurate method since it means that you are not losing any detail at the low end. With the 10% white method, you are probably "crushing" some of the low end into black.

hope this helps,


--tom

While I don't have any new HCFR files to post (yet...), I have some questions before I take a look at my system again:

1. When I calibrate color, should I use the native resolution of the display, in this case 1080p, and calibrate to the HD 709 standard, or the SD 601 of the source (a progressive scan DVD player and a Nintendo Wii).

2. What light output should I aim for on an XBR1 SXRD and a XBR5 LCD? From Tom Huffman's guide, it seems 40ft-L for the SXRD and 50-60ft-L for the LCD...good? Should I go more towards the 50 or 60 for the LCD? My viewing will be during the day in a light controlled, reasonably dim room (not a lot of ambient light) and at night in a darkened room, and I'd prefer to find a happy medium.

3. Just how important is gamma?

4. Should I calibrate the black level with a pattern, or using the .65% calculation? I did the latter before and could still see the grayscale, but was told my numbers were way to low (indeed my brightness setting was four or so below what most in the XBR4/5 thread are using).

Thanks for any help you guys can provide!

This really depends more on the display than the source. I think both of your displays had xy color points (CIE chart) nearer to 709, so if that's the case then just use 709.


This will probably also work for the LCD, but for the SXRD set brightness as low as possible so that you are displaying all information above black. When you set the brightness you can set iris to the max setting in order to see near-black bars if you want.

Once you have set brightness, then set picture as high as possible. When turning up the picture control your SXRD will probably run out of red. To see this happen do a continuous measurement on a 100% white or above-white pattern. As you turn up contrast you're likely to find that there's a point where red no longer increases along with green and blue. Somewhere in the area where that happens is around where picture should be set, but you might need to lower the control a bit below where red stops increasing to get a little more linear grayscale response.

By setting brightness as low as possible and the picture as high as possible that will get you the maximum contrast the SXRD is capable of. Once you have done this, then just set the iris to whatever you prefer (on the LCD you might want to set the backlight when setting picture and grayscale). A low iris setting might obsure details near black if there's much light in the room and a high iris might wash-out blacks if there's little light in the room. Personally I can't stand washed-out blacks so I use a lower iris setting, ignore the 100% white output, and limit the amount of light in the room during the day.


It just relates how bright different shades are relative to other shades. Gamma is sort of like contrast in a way. Contrast tells you how much brighter white is than black, and in that same sort of way gamma can relate how much brigher say 50% gray is than black. A higher gamma would mean that for the same white and black levels 50% gray would be darker. If either Sony has any way to increase gamma it would be in the service menu, so I'd say to generally ignore it unless someone who is familiar with your TVs could tell you how to change gamma. You know how you were talking about setting 100% white to a certain level, well no one suggests setting 50% gray to a certain light output even though in my opinion 50% gray light output has more difference in the subjective brightness of the display.


For my SXRDs the percentage way doesn't work, so I would suggest using a basic pattern that shows gray shades near black and set brightness as low as possible so long as you're not clipping dark grays.

Warren, everything looks fine. Only thing I noticed is that blue/red take a bit of a dive on the gamma graph but this is because at 90%, they both spike up a bit on the RGB graphs and then they come back down at 100%. So, this makes red and blue brightness at 90% a bit more than it should be relative to 100% (because there is more red and blue at 90% than there should be) - and thus a lower gamma response on the graph. I don't see any color measures in your files so that sounds like your next step.

cheers,


--tom

Ok, I think I am done. It doesnt look like I can get it any better.

I did a run of calibrations. The REC 601 looks better than the previous one I had. I also Calibrated my 1080i input. Watching LOTR right now and it looks real good to me.

Let me know what you guys think.

 

REC 709.zip 6.4111328125k . file

 

REC 601.zip 5.8203125k . file

Tom,
Thanks for taking the time to review my charts. I had noticed the spike on the RGB chart and I was going to try and level it out as well as try to raise my CR level this weekend. I hadn't separated the RGB on the Gamma chart and that was an eye opener, so thanks for pointing me in that direction. I hope to do my color measures this weekend too but I want to be sure I've got my GS as good as I can, plus I've got to burn the avshd709 disk for the red color pattern. Thanks again for looking and your input.
Warren

bassfreak, looks good. It looks like on the HD709 run, you could try ticking your high end red control up a bit or lowering your high end green control in order to see if you can get the dE values a bit lower in the higher end. Your blue response has a bit of an up and down shape to it - not much you can do about that. Lowering contrast should shift the response left to right so you wouldn't have as much of a blue dip right at the very high end but then you'll lose some overall light output. Depending on your viewing conditions, this may not be desirable.

hope this helps,


--tom

ok, I had to take a break from calibrating because I think I was letting it overwhelm me..LOL (stupid TV, stupid AVS forum with all there helpful tools and people....)

anyway, I re-did the calibration tonight and moved my PS3 over to a different HDMI port so that way I could get fresh default settings to work from.

I did about 9 calibrations and so far this one has come out the best, I measured Primary and Secondary colors (or colours which ever you prefer...)

it is about 20 after 11 and I am in a completely dark room, no ambient light at all.

anyway here are my before and after settings.

by the way in the information log on the "measurements" page I put all the settings for the TV, I went with Custom Default settings and the only things I changed were Picture, Color, Sharpness, and white balance.

I also used AVS709 for BR to get the test patterns and on the brightness the default "50" made line 17 barely visible, if I moved it down to 49 it disappeared if I moved up to 51 then line 16 became visible.

any feedback is greatly appreciated.

Thank you

 

HDMI PORT2.zip 1.8408203125k . file

50 sounds right based on what you described. My TV does the same around 49 or 50 depending on how bias is set.


You need to use a fixed iris for taking anything other than basic grayscale measurements. Your gamma measurement varies because of the auto setting, but more importantly your color setting is probably too low because of the auto iris measurements. It's possible that Live Color might be better set to off (it affects red), but it's hard to say because the Y on the primaries and secondaries are no good unless you use a fixed iris for measurements. If your green and red are still inside the CIE chart triangle in measuring with a fixed iris (likely) then it might be better to use Color Space = Wide like on some of your earlier measurements. I can't come up with any reason to necessarily doubt the xy on red and green from a display 2, so Color Space = Standard might actually be less-accurate for your A3000. I consider the general effects of detail enhancer on high to be excessive, but sharpness is more a personal preference thing and I'm not sure if the setting would necessarily change any measurements - although I know in how it works that it can alter on-screen shades. If you don't need the TV to be as bright you could probably turn on power saving (I'm guessing it's not on by the black measurement). Your gains and bias settings do seem more reasonable than when there was light in the room.

yeah I was debating about the whole iris thing, I only left it on default just to see how everything came out.

thanks for the suggestions, I will make those changes and see what happens.

I kept all the settings the same but took readings from going thru each IRIS setting.

so far I like "MIN" after I took the reading it seem to come out the best as far as the RGB scale. I did, however, make on change and retook the measurements, I forgot to save before I made the change not a big deal I only ticked the B-Bias up one tick to bring up the IRE 10 & 20 other than that all the settings for each file are the same.

I was able to cover up about 80% of the additional light coming into the room but I only got away with it because my wife isn't here and I hung blankets on the drapes in there rear of the house and covered up as much as I good on the front window. LOL it looks total redneck in here right now...HAHA

 

Advanced Iris chc files.zip 4.8203125k . file

creatine64,

The MIN one looks good as far as grayscale goes. It doesn't look like you have any color readings though. As alluring noted, you should re-do your color measures with a fixed iris setting (like MIN) and then set your Color control (I would also probably turn off Live Color before I set Color/Hue to their optimal values - if it is increasing the luminance of red then you could use it to boost red if your red is relatively dim to the other colors). If your red/green primaries are still inside the triangle, try the Wide setting for Color Space. Make sure to re-check Color and Hue after setting it to Wide just in case things change a bit.

I would probably turn off the Edge/Detail Enhancement stuff and re-check the Sharpness setting after turning them off. I would then just turn them on on an as-needed basis if you find certain standard def cable/satellite material to be blurry.

hope this helps,


--tom

My comment about measuring a fixed iris setting was mostly in regards to colors. Without a color measurement with a fixed iris it's hard to say what you need to do with Color, Color Space, and Live Color. For the most part grayscale is somewhat similar regardless of the iris setting and primarily what varies are black and white levels depending on the iris setting.



Your measurements reflect that. On my measurements High has the same 100% white as Auto2 in a dark room and your High measurement is only a bit brighter than Auto2, so you're probably not picking up much room light.

Putting a blanket or two over the TV and meter would do the same thing and might be easier. If your player displays what chapter it's currently on you can use it to make sure you're on the correct chapter and take measurements with the TV screen covered.

Question regarding the settings in HCFR, is it better to click the "average many reads on dark measurements" or to leave it unclicked?

Thanks Tom, I got focused on greyscale and didn't even think to work on the color settings with the different IRIS options after reading alluring's response and your suggestions I will redo them.


I get it now, my mistake, I meant to also tell you that when I turned off Live Color my reds dropped across the whole scale so I left it on, but I will redo the readings on each IRIS and the color options.




yeah I remember you saying that one other time but I'm using a PS3 and it doesn't display along the front, I think I've got the motions/button clicks down where I can navigate the AVS709 disc without seeing it may give that a try but it may be easier to wait til evening or do what I did before.

You only need to use one fixed iris setting to get an idea what's going on with your colors. All the iris really does is vary how much light reaches the screen, so there's no real need of measuring each iris setting. The problem with using the auto settings for colors is that the iris changes depending on which color you're displaying which messes up the Y measurements, but that doesn't happen if you choose one fixed iris and measure that way.

Generally it would be recommended to average readings. Doing that takes longer to get low-grayscale readings on my TV, so if you really wanted to you could take one run with and one run without to get an idea if the shorter time readings varied much from the averaged.

pbc,

Just to add a bit more to Alluring's reply. HCFR by default, I believe, will do a total of 8 readings or up to 10 Y (cd/m^2) when that option box is checked. So, for example, if HCFR is reading ~1 Y, it will average 8 readings. If it is reading ~4 Y, it will average 2 or 3 - in this scenario, I'm not sure if it is doing the 3rd reading and "throwing it away" because it takes the total Y over 10 or if it is using it in the average calculation and then stopping since the total is now over 10. I would guess the latter. In any event, this is all configurable in your ColorHCFR.ini file found in the HCFR software install directory. Here are the options to look for:

[EyeOne]
MinRequestedY=10
MaxReadLoops=8

So, for example, say you wanted to do 5 readings (and take the average) for every reading. You could change these to:

[EyeOne]
MinRequestedY=10000
MaxReadLoops=5

I personally found when I experimented with doing that for grayscale, that my readings were not very different from just using the defaults so you're probably fine just clicking the box in the GUI and sticking with the defaults.

hope this helps,


--tom

Thanks, boy does it take a heck of a lot longer when you click the average readings though!!

Another question about HCFR. I see a lot of people quote a range (e.g., 6300 to 6600) when referring to grayscale. I know this isn't necessarily depictive of true D6500 (as I understand a measured 6500 may not equal D6500 if I recall from my much early days of reading up on this). But does HCFR offer any graphs or a simple way of seeing this range from 10 to 100%? I think right now one has to click on each of the % stimulus columns and look at the measured T value (which I think is the color temp?), but was wondering if there was an easier way of showing this?

Err ... nevermind. Color temperature graph was sitting right in the Graphs section!!