kjgarrison, these sets do look great once they are calibrated and I am pleasantly surprised. A lot (most) digital flat panels can't produce results like what I've gotten off of the 71 and 81 Samsungs. I am elated to see that there exists a panel that can be properly calibrated that doesn't cost $8k.
I live so far out in the country in rural northwest Wisconsin that there is no way to get anybody to come to my house to calibrate.
Well, that's not true
I live in Chicago and I hit Wisconsin all the time. I'll probably be up your way within a month so if you did want me to calibrate your Samsung I could.
I have access to a probe and the ColorHCFR software and have been taking some measurements.
What probe do you have access to? If it's a spectroradiometer like the EyeOne Beamer than I'd say you could get some great results if you spend the time. If it's a "spider" or other colorimeter than you might not do so well. The colorimeters have a hard time with LCD's because of the unique spectral output of the bulb.
What picture mode were you in "before"? I assume at least for "after" you were in movie mode, since you can't get at the greyscale adjustments with DNie off other than in movie mode.
Both before and after measurments were taken in "movie" mode and both were with the "warm2" grayscale. The "warm2" grayscale is the closest one to the reference D65 grey out of the box. However, warm2 IMHO doesn't look the best out of box because the color decoder is set with a huge amount of red push... thus D65 looks too red on a stock set. In other words, cool1 might look better on an out of box set because it's a very blue greyscale and this helps to offset the very red color decoder. Once calibrated, the color decoder is flat so the D65 greyscale will look good.
I did not however choose to use movie mode because it allows for user control of the greyscale. All of my calibration work is done in the service mode and all user controls for all modes (and more) are available in the service mode. I used movie mode because there are specific controls in the service menu that allow me to make the new user defaults to go to movie mode and warm2. Movie mode is just nicer to work with in the service menu. Also, the way I calibrate, if the customer hits the reset button, the user controls return to their defaults which are the new calibrated settings.
If you decide to calibrate your TV yourself I'd suggest movie mode though for the reasons you stated. You're better off NOT using the service mode on these sets unless you really know what you are doing. You can do most things I do in the service menu with your user menu in movie mode. You are missing some of the controls I have access to in the service menu but there is much less risk.
There are literally controls in this chassis service menu that if pressed will royally hose your set. Samsung incorporated a self calibration routine in the service menu that will run through it's paces when executed. The routine expects a specific test pattern to be on the screen when run, and if it's not there the TV will calibrate its self to whatever is on the screen at the time... this is not good.
What contrast ratio did you get?
I don't recall, it was average though. I usually don't record the contrast ratio because it is not something that I can adjust, so whatever it is is what it is
Where do you get the "ideal" luminance for 100% white from? 41.680 seems quite dark. With my contrast in the mid 80s I get ~ 71 fL. Tom Huffman recommends 50 for LCDs, but that was too dark for my taste.
Ok, ideal luminance only means that it is the arbitrary luminance that I chose to make white at 100 IRE for this particular calibration. The reason it is called "ideal" in my table is because the luminance of the colors is derived from whatever the luminance of white is. In other words, Rec709 defines how bright red, blue, green, magenta, cyan, and yellow should be with respect to a given luminance of white. So if you take the measured luminance of white and multiply it by the coefficients given in Rec709 for the primary and secondary colors, you will get the "ideal" luminance for all six colors.
Red is always supposed to be (ideally) 21.3% as bright as white. So if I measure white at 41.6 ftl, I then multiply it by 0.213 to determine the "ideal brightness of red.
Ergo, (41.6 ftl) * (0.213) = 8.86 ftl
This same method is then used to determine the "ideal" luminance of the other colors by using their coefficients. There are different coefficients for different Rec and SMTP standards, but below are the Rec709 numbers for HD phosphors
White = 1.00
Red = 0.213
Green = 0.715
Blue = 0.072
Cyan = 0.787
Magenta = 0.285
Yellow = 0.928
So now how did I choose 41 flt for white on this display... well to start, the absolute luminance chosen for any give display is totally arbitrary and relative. Have you ever wonder why exactly Tom Huffman recommends 50 for LCDs? He (or someone else) basically just pulled it out of the air. Really, any display can be set to any luminance above about 12 ftl and be considered within spec, but like you, most folks would find this too dim.
Also keep in mind that most colorimeters do not give an accurate reading for the value of luminance. If you haven't heard this before then you are probably surprised, but if you are using the spider probe then you can pretty much discard the value given by the meter for luminance. If you are using the EyeOne then the luminance value is pretty good, though it's probably reading slightly high. For this calibration I used my PR
-650 which is known to have very exact readings for the actual value of luminance. Anyway, what I am getting at is that you should just decide what luminance looks good to your eye and probably take what your meter is telling you for this particular measurement with a grain of salt. Just pick a luminance value that looks good for what you want.
That being said, just remember that luminance is tricky because your eye will adjust to nearly any luminance once it has a few minutes to acclimate. Higher luminance always looks "better" immediately uppon first sight, but once the eye acclimates there is no discernible difference in perception (this has been proven in double blind studies). So as soon as you lower your backlight setting the picture will look "worse" and as soon as you raise backlight the picture will look "better." What I suggest is lowering the backlight while looking away from the screen, then go in another room and do 10 jumping jacks and come back to evaluate the image after giving your eye a break...
On this particular 46" set the customers actually told me that they wanted me to lean the calibration towards a dark room environment. They also specifically asked for me to make the picture darker and more like a movie theater picture. For the record, a movie theater image is usually between about 8~14 flt and sometimes even less.
One area of confusion with these sets is how to actually adjust the luminance. Of course it makes sense to think the contrast control is the one to reach for, but it's really not. If you measure the contrast (white level) as you change the contrast control, you'll find quickly that the luminance of white really doesn't move a whole lot when you increase contrast. Furthermore, on the 71 series sets they start to clip (in the service menu) as you increase contrast (I don't think the user menu allows you to raise contrast to the point of clipping). Also, when you move contrast and brightness controls around you change the dynamic range of the video processing and also you change the gamma response throughout the curve. Changing the dynamic range of the video processor can have other unpredictable detriments such as a dead and lifeless picture as well as artifacts at low IRE levels.
The best way to adjust luminance on these sets is with the "backlight" control. The backlight control is just a dimmer switch for the bulb. When you raise the backlight the bulb runs brighter and when you lower it the bulb gets dimmer. Since these customers were looking for a more film like picture I set the backlight to level 3 and then fine tuned my contrast to where it gave a gamma response I was looking for without clipping whites or running the chip outside of its range. Also, turning down the back light has the same affect as using an aperture on a DLP or even a neutral density filter... a dimmer bulb will make richer and better black levels.
Now, the additional bonus of using the backlight to control the absolute luminance is that it can be adjusted later by my customer without affecting my calibration at all. Changing brightness or contrast will adversely affect the grayscale, but raising or lowering the backlight setting will only make the image brighter or darker. So if my customers wants to get a really movie like picture they can turn the backlight lower and if they want to watch with the curtains open they can raise the backlight level; all without altering the calibration.
So I measured the luminance of white with the backlight at zero and got around 12 flt. I also measured the backlight at level 5 and got 52 ftl. And I decided to leave the backlight at level 3 for default because it provided a good compromise and I explained to the customers how to use the backlight control. I also told them not to change any other controls.
Can you tell us which of the color adjustments you had to go into the service menu to do? I know if there are individual color luminance adjustments they have to be in the service menu. What else is in the SM?
With what is available in the regular menu and the service menu would you say the 71s have a full featured CMS?
This is a little too complicated for me to answer and I've already spent more time this morning then I planned... but the Samsungs don't have a full CMS, but they do allow for all colors to be dialed in accurately to industry standards for Rec709 in the service mode
And, what do the color adjustments in the regular menu actually DO? I can't measure anything changes with them (I'm not talking about the White Balance color adjustments .. I know they are greyscale. I'm taling about the group that has pink, green, blue, and white.)
The color controls for "pink," "green," and "blue" don't do a darn thing in the user menu. They appear to be broken and never activated. If you look at a color bars pattern while adjusting these controls you will observe no change at all. If you measure the colors with a probe while making adjustments you will not read any change (as you observed). The only control that does anything is "white" and I can't imagine what the benefit of this control could possible be. Why would anyone want to change "white" with a slider when there are full greyscale controls available in the user and the service menu.
I hope I answered most of your questions well enough and that I am making sense to you and everyone else who read through this post.