THE SONY SERVICE CODES - Articles, Comments, Discoveries - Page 57 - AVS Forum
Forum Jump: 
 2Likes
 
Thread Tools
Old 06-18-2006, 06:18 PM
Member
 
fred33's Avatar
 
Join Date: Feb 2005
Location: Las Vegas
Posts: 186
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Anyone wanna buy a tv?

34 XBR960

Las Vegas
fred33 is offline  
Sponsored Links
Advertisement
 
Old 06-18-2006, 07:23 PM
Advanced Member
 
POWERFUL's Avatar
 
Join Date: Jan 2003
Location: The Isle of Long
Posts: 915
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Not to bug you fred, but this is the wrong forum for that. Anyway PM me as I need to replace mine.
POWERFUL is offline  
Old 06-19-2006, 10:52 AM
Newbie
 
mku917's Avatar
 
Join Date: May 2006
Posts: 14
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Just wanted to ask you guys what you think about getting an XBR960 calibrated. Specifically wanted to ask, how do you know if the guy that is calibrating it is actually doing a good job? Is it possible to get a poor calibration? because people say that you need to get it "properly calibrated". If the person that is doing the calibration is ISF certified is that good enough or could an ISF certified tech do a poor job? I know that you'll probably say read through the forums to find my answer but I just don't have the time. Also, I saw in "Widescreen Review" magazine in the front of the magazine where they have a protocol for the ideal setup or home theater. They recommend a kelvin temperature calibration. Can the ISF techs do this? I'm just worried that I won't get a perfect calibration or near perfect. Any info would be greatly apreciated.
mku917 is offline  
Old 06-19-2006, 12:49 PM
AVS Club Gold
 
GlenC's Avatar
 
Join Date: Jun 2003
Location: Los Angeles - Whittier, CA
Posts: 2,555
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by mku917 View Post

Just wanted to ask you guys what you think about getting an XBR960 calibrated. Specifically wanted to ask, how do you know if the guy that is calibrating it is actually doing a good job? Is it possible to get a poor calibration? because people say that you need to get it "properly calibrated". If the person that is doing the calibration is ISF certified is that good enough or could an ISF certified tech do a poor job? I know that you'll probably say read through the forums to find my answer but I just don't have the time. Also, I saw in "Widescreen Review" magazine in the front of the magazine where they have a protocol for the ideal setup or home theater. They recommend a kelvin temperature calibration. Can the ISF techs do this? I'm just worried that I won't get a perfect calibration or near perfect. Any info would be greatly apreciated.

Interview your calibrator, ask questions.

Glen Carter
Home Theater Calibration
www.ISFHT.com
GlenC is offline  
Old 06-19-2006, 02:53 PM
Member
 
fred33's Avatar
 
Join Date: Feb 2005
Location: Las Vegas
Posts: 186
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Yeah, I know its the wrong forum. Thats why its only a few lines

Quote:
Originally Posted by POWERFUL View Post

Not to bug you fred, but this is the wrong forum for that. Anyway PM me as I need to replace mine.


Las Vegas
fred33 is offline  
Old 06-19-2006, 06:28 PM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by mku917 View Post

Just wanted to ask you guys what you think about getting an XBR960 calibrated.

This is the wrong forum for this question. Try the Display Calibration forum, right next door:

http://www.avsforum.com/avs-vb/forum...ysprune=&f=139

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-20-2006, 06:38 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,292
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 15 Post(s)
Liked: 18
Quote:
Originally Posted by KenTech
We are assuming that the camera's response is quite linear in color response at differing brightness levels. I hope that's not expecting too much!
I performed some experiments, and discovered my camera" "response" apparently isn't quite "linear" enough to provide an accurate representation of a greyscale test pattern(IRE steps/etc) on the screen with one exposure. I believe the screenshot attached to this message as "powershotg1whitebalancelinearity.jpg", along with the description below should demonstrate this.

All 4 greyscale "slices" in attached image are all taken from the same "exposure" but processed differently via software as noted below :

Top - Greyscale (colors completely desaturated)

2nd Down - Custom white balance, "as shot" by the camera -- to calibrate Camera white balance, I used a sheet of roll paper of the type once used in teletype machines sitting on the ground in direct mid-day sunlight.

3rd Down - Camera's "daylight" white balance setting, set using Camera's "RAW" viewer/twain driver.

Bottom - Camera's "auto" white balance setting, set using Camera's RAW format viewer/twain driver.

----------------------------------------------------------------

DO note that the attached graphic is NOT meant to represent what the greyscale actually "looks like" on my screen. Although, the "auto-white balance" slice overall, and the darker IRE bars in the 2nd "slice' downmiddle are closest to what is "on screen" given the RDRV~BCUT, and other various related settings(white level/black levels/etc) used. To be as complete as is reasonably possible, for this image I was using : RDRV~BCUT= 32-17-9-31-16-11 .

------------------------------------------------------

The file attached as "exposureinfo.txt" contains detailed info on the camera exposure settings as saved with Camera's RAW format, as well as info on the TV's user menu sliders used. Note that "custom" white balance as reported in the file is only relevant for the 2nd slice from top as described above(and all slices in file attached to "part II post below). I also changed the file number/serial number to "various" since the file info pertains to 1 exposure described above(for powershotg1whitebalanceliearity.jpg), and for several exposures for Offsetcomparison.jpg, as described andattahced in the next post (part II). Suppose I should have changed the "time" as well to be more accurate for all the shots as well .. Oh well ... You get the idea ....

-----------------------------------------------------------------------

Thoughts:

In such cases when one cannot easily "move" their PC monitor(assuming a relatively close to 6500K "setting" on the PC monitor) to the same room as the TV :

I do believe it may be possible in some cases to use a digital camera screenshot of the TV screen(or at least part of it/part of a greyscale step pattern), along with a PC monitor with a 6500K color temp setting as a somewhat "seat of pants" tool in order to help one "see" or spot greyscale linearity issues, and perhaps aid in adjustment. But, I think one has to be very careful about drawing any "direct" conclusions from the digital camera screenshots.

At least in my case it seemed to help in this regard - Mainly, it helped me to "see" that something still wasn't quite right with my prior adjustments(mostly "by eye" only) although it did not necessarily show a perfect representation of what I needed to do, adjustment wise.

So, noticing that greyscale seems(from what I can tell) both close to 6500K(given the monitor's 6500K Color Temp setting), AND MOSTLY Linear on one CRT PC monitor I use - I decided to move the PC monitor(and PC), temporarily to the room with the TV so I could use it as a optical comparator of sorts to assist in adjusting the greyscale on the TV.

I also made use of another CRT PC monitor as a 2nd reference for not only the TV greyscale, but also to "sort of" check the other monitor. I assume since the greyscale is extremely similar with each monitor's 6500K setting for 2 completely different monitors from different manufactuers, and seems to match what my "idea" of 6500K is as well as seeminly matching the white of mid-day scattered "puffy" clouds lit directly by the sun, one might assume both monitors are probably both fairly "close" to 6500K, yes?

For completeness, Here's what most closely matches the PC monitor(with 6500K setting) via optical comparision. Note that I also endeavored to set Black/white levels+"gamma" as closely as possible on the PC monitor(s) to match the TV's settings, I used various greyscale patterns via Memory stick/DVD/etc :

RDRV~BCUT - 32-19-7-24-16-7 - Note: May need a bit of improvement "linearity" wise, concerning some slight yellowish "stuff" going on at certian brightness levels.

Either of the below may make for slightly "better", more "linear" greyscale - I haven't decided yet which of the three settings is best :

RDRV~BCUT - 32-19-6-23-17-9 -- Note: This may be VERY slightly +blue, and may very slightly(very very slightly) change the higher IRE's to the "cooler" side) however at this point I think the linearity is the best on this one - looks very neutral grey throughout brightness range ) :

RDRV~BCUT = 32-19-7-23-17-7 : Note: Linearity seems to be better with this one, but at this point, I believe its slightly +green ....
LL

 

exposuresettingsinfo.txt 0.8798828125k . file
Attached Files
File Type: txt exposuresettingsinfo.txt (901 Bytes, 8 views)

Jeff
Nitewatchman is offline  
Old 06-20-2006, 06:42 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,292
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 15 Post(s)
Liked: 18
Attached to this message as "offsetcomparison.Jpg" are greyscale slices from 3 different digital camera pics, each "exposure" was captured with digital camera using the exposure settings as detailed in the file "expsoureinfo.txt" as attached to last post(part I). The same static "custom" white balance for the camera was used for all 3 "exposures", which I believe will offer a "correction" of sorts to the screenshots I previously posted in Post #1661. Those previously posted images being images captured using the camera's "auto" white balance, a variable which might have caused some slight, unintended "differences" which I believe do not actually show up on the TV screen.

The custom white balance I set the camera to use for the images attached to this post was from when I was experimenting with using different "sources" for setting a custom white balance. If I recall correctly, for these images I believe I had it set rather arbitrarily from a mostly white screen from a LCD PC monitor, although it's really not relevant to my purposes here other than pertaining to, #1). the camera's white balance being set the same(static) for all 3 images, and #2). all other exposure settings for the camera being set exactly the same for each captured image, #3). All TV settings set the same except as noted below concerning 2170P-1 Greyscale settings.

So --- In no way, shape or form are these images to be considered indicative of what the greyscale "looks like" ON my TV screen, nor are they meant to be. As was my original purpose in Post #1661 , Instead, these screenshots are only meant to support my belief that the same greyscale results on the TV given the 3 "different" settings as described below - or to put it another way, my "belief" that on my set at least, +1/-1 DRV/CUT is the same "scale" as +1/-1 DOF/COF for warm/cool offsetsoffsets. :

Description of Greyscale slices in attached image :

Top - Digital camera "exposure" taken while using the following factory default 2170P-1 settings, with "warm" color temp selected from user menu :

RDRV~BCUT - 32-22-24-32-21-17

Warm Offset :

RDOF~BCOF : 31-26-16-31-27-19
SBOF = 7
DCOL = 0

----------------------------------


Middle - Digital camera "exposure" taken while using the following 2170P1 SM settings, with "Neutral" Color Temp selected from User menu:

RDRV~BCUT = 32-17-9-32-17-5

------------------------------------------------------------------

Bottom - Digital camera "exposure" taken while using the following 2170P1 SM settings, with "Warm" Color Temp selected from User menu:

RDRV~BCUT = 32-17-9-32-16-9

Warm Offset :

RDOF~BCOF = 31-31-31-31-32-27
SBOF=7
DCOL=0

---------------------------------------------

So, I still believe the SAME greyscale results on TV screen from any three of the above described settings as would be the case with "neutral" color temp setting selected from user menu while using RDRV~BCUT=32-17-9-32-17-5(middle example). Or, at least it's the same within my ability to "see" or detect any differences, if there are any. Hopefully, this time the greyscale "slices" in the posted images will better support this "belief" than was the case with my previously posted images with the camera doing auto white balance, although I still wouldn't be surprised if there were some very, very slight differences due to small differences in how the exposure turned out -- such as, for example, due to possible differences in luminance that might have resulted due to the scanning frequencies of the TV when the camera shutter was opened for each exposure(hope that makes sense!). I also have higher resolution seperate images(Tif/Camera's RAW format or JPEG's) for each pic if further, more "detailed" is required ....

Note that another factor that seems to support this "belief" : With BCUT=5 for example -- If I use warm offset, and lower BCOF lower than "26", there seems to be no effect/change on screen for any value lower than BCOF=26. which makes sense if each +1 or -1 offset value is "equal" to +1 or -1 DRV or CUT value, rather than say, .75 RDOF~BCOF vs 1 for RDRV~BCUT ....
LL

Jeff
Nitewatchman is offline  
Old 06-20-2006, 10:09 PM
Newbie
 
jjt667's Avatar
 
Join Date: Jun 2006
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
I just purchased a sony kd 32fs170 and the blacks are to dark. Night scenes, black clothing and any black objects have no detail.

I have tried adjusting picture and brightness to higher settings but this has no real effect.

How can I fix this problem? I do not know how to access the service menu nor what adjustments should be made, any help would be appreciated or I will have to take the tv back.

Thanks for your time,

John
jjt667 is offline  
Old 06-23-2006, 01:40 PM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by Nitewatchman View Post

In no way, shape or form are these images to be considered indicative of what the greyscale "looks like" ON my TV screen, nor are they meant to be. As was my original purpose in Post #1661 , Instead, these screenshots are only meant to support my belief that the same greyscale results on the TV given the 3 "different" settings as described below - or to put it another way, my "belief" that on my set at least, +1/-1 DRV/CUT is the same "scale" as +1/-1 DOF/COF for warm/cool offsetsoffsets.

My experiments confirm this, too, and so I have assumed that one notch change in the Cool/Warm offset parameters RDOF~BCOF = one notch in the base parameters RDRV~BCUT. You are using the camera as a comparator, which seems valid to me as long as the camera isn't making any color adjustments on its own.

I am coming to believe that my own Canon G2 is (1) NOT an accurate color-temp (CT) meter (e.g. taking a RAW image, and reading the color temp slider in Photoshop after balancing the color-patch in question to perfect gray). Nor is it (2) linear in grayscale for different exposures -- at least not enough for calibration purposes. On this latter issue, I have yet to run a detailed test, but that's my impression so far.

I have to say: There is no question in my mind that the most difficult task is that of balancing the grayscale (i.e. _DRV vs _CUT settings), as the eye is so accommodating that errors can look mighty fine! I found that, when looking at b/w photographic or movie images onscreen (Color slider set to Min=OFF), I was really happy with slightly cool or magenta shadows for aesthetic reasons, fer Pete's sake! Gray patches were much better for testing, as I couldn't relate to the content. Skipping among the various percent-white patches on the DVE disk is also useful, as you can make a color judgement before the eye has a chance to adapt. The four-patch grayscale-plus-pluge bars pattern is good, too.

I think there is a bigger issue; I'll explain in personal terms: In the context of normal evening TV programming and with ambient lighting set low enough so it doesn't affect the eye's color-adaptation, my eye has a fairly critical and consistent perception of real white. Blown highlights, white text and graphics, cartoon eyeballs -- you name it -- are all perceived consistently and are easily perceived as "cool" or neutral or "pinkish" or "warm."

In my living room in the evening, after watching the screen for a while, I simply want digital/analog white as broadcast and recorded on DVD to appear neutral white. If that is true, all else follows naturally -- bad video color, subtle white variations, a film director's color-balance variations, and truly gorgeous natural video. I know that the CT is 6000-7000K, but I don't know what it is exactly. And I don't care. If a precision 6500K "expert" calibration by one Joe Shmoe would result in my set, under those conditions, showing broadcast or recorded 100% white as anything but perceived white, that calibration would be simply wrong. There is no rational argument that video-white should be calibrated so as to appear anything but white. (What would be valuable, however, is the linear grayscale that would result. Then I would tweak the overall CT.)

The logic is simple. The world as represented in video is encoded using a specific signal level as 100% or saturated white, representing an intended true white. How that is assigned is up to the production company, but I receive it as broadcast or recorded as a specific electrical representation. This 100% white representation must be displayed on my screen as perceived white. I don't know how this varies from person to person (assuming uncrippled color vision), but who else must be satisfied, here? Some external authority? I don't think so!

So, over time, and having made a good attempt to get it right initially, my procedure has been thus: After watching my TV under normal evening circumstances, I ask "How are the whites? Hmmm. Maybe slightly and consistenly yellow." So I drop into service mode, add 1 click of BDRV and BCUT, write the settings, and go back to watching the TV. After these minor tweaks over a few days or weeks, I get a screen that makes me think "Wow, that's beautiful!" when I sit down to watch. And that's what matters! (I do write down and date the changes each time, too.)

It follows from the above that I am really curious if one notch change in _CUT tracks one notch in _DRV. That is, if I have a perfect correspondence among the various grays but their CT is off, can I make equal adjustments to _DRV and _CUT and maintain that precious linearity. There's no particular reason to assume this, although I have done so as a first approximation,

I may use my digital camera as a comparator to see if this is the case. So far I haven't gotten around to it.

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-23-2006, 01:47 PM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by jjt667 View Post

I just purchased a sony kd 32fs170 and the blacks are to dark. Night scenes, black clothing and any black objects have no detail. I have tried adjusting picture and brightness to higher settings but this has no real effect.

Do you mean that, if you increase the Brightness slider setting, the "blacks" do not turn grayish and begin to glow? If you can't do this, you have a defective set!

The certain way to test for this is to use one of the excellent test DVDs, AVIA or Digital Video Essentials. (Requires a DVD player, of course.) This would prove the point without your having to rely on some broadcast video images that may really NOT have any black detail.

But if you can't bring the black level up to a dim gray with the Brightness control, something must be wrong with the set. (The Picture control won't affect this.)

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-23-2006, 11:07 PM
Newbie
 
jjt667's Avatar
 
Join Date: Jun 2006
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Kentech,

I realized that although the info from your postings were not about my tv, the sony codes are universal. Although my tv doesn't have all the service codes talked about, my tv does have most, so I adjusted those. I have also bought a service manual for last years model the FS130 (basically exactly the same set) so I now have that to reference.

I am amazed it is much better, the picture is breathtaking and I can see fabric and dark scenes excellent.

The picture is now excellent, and wonder what else should I do. Some stations below channel 20 have a little snow or static in them but I attribute that signal since all other channels and signals are great. Are these the best settings, are there any other settings I can and should adjust to maximize performance?

Although you recommended SBRT 31-33 I found that level too low and could not see dark objects or details in black objects. Can I lower SBRT and adust something else (my tv does not have an SBOF adjustment)?

Is my tv defective because I have to set the SBRT to 54 (max setting of 63)?

Here's what I changed these are my current settings old setting in ().

Under VP1:

SBRT 53 (12)

RDRV 42 (84)
GDRV 28 (81)
BDRV 25 (79)
RCUT 31 (100)
GCUT 18 (51)
BCUT 14 (42)

Under Pallete:

VGMA 2 (1)

Under VP2:

GAMM 2 (1)
BDOF-BCOC 31

and then adjusted for warm and cool setting

Warm

GDOF 31
BDOF 29
GCOF 31
BCOF 29

Cool

GDOF 31
BDOF 33
GCOF 31
BCOF 33

then I repeated the preceding in all inputs (Video1, 2, etc.) and modes (pro, normal, movies) so when I switch I can have close to the same picture.

I plan to buy the Avia disc Home theater setup, to see if that can improve performance any further.

Thanks Kentech for your dedication and excellent usefull info.
jjt667 is offline  
Old 06-24-2006, 10:03 AM
AVS Special Member
 
RWetmore's Avatar
 
Join Date: Jan 2006
Location: Brick, New Jersey
Posts: 3,297
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 102 Post(s)
Liked: 83
Quote:


I have to say: There is no question in my mind that the most difficult task is that of balancing the grayscale (i.e. _DRV vs _CUT settings), as the eye is so accommodating that errors can look mighty fine!

Funny that you mention this, as I have recently found the same problem. I recently had to redo my greyscale because I discovered the GAMB was jacked up to 3 from zero without me knowing it . I don't know how this happened, but it did. While trying to redo the greyscale and finding pure white, I went back and forth from too much red to too much green, etc., and noticed that the eyes quickly adjust to perceive white when it is still color tainted. All I can do is go back and forth between too much saturation of blue, green, and red and try to find a compromise, but it's really difficult since the eyes adjust so quickly. One minute I think I'm looking at near pure white, then try to check it by oversaturating one color and then quickly returning back to the previous settings, but it looks different than it did before!

It's too bad colorimeters are so expensive as they would make the whole process so much easier. Are there any consumer grade colorimeters under $500 that are decent?
RWetmore is offline  
Old 06-24-2006, 11:34 AM
Senior Member
 
todd95008's Avatar
 
Join Date: Oct 2004
Posts: 224
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 19
Quote:
Originally Posted by RWetmore View Post

It's too bad colorimeters are so expensive as they would make the whole process so much easier. Are there any consumer grade colorimeters under $500 that are decent?

I use the SpyderTV with the support mode (reads x,y,Y data).
You can also use raders spreadsheet or the new calman software.
See the calibration thread for details...
todd95008 is offline  
Old 06-24-2006, 11:48 AM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by RWetmore View Post

All I can do is go back and forth between too much saturation of blue, green, and red and try to find a compromise, but it's really difficult since the eyes adjust so quickly. One minute I think I'm looking at near pure white, then try to check it by oversaturating one color and then quickly returning back to the previous settings, but it looks different than it did before!

Yep! Unless you practice this, simply looking at a continuous-tone b/w picture (Color=Min) turns out to be a poor test, no matter what I have said before. I now much prefer something like the 4-patch gray plus pluge on DVE. The adjacent gray patches are different enough, and white and dark gray are close enough, that you can much more easily see whether, say, the dark gray is bluer than the white or lightest gray. The dim pluge bars represent a deep "black" if you up the Brightness a little, so you can unify the whole pattern. Then you can check your work by skipping thru channels with Color=Min to see if you're reasonably on-target.

It takes a couple of go-arounds to get it right: approximate a good color temp for white, then balance grays, then watch normal programming and touch up the white (as described a couple of posts above), check the grays, etc. Don't rush it or you'll get frustrated. No matter how badly I have screwed up what were actually decent settings, in serch of "perfection," I have been able to return to near-perfect settings by referring to records of previous settings and touching up the whites while watching real program material.

Note that this does not work unless the TV is dominating your vision and your eyes have not become dark-adapted! An evening's dim ambient light and a suitably bright screen are perfect, not a dark room (hopeless!) or a bright afternoon when the light is colored by the green trees outside your windows (TV looks pink) or the blue sky above (TV looks orangish).
Quote:


It's too bad colorimeters are so expensive as they would make the whole process so much easier. Are there any consumer grade colorimeters under $500 that are decent?

Maybe a posh unit would make it easier. Inexpensive units tend to have light vs dark errors, and so they would mislead you on grayscale calibration. I have two such units from different manufacturers, designed for computer screens. Both get me close to 6500K if I request that, but then I have to remove a greenish cast manually by looking out the window at bright clouds (no glass) while tuning the color. Then I have a perfect, white screen, important for Photoshop digital-photo tinkering. Doesn't work on my TV, but I can calibrate a spare CRT-iMac computer, and have it nearby as a reference in the living room.

This is not a one-time process! You will want to re-visit the grayscale and white calibration as the set ages (or as you get smarter). I highly recommend learning how to do it, how to spot light/dark-gray color discrepancies, and to trust your eye while watching programming in a properly dim room. This empowers you, the owner of the set, and you can take pride in its, ahem, perfection! (See DVE or AVIA for appropriate ambient-lighting rerferences.)

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-24-2006, 12:25 PM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
TRIED-AND-TRUE SELF-REFERENCING GRAYSCALE CALIBRATION

There is a technique, easy to implement on a computer because of its fine scan lines, that will result in a perfect grayscale linearity. I have experimented with it on my TV using the Memory Stick image-viewing feature and maximum-quality jpeg encoding.

Using a computer and whatever graphics-program skills you have, make up a test pattern of several blocks against a black background. These blocks should be made up of horizontal white lines in various spacings, including zero-spacing (pure white). E.g. all lines white (solid white), every other line, every third line, every fourth line, etc. If these blocks are viewed at a distance with blurred vision, the lines appear to blend into a nearly continuous tone, and the blocks appear to be various shades of gray. But since they are made up of white lines against black, the "color" of each "gray" exactly matches the white. (Warning: Vertical lines or dots will interact with the color-phosphor pattern on the CRT, producing false colors. Use horizintal lines, or long dashes if you must.)

Now position several continuous-tone gray blocks next to the ones made up of white lines. Their gray percentage should match the perceived gray of the lined blocks when blurred by the eye. Bottom line: These blocks' colors are determined by the grayscale accuracy of the display, and the lined blocks are an adjacent perfect reference, linked only to white.

Now one simply adjusts the grayscale calibration until there is no perceivable color differences between the continuous-tone blocks and the adjacent blocks made up of white lines.

On my 36XS955, the MS display is exactly 1440 X 1080i, and so an every-other-line pattern flickers at 30Hz -- very distracting, but you can sorta ignore it. Wider spacings are hard to blur, and you have to walk a ways back from the TV to get adequate blurring from your eyes. (The patters are much coarser than on a computer.) Trying this via still-patterns on a DVD is even more difficult. But I wanted to present the basic idea; maybe creative readers will come up with some new methods that are practical for a TV.

If you can make up a pattern like this, it is a secure and accurate indication of your grayscale linearity, no matter what color you have set for white. I will attempt to clean up my experimental pattern and post it.

UPDATE-1: Zip file has now been attached, native for 1440 X 1080. Extract parts of it or cut it down, as needed. The 50% patch is 2+2 alternating black and white, since 1+1 flickeres so badly.

UPDATE-2: Just tested the file on my 36XS955, and it works perfectly from a memory stick. (The thumbnail is screwed up; full display is fine.) The most useful patterns are those in the lower-right, particularly the next-to-darkest pair. Make sure that the background is pure black, with no background glow. Turn Color all the way to Min.

This pattern should display fine on a 16X9 set, since it's the 1:1 mapping of scan lines that's important, not the horizintal scaling. But I can't test this. Let me know if a 1920 X 1080 version is needed, and I'll make one up.

 

GSLinearityTest2.zip 15.140625k . file
Attached Files
File Type: zip GSLinearityTest2.zip (15.1 KB, 11 views)

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-25-2006, 12:12 AM
Advanced Member
 
MechanicalMan's Avatar
 
Join Date: Nov 2005
Location: Missouri
Posts: 983
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 20
Quote:
Originally Posted by jjt667 View Post

Although you recommended SBRT 31-33 I found that level too low and could not see dark objects or details in black objects.... Is my tv defective because I have to set the SBRT to 54 (max setting of 63)?

I doubt that recommendation applies to FS170s. And FWIW, I definitely think it's strange that you have SBRT at 54. I have it at 24 on my 27FS170.
MechanicalMan is online now  
Old 06-25-2006, 05:52 AM
Newbie
 
jjt667's Avatar
 
Join Date: Jun 2006
Posts: 12
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:

"And FWIW, I definitely think it's strange that you have SBRT at 54. I have it at 24 on my 27FS170."

1.) what does FWIW mean?

2.) What should I do (is there something I can adjust), are my drv levels too low why is my set soo dark and have to turn up SBRT so high? (the current picute is great)

Thanks for your time MechanicalMan
jjt667 is offline  
Old 06-25-2006, 10:49 AM
Advanced Member
 
MechanicalMan's Avatar
 
Join Date: Nov 2005
Location: Missouri
Posts: 983
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 20
fwiw = for what it's worth

I can't advise you. But I think I'd be a little nervous if I needed to put my SBRT so high. That doesn't seem normal. As you said, the default is 12. If your TV looks right at 54, it must have looked completely wrong at 12..? So without going into the service menu (which we aren't expected to do), you wouldn't have been able to get a satisfactory picture on your TV? I'd be afraid that something is wrong with the set. I wouldn't want GAMM at 3 on my set either, but that one is subjective. Still, the fact that you have SBRT at 54 and GAMM at 3 makes it sound like you are compensating for a problem. *shrug* Changing your DRV and CUT values won't help.
MechanicalMan is online now  
Old 06-25-2006, 12:43 PM
Senior Member
 
irhxcbcziuzxs's Avatar
 
Join Date: Jan 2006
Posts: 395
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by BoloTheRomeo View Post

Once again I will say that I am saying only to use Vivid while gaming, at no other time do I use it. The image that Vivid provides with the right tweaks (and picture and brightness sliders at 31 clicks which is the middle) is unbelievable while gaming, it looks terrible for anything else though. I use pro mode for regular TV watching and Movie mode (tweaked) for DVDs, and Vivid for my PS2 and Xbox 360.

i was under the impression that vivid was bad for the tv, what are these "right" tweaks, and how can it make games look great but everything else crap, cause if you look at my "suggested thread" youd see that i am having black crush problems in pro... anything 480p and above, so that would be what, 1080i conversioned up?
irhxcbcziuzxs is offline  
Old 06-25-2006, 01:08 PM
Advanced Member
 
MechanicalMan's Avatar
 
Join Date: Nov 2005
Location: Missouri
Posts: 983
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 20
Quote:
Originally Posted by irhxcbcziuzxs View Post

i was under the impression that vivid was bad for the tv, what are these "right" tweaks...

AFAIK, the only thing that could be harmful about using Vivid (with the default settings) is the picture setting. On my Sony, Vivid had picture maxed, and Standard was almost as bad. Bolo said in his post that he has "picture... at 31 clicks which is the middle."
MechanicalMan is online now  
Old 06-25-2006, 02:29 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,292
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 15 Post(s)
Liked: 18
Quote:
Originally Posted by KenTech View Post

This is not a one-time process! You will want to re-visit the grayscale and white calibration as the set ages (or as you get smarter).

Given my set is used by everyone in the family and is generally on 12~15 hours a day, I wonder how often I should plan on revisiting greyscale(at least checking it to see if I can detirmine anything has changed), once I'm satisfied I have "good greyscale", currently?

I've pretty much decided regarding my current "greyscale revisiting" that I'm going to spend the time, "no matter what" and how long it takes to get it "exactly" how I want it, and I believe I'm very close to achieving that point, now ....

On previous occasions I was more interested in spending time with other improvements via SM such as image processing/etc, I had only been interested in getting the greyscale a little more "in the ballpark" vs. the factory supplied settings. But, as time went on, I also made some seemingly small improvements to those settings when I noticed something about the greyscale that particularly "bothered me".

Hence, since I have occasionally made adjustments from time to time, it's difficult to get any sort of an idea concerning how often I may want to "revisit" the greyscale settings, or whether or not anything has "changed" so far, after a year of useage ...

Perhaps, in the future I should plan to look at greyscale/test patterns at least once a year or so?

Quote:
Originally Posted by KenTech View Post

There is a technique, easy to implement on a computer because of its fine scan lines, that will result in a perfect grayscale linearity. .....
Now one simply adjusts the grayscale calibration until there is no perceivable color differences between the continuous-tone blocks and the adjacent blocks made up of white lines.

Excellent! This turns out to be very useful for me. Looked at this a little last night using your attached pattern and this does seem to work very well!

Quote:
You have to walk a ways back from the TV to get adequate blurring from your eyes.

What seems to work well for me is looking at it with my glasses removed(I'm quite nearsighted --- beyond about 15" away, things are quite blurry without the glasses!) a little closer to the screen, and then doublechecking that with glasses on farther away .... about 10 feet is as far away as I can get without knocking out a wall, and 10 feet didn't quite seem far enough away at first to allow my brain to interpet what my eyes were seeing. I did notice it was probably a good idea for me to closely examine the pattern for very short periods, as it seemed quite a strain on the eyes - especially with my glasses off!

I didn't try it, but I wonder if lowering sharpness slider, adjusting the various EE related image processing SM settings, and/or defocusing with the focus pot on HV transformer may also be beneifical to achieve the necessary blurring if one doesn't "wear glasses", or can't quite move far enough from screen, so to speak?

Quote:
This pattern should display fine on a 16X9 set, since it's the 1:1 mapping of scan lines that's important, not the horizintal scaling. But I can't test this. Let me know if a 1920 X 1080 version is needed, and I'll make one up.

It works fine on KD34XBR960 at 1440x1080 --- However, I found a 1920x1080 version of it a little easier to work with/look at, given the patterns are a little larger horizontally. I didn't try it, but I wonder if one could utilize the viewers "zoom" feature as well?

All I did was resize your pattern to 1920x1080, which seemed to work fine. I compared it with the 1440x1080 pattern as you sent it(and worked with both images while making greyscale adjustments), and didn't see any differences except the "wider" version was easier to look at/work with. Oh, I also "resaved"(using "save as"/different filename) the file in Photoshop, including the thumbnail before moving it to a memory stick and then the thumbnail(including for the 1440x1080 version) then worked OK.

BTW, OT perhaps since it's not SM related .... Does the MS viewer not work with progressive JPEG encoding? I did notice last night it worked with the version of Photoshop(for PC, just an "LE" version) and it's "optimized" standard JPEG encoding.

I had ran into some problems(big question marks when the viewer couldn't recognize the files - the thumbnails+full images) when I created some JPEG files for looking at various greyscale patterns/grey fields/etc via MS several days ago. Rather than take the time to figure out why some of them weren't working, I "minimized" the time that would have been required to "get it to work" by doing several things at one time .... shortening file names/Making first letter of file names capitalized/removing any spaces/changing 8bit greyscale images to RGB(although of course without adding any "color"), making sure I saved thumbnail along with the file, and switching from progressive encoding to standard/etc, so I'm not sure which one of those "did the trick".

Prior to the last week, it had been quite some time since I'd transferred any images to MS to use with the set's MS viewer, and so I couldn't recall any "weirdness" with JPEG images I may have run into on earlier occasions.

Quote:
UPDATE-2: Just tested the file on my 36XS955, and it works perfectly from a memory stick. (The thumbnail is screwed up; full display is fine.) The most useful patterns are those in the lower-right, particularly the next-to-darkest pair. Make sure that the background is pure black, with no background glow. Turn Color all the way to Min.

The lower right/next-to-darkest pair -- being bottom right second pattern up from bottom, and the top right, 2nd pattern down from top seem to be the most useful for me at this point, although, even though, I also found all of the patterns useful in some way to see what was "going on".

Still have to spend some more time on this(I've included more info and thoughts concerning how this "relates" to previous greyscale experiements in my next post), but at this point it appears like on my set, and for my eyes, as I've suspected before with other experiments -- a bit of a slight greyscale linearity "compromise" may be in order. With use of this new technique/pattern, however, it seems it will end up involving less of a compromise than I've been able to achieve, so far - although, only recently has it been the case I've spent as much time+effort on "getting good greyscale" as it requires in my case without test equipment, or a "perfect" optical comaprator to use.

With the new pattern, It seems I can 'get' everything except the very bottom lower-right pattern to match perfectly, or almost so. While I can lower RCUT quite a bit and get the lower-right pattern to match, this causes problems elsewhere.

Either that(greyscale linearity "compromise" required), or perhaps there may be a bit of an issue with "my eyes". I have noticed, for instance with only the Red gun turned on via RGBS setting --- while looking at AVIA "color decoder" test, My left eye seems to see a very slightly different shade of red than my right eye.

Jeff
Nitewatchman is offline  
Old 06-25-2006, 02:33 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,292
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 15 Post(s)
Liked: 18
Sorry for length of below comments, as some(or much) of it is probably more than anyone wanted to know - I do find it difficult in such cases to decide/detirmine which of my comments may be useful for others, and what comments are unnecessary. For example, the actual values for various settings aren't important for others' sets, and are only relevant in some cases+to some extent concerning how they "relate" to other values used on *my* set. So I simply endeavor to "write it down" as I "experienced" it, including at least with the details of seeming importantace for my "endeavors" ....

So, in case anyone is interested thought it might be useful at this point If I not only update my recent posts concerning my recent "greyscale" experiments, but perhaps I should also attempt to explain/summarize and report in hopefully mostly "plain english" my experience so far in attempts to get "good greyscale" on Sony TV, without necessarily going into all the relevant details :

While, in the past it has seemed apparent I can achieve very good greyscale linearity via examining all sorts of various test patterns ------ (for example, the Pluge+Grey pattern, grey fields, step patterns, the window patterns on DVE of varying brightness, using video black and looking for neutral greys with the brightness slider in the higher ranges/etc/etc) :

It's also been the case using those methods to attempt to improve greyscale linearity that I've seemingly(to my eyes at least) often ended up erring a little on the "cool" or +blue side, therefore ending up with a different overall white balance than I had intended. Or, a little less often, ended up with a nice "warmer" "linear" greyscale, but one that was intolerably yellowish/greenish throughout brightness range. Apparent to me not only with white text, but also often during programming(or white text in menu graphics such as my Dish Network receiver) when bright white backgrounds mostly fill the screen. Oddly enough, often not as apparent to me with test patterns, such as the QM "white" or "white+Blue"(the latter #5 or #6 if I redcall correctly) screens, or white fields from the calibration DVD's.

It probably bothers me a tad more if "white" looks more bluish-greyish(or +red and +blue) than if it is slightly yellowish, or slightly yellowish~reddish, although it's better to my eyes if it's a bit cool or, a bit +blue vs. "greenish", especially for B+W content.

Then, when I've went about attempting to correct for the "too cool" or "too blue", or less frequently "too warm" or "too yellowish", but with good greyscale linearity -- I've ended up "upsetting" the linearity somewhat, to some extent there is a bit of annoying reddish or yellowish contamination at certian brightness levels, perhaps especially at the low end(sometimes "upper middle" brightness, say 60~80 IRE with test patterns), or then erred the "opposite" way overall (too cool/too warm/etc). In other words, while it hasn't been difficult to "get in the ballpark" for what I want(and I also don't particularly care it that's 6500K, 6400K, or 6900K CT, although I suspect it's somewhere between 6400K and 6800K), it has been difficult to find the "sweet spot" that wasn't too cool, too warm, and with the best possible Linearity, without any(or at least as minimal as possible) color contamination at certian brightness levels. So, at this point, it Looks as if this new pattern will help greatly with this! Thanks Ken!

As Ken mentioned, granted, while it does not seem very difficult to set for pleasing "warm" whites-* without apparent color contamination ------ Getting a good "linear" greyscale throught the rest of the brightness range(say IRE 0~80) while not "messing up" my nice, warm whites seems much more difficult. Oddly enough, I seem to prefer white being perhaps just very, very slightly yellowish, although, with B+W programming, I'd rather it be a little cool or a bit +blue(probably moreso than the 5400K "look"), which definitely seems to be a "contridiction" of sorts ....

*- In addition to RDRV~BCUT, I think the "related" various White level settings are a factor here as well - In my case, white levels have to be set such that they do not give me that "burning your eyeballs out" feeling, and also can't be "too dim" in a dim room with appropriate ambient lightning - If I can't watch the TV in a Dim/Dark room for hours on end without my eyes becoming noticably fatigued, it's just not acceptable to me, and my eyes do seem rather sensitive in this regard -- Of course, OTOH, If it's "too dim", it's not good either, so,he white levels have to be just right for my eyes, it seems, for all input sources -- Just at the point where moving from a dark scene to a brightly lit scene is very close to what happens when you, say, leave a movie theatre on a sunny day.

------------------------------------------------------------------------------------------

Concerning The experiments I had been doing with the #1), digital cameras and #2). the PC monitors(with some somewhat "unknown" factors, such as their actual CT) as optical comparators --- Really, the main reason I was doing those experiments were to use those as additional "references" of sorts, or perhaps, more specifically, more "data points" in hopes they would help my "eyes" try to see a little more "closely" what was actually going on with the Sony TV's greyscale, in hopes it would allow me to adjust its greyscale to get closer to what I want to see -- As mentioned earlier -- I do believe those experiments were of use, in short, for lack of a better way to say it, at least in a "intuitive" way, and to help me get a better idea of whether slight adjustments erred "one way or another"(plus blue, plus red/plus green, too cool,too warm/etc) ...

Also An update on that concerning my previous posts -- the optical comparisions using the PC monitors seemingly turned out, in generaly to result in something a little "too warm", and/or on the "greenish or Yellowish/reddish or +red - **(see bottom of post for more details concerning some more detailed notes concerning the PC monitors+related hardware)

So, the next thing I did was to go through through to find the "best"(to my "eyes" and from a linearity standpoint) of the set of values for RDRV~BCUT from the experiment with the PC monitor(s) as optical comparator. The values I had listed in an earlier post from that experiment all turned out to be slightly "worse" than the first set of values I'd come up with using one particular PC monitor, and those "best" values I had came up with were :

RDRV~BCUT = 32-18-6-26-16-7

Then, from those values I worked on making improvements in linearity and getting rid of "yellowish/Greenish" appearance to whites which resulted from above values, and "best of" what I could come up with using the monitor as an optical comparator -- while at the same time being as careful as possible not to go +blue or "too cool". I had thought I'd gotten about "as close" as I could with the following values, and indeed, I've been living with the following values the past several days and have generally been very pleased, and believe it was an improvement(best greyscale I've seen on the set so far), and I believe it was worth the effort :

RDRV~BCUT = 32-17-9-24-17-8

So, after much examination, I thought I did a fairly good job of it to find a good "compromise" between CT and linearity concerning what my eyes "like to see", with those settings. White's were nice and white, and *almost" nothing looked "too cool", although it was apparent at times the greyscale linearity wasn't quite "perfect".

For instance, I did notice that when I had video black up using higher brightness slider(color slider at 0) settings, there was a bit of yellowish/greenish tint going on, especially noticable with brightness slider betwen about 47 and 55. OTOH, I *thought* at times the generally white text in user menu's had a slight "bluish" tint to it, and at times, it appeared a bit "reddish/pinkish" -- I pretty much chalked that up to something mentioned earlier by Ken+RWhetmore -- the "biological" white balance adjustment being "done" by eyes+brain --- I also noticed much the same thing(although either reddish/pinkish or slightly yellowish, never "bluish") with programming with quite a bit of white(but not a "full" white screen), or a menu display with mostly white background from my Dish network receiver ....

I believe I was also able to detirmine the following was "really happening" and was not something due to my "eyes" adjusting ---- This was, a bit of "yellowishness or pinkishness"(seemingly +red) mostly apparent in low~low middle brightness levels with various test patterns/grey fields --- So, I set up "cool" offset with RDOF~BCOF = 31-31-31-30-31-32 - And, this seemed to improve the linearity to being quite close to neutral grey throughout brightness range - but, also seemed to perhaps push things just a very tad too far to the "cool" side, including for higher brightness levels ... I hadn't gotten to the point of deciding whether or not that was really happening or if my eyes were just "adjusting", and to try to make other slight improvments, when I looked at the new test pattern ...

Then, last night, I looked at the new greyscale "linearity" pattern Ken provided in his last post, and came up with the following values :

RDRV~BCUT = 33-17-9-22-17-11

Except for a slight difference in perceived "color" for the darkest pair of boxes in the very bottom lower right - according to the pattern+matching the "color" between left+right sides of each pair --- this seems to provide very near perfect greyscale linearity, including with various other test patterns. When I was doing this last night, I *thought* it might also be pushing things a little too much on the "cool" or "blue" side, and, that may very well be the case, but looking at it today so far(albeit in daytime with too much light being reflected around), it appears like it *may* not be too cool. If it is, I'll just have to try adjusting for a "warmer" white+readjust for linearity accordingly -- although, I have a feeling that I won't be able to achieve much improvement while keeping good linearity and without ending up with things ending up a little "too warm" or "yellowish/reddish", I do believe it is going to allow for some improvement vs. what I've been able to do without this nice tool!

Gamma/Black level notes: I didn't check this(or look very closely at the changes in Gamma - also note I have GAMM=0/GAMR~GAMB=0 set up for all but one Pic mode, GAMM=1/GAMR~GAMB=3 for "movie" pic mode for a lower gamma) yet, but I expect It is possible I may need to lower SBRT to 28 from it's current value of 29, or lower brightness slider one click, or lower SBRT to 28+raise brightness slider one click.

I had raised SBRT to 29 from previously used 28 given my combined CUT values had decreased by -5~-10 or so from previous values(occasionally a change of 1 or 2 for the combined DRV values also resulted for my experiments). Specifically, for various experimental values during my recent revisiting of attempting to improve greyscale from my previously(generally) long used values of RDRV~BCUT 32-17-9-32-16-9 (=58 DRV/57 CUT), which were I believe was at least more in the ballpark compared to the factory "greenscale"(as I call it) warm offset values equating(I believe) to RDRV~BCUT 32-17-9-32-17-5 (=58DRV/54 CUT). Although I haven't compared it directly to the "new values" yet, or checked those with the new test pattern for linearity I believe those long used "adjusted" values were a bit +blue and +red(more +red, and perhaps slightly on the "cool" side), with the factory warm offset values I believe being quite a bit +red and a bit -blue.

Right now, adding up the DRV/CUT values for the lastest set of values after seeing Ken's test pattern for greyscale linearity, I'm -3 from the factory warm offset for combined DRV/CUT's(-4 for the cuts specifically), and -3~-6 or so from most of the values I've experimented with on previous occasions previously, so I'll have to look at black levels+gamma closely(best I can check gamma is with various gamma charts I have on MS) to see if I may need to change anything, slightly. I did do a check of white levels, (picture slider/etc), and don't think I'll need/want to be changing it unless perhaps I end up with different DRV values than current, or much more difference than +/- 2 or 3 or so.

----------------------------------------------------------------------------------


** - Add'l PC monitor info from earlier "experiments" - Note: Probably should have included this in the earlier post about PC monitors --- "one of these days" -- I now know I need to work on at least one(probably two) of those PC monitors "of particular importance" a bit as well - particularly perhaps linearity wise. One of them is also a little on the +red side, maybe a bit "warm" as well - It sure does look nice though, especially with greyscale images.

Luckily, that particular monitor actually has R/G/B (drive?) controls accessable via it's "settings" menu. Another perhaps odd thing I ran into during these experiments when hooking up monitor to external "VGA" port on a notebook PC, and another PC as well, was that the output of the graphics controller(even concerning greyscale) seems to vary a bit with this differing PC hardware(IBM using Windows) as well - even though I also made sure to use a "driver" for the monitors in question as supplied by the hardware manufacturer, which also includes a color management profile specific to the monitor(s) used.

I do have "gamma" controls available(individual R/G/B sliders can be locked together or adjusted independantly) with the driver for the video controllers in 3 of my PC's(including the notebook), perhaps those can be useful in this regard as well. The best color/greyscale output I can seem to "detect" from the PC hardware I have currently, seems to come from a 10 year old(or more?) Matrox video card installed in a server(which would be very inconvienient to move! I didn't move that next to the TV!) I have sitting in an attic hallway, go figure .....


---------------------------------------------------------------------

Shew! I do believe that is about all I'm going to have to offer concerning my experience getting good "greyscale" without test equipment or a extremely good optical comparator! Can't thank Ken+everyone else who has contributed here on this issue enough, everyone's comments+the discussion on this was really of great help !

Jeff
Nitewatchman is offline  
Old 06-25-2006, 03:10 PM
Member
 
dshreter's Avatar
 
Join Date: Feb 2006
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
I made the sorry mistake of resetting 30XS955 to the factory/newborn state. I know I'm an idiot but there's no taking it back now.

Anyway, I'm trying to get the TV back to a watchable state and I'm having some trouble. I've copied in as many of the defaults as possible. The problem is that my menus are a bit different from those in the spreadsheet attached in this forum. Other people have listed the same issue. Notably, my MID3 is entirely different, and I think this is key to why I can't get things to look right.

My mid3 is as follows: And I realize that my settings for several of these are probably totally wrong as I haven't been able to attempt to calibrate them yet with the starting far out of whack.

0 YCPO 40
1 CCPO 40
2 PRPB 1
3 DOSA 0
4 YCWD 1
5 MYCD 0
6 PSTP 136
7 PSTT 0
8 VHSC 50
9 VHSL 1
10 PLHC 100
11 PLHL 2
12 MDTC 2
13 MFRV 0

Question 1: Can anyone with the same MID3 menu share their values with me? Also, what do you have as your ID7 setting, and will changing mine make any difference? Mine is set at 11.

Question2: I'm under the impression that its possible to copy the firmware/user data from one tv to another via memory stick. Does anyone know how to do this and/or is willing to copy their data for me in the bay area if I provide the memory stick?
dshreter is offline  
Old 06-25-2006, 03:47 PM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
MEASURING SCREEN BRIGHTNESS WITH A PHOTO LIGHT METER

If you have a hand-held light meter that has an EV scale, you have a very easy way to measure screen brightness. If there is no EV scale, you will have to convert from shutter and f-stop settings, but thats only one more step. Even if you have only a film camera whose light meter you trust, you can do quite well its just more cumbersome. (Maybe you have a friend with a light meter he no longer uses, having purchased an automatic/digital camera. Thats was my source!)

Youll want to know what area you are metering, and so the light meter should indicate somehow what its coverage is. A spot meter is best, or a camera with a defined-area spot-metering feature.

Using an EV (Exposure Value) scale is easiest. You set your film speed to DIN 100 (or ASA 100), and read the brightness of an area in EV. Or, with that same DIN 100, set the shutter at 1/15 and note the f-stop readout. (See chart below.)

There is an attached chart that relates EV at DIN 100 to brightness in either the older system of foot-Lamberts (ft-L) or candelas/square-meter (cd/sq.m), a more modern brightness-unit. On the chart, the black diagonal line is foot-Lamberts and the red line is candelas/square-meter. Simply note on the chart where a measured EV vertically intersects the red line, then look horizontally to the scale at the left to get the cd/sq.m value. Same with the black line and ft-L. (Sorry for the hand-drawn look. I did it on old semi-log graph paper with pens and pencil.)

Example: Put up a white test pattern that covers only about the middle 20-30% of the screen, and measure the EV. Suppose you measure 9.0. That values intersects the black line at 21 ft-L or the red line at 72 cd/sq.m.

(A caution: Whole-screen or half-screen white will activate the brightness-limiting feature on the Sony DA-4 TVs we have, and so your measurement will be erroneously low! Further, all of these CRT-TVs have a bright spot dead-center, and so you may wish to use a pattern that has an off-center white patch, such as the gray-steps-plus-pluge screen on DVE. Dont use the TVs built-in white block test pattern its not 100% white.)

The gray-shaded boxes represent certain accepted standard brightness levels: the larger box for direct-view and otherwise bright televisions (like our Sonys), and the smaller box the THX certification standard for movie theaters.

Here are the equivalent f-stops for a shutter of 1/15 second at DIN 100. As you can see, an EV difference of 1 = 1 f-stop = a factor of 2 in shutter speed. I chose 1/15 because it results in reasonable f-stop values:

EV 10.0 = f8.0
EV 9.0 = f5.6
EV 8.0 = f4.0
EV 7.0 = f2.8
EV 6.0 = f2.0

Now, working to a brightness target value, suppose you want your TV to have a bright-white level of about 30 ft-L (= 100cd/sq.m), a good viewing brightness for usual evening TV-watching. Colors are bright, the TV isnt overstressed, the electron beam probably bloats a bit on white, but not enough to matter. Heres how to proceed.

Put up a part-screen white box from one of the test DVDs. Set your Brightness slider for a decent black level and increase Contrast until you get a measurement of EV 9.6 on your meter or the equivalent of 1/15th second shutter at halfway between f5.6 and f8. Thats it. That is very close to 30ft-L no need to split hairs.)

If you are in service mode 2170P-1, trying to establish a reasonable relationship among the _DRV, _CUT, and SBRT settings, here is a good target method: Set the Contrast slider to, say, 31 (midpoint), the Brightness slider to 31, and set SBRT to 31. Now use whatever _CUT settings give you an approximately correct black level and neutral color. Red will likely be 15-25 points higher than green or blue. (Theres no magic range for SBRT; I set mine, and the charts suggest, somewhere near the middle of the 0-63 range. My _CUT settings ended up at 43-21-23 with SBRT at 28 after some weeks of tinkering.)

Finally, increase the _DRV settings to get you 30ft-L on that white test patch. I finally got the _DRV settings right for my TV at about 44-26-23, with a really neutral white point. No question that there was much tinkering over the weeks getting the color and grayscale right. But as long as my _DRV, _CUT, and SBRT are close to the original rough settings, I know that Contrast = 31 gives me an accepable 30ft-L screen brightness. I can then routinely adjust Contrast to suit the video material and the room conditions on-the-fly.

A reminder: Different models of the DA-4 chassis and different copies of the same model will end up with different settings, so anyones suggestions are very rough! Those midpoint settings arent magic, either. Theyre just sensible places to start, and calibrating the different sliders to the 31 midpoint makes it much easier to remember when returning your set to its normal settings after changing them.

Here again is the location of the chart I previously uploaded that consolidates in one place the service-menu parameters that affect black level, white-brightness, color, etc.

 

EV-BrightnessSmall.zip 115.294921875k . file

 

EV-BrightnessSmall.pdf 116.69921875k . file
Attached Files
File Type: pdf EV-BrightnessSmall.pdf (116.7 KB, 7 views)
File Type: zip EV-BrightnessSmall.zip (115.3 KB, 5 views)

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-25-2006, 06:24 PM
AVS Special Member
 
RWetmore's Avatar
 
Join Date: Jan 2006
Location: Brick, New Jersey
Posts: 3,297
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 102 Post(s)
Liked: 83
Quote:
Originally Posted by KenTech View Post

Yep! Unless you practice this, simply looking at a continuous-tone b/w picture (Color=Min) turns out to be a poor test, no matter what I have said before. I now much prefer something like the 4-patch gray plus pluge on DVE. The adjacent gray patches are different enough, and white and dark gray are close enough, that you can much more easily see whether, say, the dark gray is bluer than the white or lightest gray. The dim pluge bars represent a deep "black" if you up the Brightness a little, so you can unify the whole pattern. Then you can check your work by skipping thru channels with Color=Min to see if you're reasonably on-target.

I tried this, and did help - thanks. I guess this is about all we can do without a colorimeter. I have gotten mine acceptable now....how accurate it is, I don't know.

Color temperature aside, my set looks better than ever with the high _DRV and _CUT values. I did some more tests and found virtually no correlation between SBRT and the _CUT settings. No matter how high I have the _CUT settings it does not require a reduction in SBRT according to my black level test patterns. The image I have now is amazingly 3 dimensional with awesome shadow detail and deep blacks.

Did you ever figure out why this is (or why I'm perceiving it is)???
RWetmore is offline  
Old 06-25-2006, 08:00 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,292
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 15 Post(s)
Liked: 18
Quote:
Originally Posted by KenTech View Post

Even if you have only a film camera whose light meter you trust, you can do quite well -

Great idea! It also occurs to me this would likely be a good method to use to "double check" and perhaps improve my white level "balancing" among different inputs/soures with SPIC/etc, and make any adjustments as necessary. Given of course a suitable level of white on screen from test pattern(easy enough to do for MS, and for inputs I can hook up a DVD player).

Anyhow -- Ok -- I found this to be easy enough, except for documenting my results in this post! I tried this with the light meter on my trusty Canon FTb SLR, using 50mm Canon FD lens(note to self - although it's fine now, need to get a new battery for the light meter soon!). In this application, It's light meter(uses CdS photocell/coupled to shutter speed, f-stop+film speed) uses Full aperature Metering, but with a central area metering system, which measures the center 12% of the frame or "picture area". You can see the 12% area of measurement directly in the viewfinder. As suggested, to take the measurements I used 1/15 shutter speed, ASA 100 film speed setting.

For the test pattern, I used DVE chapter 12, #12 "window" pattern, which is labeled "100% amplitude"(100 IRE I presume), and set the camera such that the entire metered area was "covered" with the white from the test pattern, with the TV set for "full" mode(test pattern(including all the black in the test pattern uses the entire 16x9 screen). I believe this results in somewhat close to about 20~30% or so of the screen area. I didn't double check it to make sure, but I hope it wasn't too big of an area of the screen to the extent ABL kicked in. *Update 02:45UTC Mon* : I just confirmed the below measurements with "Pluge+grey" pattern from DVE.

TV settings used: Other than the differences in Picture slider noted farther below, The following relevant settings were used for TV : 2170P-1 - RDRV~BCUT 33-17-9-22-17-11, Brightness slider at 31. SBRT=29(factory adjusted default was SBRT=28). 2170P-4 SPIO at factory default value of 10. Since I was using 480p via component video input from DVD player, I have 2170P-3 UBOF=0 for this source(note: I have UBOF settings ranging from 0~5 to balance black levels among different sources/inputs/etc, just so happens 480p from my DVD player needs the lowest setting). I have 2170P-4 SPIC=6 for this source, SPIC ranges from 0~6 to balance white levels among all different sources/inputs. This DVD player at 480p just happens to need the highest SPIC value, for 480i from the same player to "match" this it needs/is set at SPIC=0, as do most other sources. Also, of some slight relevance here are my following current settings for CBGN~YGN, which I adjusted necessarily to correct for a color imbalance for 480p/720p+1080i vs. everything else as discussed earlier in this thread, as well as to lower luminance slightly for 720p/1080i from ATSC tuner, given that I "need" SPIC=0 for those to balance among sources, and can't go any "lower" with SPIC!. Current settings at left, factory defaults in parenteses :


CBGN = 4 (4)
CRGN = 4 (5)
YGN = 4 (5)

GAMM was set at 0, GAMR~GAMB also all at 0, the default for GAMM=0. BLK also at 0. ABL settings for all pic modes were adjusted per Ken Tech's recommendations early in this thread for "customizing" pic modes involving related ABL settings, without looking it up in my notes If I recall correctly, those are the pro pic mode ABL defaults. Any other relevant settings involved for Black/white levels match service manual listing for "pro" Picture mode defaults.

------------------------------------------

Light meter measurements - assuming the measurements are actually quite close to "reality", what I'm measuring seems very much to match what I "expect" I've noticed with my eyes ..... :

Follows is what I measured, listed by "picture slider" setting, unless otherwise notided, the below F-stop readings denote the f-stop setting on the lens required so light meter needle is directly in the middle(or more or less) of the ring, which is "attached" via a apeture signal level+pin to the FD lens, and "moves" inside the viewfinder accordingly with different F-stop settings :

KD34XBR960 -- Current, preferred settings, as noted above, and with picture slider at 23 :

f/4.8~5.2(note there is no label on the lens for this f stop setting - it's in between f4.0 and f5.6 - perhaps the "notch" is a little closer to F 5.6) - According to the info you provided, that would seem to put EV in the range of about 8.4 to 8.8 or so. and, according to your notes and the chart, EV 8.4~8.8 is considered "THX standard" for movie theatres. This setting for "screen brightness" is preferred by me in the sense my eyes *do* seem to like it ....

Honestly, anything brighter given a dim room(with ambient light source behind the TV matched to around the ambient light reference from DVE, mostly being reflected off the walls) Is quite fatiguing to my eyes after more than a few minutes of viewing. Anything dimmer appears "too dim". In fact, even in the daytime with "too much" ambient light of the wrong kind involved, it's more than bright enough, to the point of actually being a bit surprising(to me anyway) given what I've experienced in the past with direct-view CRT's in this regard.

Picture slider at "31" :

f5.6/EV 9.0/~20FL per your chart/Info. It seems a little too bright for my preferences on this TV.


Picture slider at "44" :

Approx f6.8(in between F5.6+F8.0 - there is no label for the actual value on the lens) - This is the target value you suggest for EV around 9.6, as you suggest something around 30FT lamberts -- but, to my eyes, this is just way too bright for me .... but, I suppose to a certian extent it's a matter of preference more than anything else ....

Picture slider at "53" :

f8.0, EV of 10 per your info. Certianly way, way too bright.

Additional measurement on another TV : Note I measured with the same test pattern via the camera's light meter on another set I use(Toshiba 34HF84 - also 34" "HD" 16x9 direct View CRT) with my preferred settings set up for that TV. This measurement resulted in f-stop of somewhere right between about f4.8~5.2(again no label on the lens for that one - I probably have documentation on the lens somewhere which would tell exactly what it is) and 5.6, so, the "needle" for the meter would be "closer" to "in the middle of the ring" given say a ~f5.2~F5.4 setting - Which I'd think should be in the range of 8.6~8.8 EV or so. Not surprisingly, very much along the lines of what I prefer "screen brightness" wise on the Sony .....


----------------------------------

Quote:


, I know that Contrast = 31 gives me an accepable 30ft-L screen brightness.

Well, not meaning to be "contrary"(at all, so please don't take this the wrong way!) here, as each to his own, and what is preferred by YOU is of course the important thing in your circumstance ... However, just to comment on this from my "perspective" ---- according to my "measurements" as noted above(if correct, and I suspect they probably are, or are close enough given the results I've been consistantly getting from exposures with this camera+light meter for many, many years, although I've never actually compared it with another light meter ) : I don't think with this set(or any other) there is any way anything that bright would be acceptable to my eyes, they would become easily, and quickly fatigued with that much "brightness".

I can also just as easily remember Pic slider="44"=30FT lamberts as I can "31" for EV 9.0, or "23" for my preferred settings -- Which again, evidently are right along the lines of the "THX standard" per your chart - which, although as you say it applies as a standard for movie theatres(presumably not necessarily Direct View CRT's), I find it interesting, perhaps especially as it also seems to "match" my results from the THX optimizer "contrast" setting screen - which is admittedly, not a "precise" way to the "picture slider", and is only one of many other references I've used to set "white levels" and/or "contrast".

Of Most importantance to me is not conforming to any sort of standard for screen brightness in "foot lamberts" - Instead, for me is that white level/screen "brightness" are such that 1). an excellent picture results(a film-like, very "detailed" pic or whatever you want to call it) and #2). that watching a few hours of TV doesn't become fatiguing on the eyes in a "dim" room enviornment because it is "too bright", But, OTOH, not so "dim" that the pic appears "dim" or "lifeless". Generally, seemingly equivilent to striving for a similar effect "on the eyes"(but not quite as pronounced) when going from a very dark scene to a very bright scene as one gets when they go outside on a bright day after being in a darkened room(such as a movie theatre).

I could of course redefine the DRV values for a mid-range slider setting of "31" to be the "same" as my current preferred "23" setting(EV ~8.4~8.8), although I suspect I'd probably be getting awfully close to "0" for BCUT, which wouldn't work very well for experimentation purposes, including using the Offsets ... And again, I really don't see any reason for me to do this ....


Quote:


A reminder: Different models of the DA-4 chassis and different copies of the same model will end up with different settings, so anyone's suggestions are very rough!

Definitely! VERY rough! Especially perhaps as it seems what some of us prefer "screen brightness" wise apparently differs. Black levels OTOH are easy to "talk about", and in my experience, fairly easily to set "properly" with a pluge pattern(such as the DVE Pluge+grey or white, or THX optimizer black level test with the "drop shadow").

Jeff
Nitewatchman is offline  
Old 06-25-2006, 11:17 PM
Senior Member
 
irhxcbcziuzxs's Avatar
 
Join Date: Jan 2006
Posts: 395
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by MechanicalMan View Post

AFAIK, the only thing that could be harmful about using Vivid (with the default settings) is the picture setting. On my Sony, Vivid had picture maxed, and Standard was almost as bad. Bolo said in his post that he has "picture... at 31 clicks which is the middle."

anyone else have any say
irhxcbcziuzxs is offline  
Old 06-26-2006, 12:17 AM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by Nitewatchman View Post

I tried this with the light meter on my trusty Canon FTb SLR, using 50mm Canon FD lens(note to self - although it's fine now, need to get a new battery for the light meter soon!). In this application, It's light meter(uses CdS photocell/coupled to shutter speed, f-stop+film speed) uses Full aperature Metering, but with a central area metering system, which measures the center 12% of the frame or "picture area". You can see the 12% area of measurement directly in the viewfinder.

I have intimate familiarity with this camera, having owned two FTb and one F1 (brass!) bodies. IMHO, all of them had inacurate light meters, and they disagreed with each other! The F1 was closest, and I generally compensated with a 1/3-stop change in ASA. The other two consistently overexposed slide film, as I recall, and I recalibrated them to some good independent standard (adjusters under the top case). Then they were fine.

They may be under-reading the brightness of your TV: You think it's 20 ft-L, but it's actually 30-ish. Maybe the reverse. Moreover, make sure that you way over-cover the spot-metering area with the white patch on the TV -- i.e. the known boundary of the spot should be smaller than the white patch as seen in the viewfinder.

My point about the "31" settings was not to recommend a specific value for them but that they were convenient reference points for whatever settings you want them to represent. For you, that should be a comfortable "normal" viewing brightness. My Minolta digital Spotmeter-F sees my "comfortable brightness" as about 30 ft-L for white. Of course, YMMV.

KenTech
"We all get smart slowly."
KenTech is offline  
Old 06-26-2006, 01:10 PM - Thread Starter
Advanced Member
 
KenTech's Avatar
 
Join Date: Feb 2004
Location: Portland, OR
Posts: 714
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Quote:
Originally Posted by RWetmore View Post

I did some more tests and found virtually no correlation between SBRT and the _CUT settings. No matter how high I have the _CUT settings it does not require a reduction in SBRT according to my black level test patterns. The image I have now is amazingly 3 dimensional with awesome shadow detail and deep blacks.

Did you ever figure out why this is (or why I'm perceiving it is)???

Your statements challenge my assumptions about how the _CUT parameters work in these TVs, and I'm really grateful that you posted your comments!

My experience is from years of dealing with the physical "cutoff" controls for the red, green, and blue CRT guns. These establish a "turn-on" threshold for each color so that, for a small video signal, say 2 IRE, all the guns begin firing at the same time, preserving correct color balance at the black end of the grayscale. If you crank each one by a bit, you can significantly increase the black level without affecting color.

But it looks to me, too, that the Sony _CUT settings seem to affect the "black-end" color without affecting the brightness of those tones much -- very counterintuitive for me! I've run a couple of brief experiments and I mostly concur with your observations: I jacked my _CUT settings temporarily by 15 points each, pushing red to 61. Increasing each color by the same amount resulted in virtually identical color balance. (At least these parameters seem to track!) Then, while observing a full-range broadcast picture with large dark areas, and with the Color set to Min, I performed a READ on the remote (0 - Enter), which instantly restores the previous settings. The picture did get a bit darker in the shadows, but the color remained the same. So there is some modification of black-level brightness with these _CUT settings, but not much. The color changes brought about by differentially changing the three parameters seem much more obviious than any black-level change. And it seems that the three color settings do track well: equal increments in each of the R, G, and BCUT settings keeps grayscale balance very much the same.

My observations are preliminary, and I have no opinion on what quality changes I may see in the video image, as you have mentioned. I will spend some time investigating this. But my initial tests show that there is an apparent small increase in black level with a large change in all three parameters, and I wonder if this also represents a change in the way video is presented on-screen, not just a level increase.

Next experiment: Use one of the Cool/Warm offsets to replicate those +15 offsets (make each 46) and maybe a compensatory black-level tweak with SBOF. Then, with a simple flick of the remote, I could invoke this series of "new" cutoff settings and check for an image-quality change, too.

I just wanted to validate what you are seeing, RW, and now I must re-think the role that the _CUT settings (and their offsets in _COF) play in setting the dark-end color of the grayscale. (If black-level brightness is little affected, it makes it easier for us to tweak these _CUT settings!)

KenTech
"We all get smart slowly."
KenTech is offline  
 
Thread Tools


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off