Originally Posted by KenTech
Hah! You caught me! Yes, the Obsessive Diddler has changed a few things.
FWIW, I wasn't tryin to catch nothin. Just interested in what numbers you're plugging in for comparisons sake.
I used a kind of homespun method of adjusting greyscales on my 34xbr800 as well via computer. So it's interesting to see how someone else approaches it. Judging by how Glen responded to your cloud approach, I'm sure my method (which involves various incantations and sacrifices to the HT gods) would probably have him rolling on the floor in agony.
If it's all right with you though, I'll throw caution to the wind and offer just a few observations on this stuff, which may or may not be worth the cyberspacial parchment they're typed on.
Like you I also chose the Neutral color mode for the exact same reason, because it appeared to have no offsets like the other Temps.
In addition to that, however I also decided to "zero-out" (as it were) 2170P-1 YOF, CBOF, and CROF for my primary input (DVI), because I was concerned that those adjustments might also be contributing to the red and blue bias of the TV. (The color decoders I'm afraid are not the cure to all of Sony's signature color bias.) After calibrating my color decoders, I actually went through all the possible color adjustments for DVI 1080i that I could find, hunting for other potential offenders/contributors to color bias (which is how I also happened to experiment with the CXA2151 offsets). And aside from the grey scale adjustments, YOF, CBOF and CROF were the only ones that looked like potential candidates. The default values on my TV for YOF, CBOF and CROF were 7-44-41, which (everything else being equal) to my mind said-- red and blue push. So I set them to YOF=7, CBOF=31, CROF=31, and for better or worse, have left them there ever since (for 1080i DVI).
The adjunct to this story however is that I'm also using the 1080i that bypasses the DRC/MID circuits... which is why I'm sort of resistant to offering others who may not be using the same signal path much input on this subject... because I can't promise that my approach will work with the regular 1080i signal path due to the add'l processing it receives in the DRC/MID circuits. For the sake of intellectual curiosity, we'll just ignore that though for the rest of this post, and assume it don't... er, doesn't matter.
OK. So I've already done my color decoders. I've Neutralized my Color Temp and the YOF, CBOF, CROF offsets as best I know how. So what's next? Well, grey scales naturellement... which is of course the real nightmare. After reading a couple FAQs about clouds and light meters and trying a few experiments, I decide to take a leap of faith and fashion my own method since I don't have either of those things readily at hand, with the goal of simply trying to elicite the widest and subtlest range of color possible from the tube. Whether that's the same goal as the 6500k and cloud approach, I dunno at this point, but it's all I have. So I forge ahead. (Notice we've now moved into a more present tense for this bit of the flashback. Whether it's present participle, past participle or past imperfect I'm not sure though, as english was never my forte. Anyway...)
Step #1, rather than worrying about using the grey scale to control contrast, I decide instead to begin with normalized values, namely 31 across the board, and to leave the Red Drive and Cutoff at 31, and only adjust the green and blue. This may have been one the suggestions I read somewhere else, like perhaps in the grey scale section around page 6 of the GWII FAQ
, or not, I don't really remember. Since it seemed to simplify things though, I go with it. Anything that simplifies is good, since I really have no idea what the **** I'm doin anyway. It seems a logical (and perhaps also providencial) way to proceed though.
This brings us to our (or rather, your) current RGB Drive/Cut settings...
...which look very untidy and unsimplified to me because the Red values aren't normalized for 31. So let's try to fix that. My math was never too good either, but if I transpose your values so the Red drive and cutoff is 31, I think that gives us around...
Still not the same as my 31-15-21, 31-15-21... but closer. Blue is still misbehaving, but the red and green relationships are now settling down, and beginning to look more within the margin of error for something being assessed purely by eye.
The first thing I notice when doing my own grey scale adjustments (and just to keep our timeline straight, this is long before your very helpful and informative SM thread came into existence) is that I'm getting very similar values for both my drives and cutoffs for a given color component. So I take another leap of faith (since I've done my best to neutralize the color bias everywhere else I can think to do so) and decide to try entering the same values for both drive and cutoff on each component and then adusting them up and down in tandem like one single control. It's just a dumb linear voltage adjustment I say to myself. And if I were a smart engineer, I probably would have implemented them that way to simplify things. During my reading for example, I ran across some TVs which didn't even have cutoff controls.
To my stunned amazement this tandem approach actually seems to work!... So now I'm like really stoked. I think I've cracked Sony's elusive color code and may actually be able to see colors on their TV that resemble something close to what the original video authors intended. So I proceed with renewed vigor and vitality to conquer my 34XBR800's greyscales...
Looking at your transposed greyscales now (which were also done via the Neutral Temp)...
...I see what seems to be a similar pattern, namely the value for the drive setting is not that different than the cutoff value for that component. They're nowhere near the wild and seemingly random grey scale values I've seen on other people's sets anyway. Again, my math ain't too good, but if you average the drive and cutoff values out (like I would have for my tandem approach), then you get:
Now our reds and greens are both right on the money. And at this point I feel comfortable revealing that I used an entirely different method to adjust my red and green components than yours or Glen's. For one thing, the Color was on (at the color decoder calibrated level). For another, I used a pattern of color bars. Then all I did was go into the SM and disable the blue gun with RGBS, and tweaked GDRV and GCUT together until I got the most distinct looking reds, greens, and most importantly, yellows. Yellows that did not seem to favor either red or green but which were as distinct as possible from them. This got me into the neighborhood of 15, which was then further refined after I looked for color bias in my blacks.
Back to the present... and our TVs still seem to be disagreeing on exactly what blue is. Yours says its around 13 or so, mine says around 21. Maybe our blue guns or phosphors just work differently? That's quite possible. Or maybe the discrepancy lies somewhere else. You're looking at clouds after all, while I'm busying myself (long before the fact) just looking for color differentiation. So I begin to take a closer look at the two methodologies. There must be an error in my approach I figure, so I try setting blue to 13 instead of 21... No-go. Now my whites are lookin too yellow. And other colors seem to overpower my blues. So I look elsewhere for an answer. And a few other possibilities shortly present themselves...
First (I think to myself), Ken's adding blue via his Gamma controls. Which means his greys must be lookin a bit yellowish at 13 too... I dig a little deeper and also find that by his own admission Ken's white reference is actually a bit off the ideal 6500k and probably more in the neighborhood of 6100k... A little more digging and I find Glen's chromaticity chart, which (if I'm reading it correctly) tells me that 6100k has more yellow in it than 6500k. So now I'm beginnin to suspect there's a rat, or a fly somewhere in the ointment, and maybe my 21 setting isn't that nutty after all. Ken may be trying to match a more yellowish reference to begin with (I have no white reference btw, all I'm looking for is maximum color differentiation), and also seems to be trying to compensate for excess yellow with the GAMB control. So now I'm thinkin his 13 is a little low and maybe he needs to add some more blue in there like I got, to bring his somewhat yellowish 6100k more into the neighborhood of "6500k" (whatever that is).
Then I look at his YOF, CBOF, CROF settings, and see he hasn't neutralized these the way I did. Both his red and blue are elevated compared to mine. More blue (and red) being added in. So maybe that's also why he's dialing blue back down via BDRV & BCUT. And the pieces (some of them anyway) seem to be falling into place... And yet I still don't know whether to speak up because this could all just be in my head, and bringing this stuff up might sort of offend some folks, and make people think I've got some kinna ego problem. And for all I know the MS may work differently than DVI (though I secretly sort of doubt that, since they're probably both essentially digital RGB inputs). So I vascillate briefly and then decide to post this message to see if Ken and Glen (who's names I just now notice rhyme) think old ADU is some idiot video savant, or just having some bad hallucinations brought on by the monsoonal SoCal heat. This post dedicated to R.W. and the other southern AVSers who've hopefully made it through another nasty hurricane in one piece somewhere in the redneck Riviera.