THE SONY SERVICE CODES - Articles, Comments, Discoveries - Page 31 - AVS Forum
Forum Jump: 
 1Likes
Reply
 
Thread Tools
post #901 of 2969 Old 02-02-2006, 02:44 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
Made small changes to the GS&ColorChart.pdf file, eliminating superfluous column in _DRV and _CUT tables and adding color-matrix table (RYR~GYB) to section III. See original post to download current file version 04.

These files are openable by Adobe Reader, a free application from adobe.com. I have no idea why Explorer insists on trying to open it in Windows, as it clearly has a ".pdf" suffix. Another choice: Download it to a specific location as a ".pdf" file, then open that.

KenTech
"We all get smart slowly."
KenTech is offline  
Sponsored Links
Advertisement
 
post #902 of 2969 Old 02-02-2006, 02:52 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

Note that for "warm" offset, I have values set up so the RDRV~BCUT values are the same as what resulted with the factory defaults for "warm" offset(which were same as service code listing)

I can't imagine where Sony got these values, as they violate common sense. If Warm using approximately these settings were tweaked to be correct, then why would you ever switch to Neutral, as it would be way too blue? Just keep in mind that these are *offsets,* not absolute _DRV and _CUT values. Setting Warm or Cool to 31-31-31 has *no* effect.

Like you, I "take over" Cool for experimentation so I can try out changes and live with them for a while before committing them to the Neutral _DRV and _CUT parameters. Not that there's any potential problem, as I log all of my previous settings. But one can easily switch between "old" and "new" settings that way.

KenTech
"We all get smart slowly."
KenTech is offline  
post #903 of 2969 Old 02-02-2006, 02:55 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by Jediphish View Post

If not, I guess "Standard" can become the picture setting I use for SD, and "Pro" or "Vivid" could be for HD, while "Movie" could be for DVD.

Perhaps I might be missing something, but, unless you want to use slightly different RYR~GYB settings for say HD vs SD for some reason -- for the most part I don't know why you'd really need different Pic modes for "SD or HD"/etc, as for example, The 2170P-3 settings allow you to set different image processing(and Color/Hue or black level offsets) which are BOTH scan rate and input specific. In fact, they are already set from the factory "differently" for different scan rates(HD or SD/etc), even via the same input. For instance, there are different SYSM settings available for 720p, 1080i and 480p from the internal ATSC or QAM tuner, and there are different settings available for SD from the internal NTSC/cable ready tuner. You don't have to "switch" between them, the change in "settings" happens "automatically" when with the internal tuner you say, switch from analog channel 7 to a digital station on virtual channel 7.1 sending HD. For more info,

See Ken Tech's article earlier in this thread on customizing picture modes, as well as his articles/posts on the image processing settings.


I don't know if you can "rename" those or not, I've never looked into that. But, that is one of the reasons, I have set up "standard" as my "main" viewing mode for everything. Actually, I rarely use anything else, except for my lower gamma settings for "movie" pic mode, or for expermimenting with various image processing settings with "vivid", or making comparisions to "pro" which I've left at it's defaults.

Jeff
Nitewatchman is offline  
post #904 of 2969 Old 02-02-2006, 03:06 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

It could never happen, but I was thinking how interesting it might be if we could do a "side by side" comparision of our sets Picture quality, using the same sources(of course, for example a cable STB's output might of course vary and require different settings from what I need for say, YV/VC for my Dish network receiver, or JVC S-VHS deck).

What? You don't want to do a Giant Schlep between Oregon and Ohio? (I hail from Youngstown, BTW, thru high school, then on to Pittsburgh and CMU.)

What one *can* do for show-and-tell, however, is digitally photograph the screen -- or, better yet, parts of it so the picture files don't get huge. It won't prove anything regarding color, grayscale, etc., but for image quality, arguments about "edginess" and sharpening -- yes, it works. I keep thinking I will do this the next time I see someone complain about the quality of SD-analog broadcast on their set! (The DA-4-based sets are unparalleled at SD display, since they only minimally upsample 480i for 960i display. Garbage in, garbage out, however!) Trouble is, you have to use a freeze-frame, as most camera exposures are a bit long for anything moving.

An example for anyone willing to try: Display the attached freeze-frame from Monsters, Inc DVD (note the time in the filename), and take a picture. It's one of the best examples I know of minimally-enhanced video at the limits of DVD quality. View at 300-400% to see the gory details. (The picture is a pixel-accurate computer-grab from a DVD freeze-frame, *not* a photograph.)
LL

KenTech
"We all get smart slowly."
KenTech is offline  
post #905 of 2969 Old 02-02-2006, 03:08 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

Just keep in mind that they are *offsets,* not absolute _DRV and _CUT values. Setting Warm or Cool to 31-31-31 has *no* effect.

Absolutely. The warm offset values I provided, as well as the DRV_CUT values the "correspond to" should, at least in part demonstrate this.

Perhaps I didn't explain myself fully enough. To "get" the DRV~CUT values that "corresponded" to the factory "warm offset" values, I looked at what the offsets "were" in comparision to what the factory "neutral" values were. For instance, Factory BDRV value was "24". Factory value for BDOF for "warm Offset" was "16". So, looking at the number you need to subtract(15) from "31" to get "16" for the offset, means that the factory BDOF offset of 16 means the warm offset from the factory (BDRV = 24 - 15) was using "9" for it's "actual" "BDRV" value.

So, I figured that out(and double checked it numerous times) and wrote down the actual RDRV~BCUT VAlues that would produce the SAME results as the factory "warm" offset if I actually used them with RDRV~BCUT and "neutral". Then, given my new "neutral" values(actually what I'm using for RDRV~BCUT) I redefine new offsets for the "warm offsets (+ or - from 31 as necessary) so that the "warm offset" still gives me the exact same results as was the case with the factory "warm offset".

I know that seems like a lot of work, but wanted to be able to check the same color temp that resulted factory warm offset values for reference, and I just prefer to use "neutral" and the actual RDRV~BUT settings with my improved settings.

Jeff
Nitewatchman is offline  
post #906 of 2969 Old 02-02-2006, 03:16 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

What one *can* do for show-and-tell, however, is digitally photograph the screen
.)

Yes, I do that often for various reasons in my local threads, and for my TV/DTV Dx'ing Hobby. You'll find some(but mostly from other sets) posted from me In Cincinnati or Dayton threads.

Quote:
Originally Posted by KenTech View Post

It won't prove anything regarding color, grayscale, etc., but for image quality, arguments about "edginess" and sharpening -- yes, it works.

And, since I'm using the same settings there(your lastest image processing "Findings"), including If I recall correctly from one of your posts even the same value for the user menu "sharpness settings" -- I don't think we'd see any difference in the image quality

The user's PC display might be something to think about though.

Quote:
Originally Posted by KenTech View Post

The DA-4-based sets are unparalleled at SD display, since they only minimally upsample 480i for 960i display.

I agree. Unlike many HD displays(not that some of the others are all that bad -- the farojuda chipset in my Toshiba 34HF84 does an awfully good job of it as well - just not quite as good as the XBR), SD (even from analog OTA sources) looks fantastic on this set.

Jeff
Nitewatchman is offline  
post #907 of 2969 Old 02-02-2006, 03:18 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

Absolutely. The warm offset values I provided, as well as the DRV_CUT values the "correspond to" should, at least in part demonstrate this.

Perhaps I didn't explain myself fully enough.

No, I got it. My comments were not aimed specifically at you, but were to clarify for some readers who, not being engineers in their DNA, don't relate intuitively to the whole "offset" notion, especially when "zero effect" isn't zero but 31. For those who get it: I'm trying to walk that line between snowing some readers and talking down to those who already understand. Hope there's not too much repitition.

KenTech
"We all get smart slowly."
KenTech is offline  
post #908 of 2969 Old 02-02-2006, 03:27 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

The user's PC display might be something to think about though.

For black level, contrast, and especially gamma, you're right. But it's fine for seeing the actual image on the DVD. Here is a method that can be used on a Mac or PC with free software to get pixel-accurate 1:1 frame grabs for any DVD you can play on your computer. There are a few more frame grabs posted there and in the next few messages, too.

KenTech
"We all get smart slowly."
KenTech is offline  
post #909 of 2969 Old 02-02-2006, 03:57 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

Hope there's not too much repitition.

Not at all. I also didn't mean to "moan and complain" that you were pointing something out I already knew -- but unfortunetly via this "medium" I think it's sometimes impossible to know "exactly" what is meant by any given comment or point(not that it is allways necessary to - as people come here I think mostly for the information that will help them), and it is also difficult (and really unnecessary) sometimes to "word" your posts so that the actual "intent" of the comment is known. If THAT makes any sense.

In any event -- I think add'l explanations/clarifcations/etc are useful in some cases - along the lines of what I was discussing with ADU concerning his earlier posts about "gamma" -- for instance Perhaps hopefully my additional clarification might also help others who are looking at utilizing their Cool/Warm offsets in similar ways. If I recall correctly, a post similar to mine helped me understand it back at the time, and saved me time trying to figure out what was going on.

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

Jeff
Nitewatchman is offline  
post #910 of 2969 Old 02-02-2006, 04:02 PM
Newbie
 
digimiX's Avatar
 
Join Date: Jan 2006
Posts: 4
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
i recently bought a KD-34XBR960 and need to adjust the picture size/placement. all sources are cut off a bit...about 5% or so. isnt this overscan?? i cannot adjust in the normal tv settings. anyone know how to do this in the service menu?
digimiX is offline  
post #911 of 2969 Old 02-02-2006, 04:16 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

For black level, contrast, and especially gamma, you're right. But it's fine for seeing the actual image on the DVD.

I should also mention -- One issue I have run into when photographing a CRT screen, is "focusing the camera" issues and visable "moire pattern". That's as true for a "digital" camera as it is a film camera.(Yes, I'm also into photography+photo editing/etc - just not so much lately). In fact I have, and sometimes still even use a 35mm film(negative) scanner).

You don't see it with your eyes, but if you focus the camera for "sharpest" results, or use the "auto focus", I've found moire pattern will usually be present, and that has been the case for every CRT display I've ever photographed. Placing the camera a certian distance from screen can help, but how I usually handle this is slightly defocus before making the exposure - although, with my digital camera the little LCD display isn't the best for this. Generally, I use relatively low F-stop and shutter speeds of ~1/15~1/60 per second with settings on my digital camera which supposedly approximate ASA200. As for the shutter speed/etc, if I under or overexpose something/etc, I'll fix it by raising/lowering Gamma/etc. with photo editing software, since it isn't generally a priority to be exact, I just change it so it "sort of" looks good.

FWIW, attached is a screenshot I took of a portion of KD34XBR960 screen while in one of the "QM" "info" screens for ATSC information(PSIP info from the transport stream/etc), in order to demonstrate the station wasn't sending some proper info concerning some PSIP tables. The abbreviations "light up" for those tables if the receiver can properly recognize them/etc. In this case, it's a 1080i signal, and 1080i "full" mode being used.

I took this screenshot several months ago, when I was using your "old" image processing recommendations. I believe I was experimenting with greyscale also at the time, and don't recall what "other" sorts of settings I was using. This is very much a "close up" shot of screen, and is a part of a higher resolution image which I didn't resample/resize or use unsharpening or sharpening filters on -- which is something I often do for screenshots of CRT for other purposes(such as involving my TV Dx'ing hobby).

As you can see, there is a slight difference in focus on what is the far left hand edge of screen. I can't really see any of that to as much of a degree with my "eyes", even this close to screen, presently, so perhaps, unless it has to do with having the lens so close to screen/etc.(I just didn't take the time or effort to pay any attention to that) -- I might have also been experimenting with the dynamic focus adjustments a bit at the time. Also, note I wasn't really trying to get a "good shot" other than than to get a clear shot, just so the text was readable.

Update: oops! Just realized the attached file is not a very good example .. as Looking at that again, looks like I must've used a little too high JPG compression ratio when I took that image. Don't know when/if I'll ever get around to it, but I'll put it in my notes to try to get a better screenshot from something such as the captures from Monsters, INC Ken posted(can't do it from DVD, though), and I'll use the camera's "raw" format to create a uncompressed tif and be very careful about the compression ratio/etc.
LL

Jeff
Nitewatchman is offline  
post #912 of 2969 Old 02-02-2006, 05:15 PM
Member
 
ptchristensen's Avatar
 
Join Date: Nov 2004
Location: LA
Posts: 191
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by digimiX View Post

i recently bought a KD-34XBR960 and need to adjust the picture size/placement. all sources are cut off a bit...about 5% or so. isnt this overscan?? i cannot adjust in the normal tv settings. anyone know how to do this in the service menu?

If the picture looks great and its an even amount of overscan on all sides...leave it alone. You never know whats behind this overscan: bending lines in all directions, corner spots and other nasty stuff.
ptchristensen is offline  
post #913 of 2969 Old 02-02-2006, 07:57 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

HELPFUL NEW CHARTS
I have prepared some new charts to organize this stuff. Fill in your own values.

Sorry to post twice on this but I just wanted to thank Ken again for putting the work into this and posting it, and for providing/adding the excellent descriptions. It should be very useful for folks, I've printed out the latest update and already found it useful especially as it is much less "confusing" to look at to see all the relevant parameters for "Black Level", "Contrast", "Color" All on one page in an organized way, along with the different columns for inputs/etc.

Jeff
Nitewatchman is offline  
post #914 of 2969 Old 02-02-2006, 09:53 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
I just did a little tinkering this evening with CXA2171/CBGN~YGN. I put up a completely balanced-looking DVD still frame from my V5-480p input. (I prefer 480i for DVD, but that doesn't go thru the CXA2171 chip.) I watched the picture while I tweaked the values of CBGN and CRGN to see what the "signature" of maladjustment was.

Other "misadjustments" are well knows: Wrong Hue setting = greenish or pinkish skin tones. Color amount wrong = garish or insipid colors across the spectrum. Red push = colors balanced, except for red, which stands out as exaggerated.

But why did Sony "unbalance" the factory settings for the two components Cb and Cr in the service-manual chart -- and apparently also for Nitewatchman's 34XBR960? Mine were set alike: CBGN - CRGN - YGN at 4 - 4 - 4 or 3 - 3 - 3, in different columns.

The effect of unbalancing the first two is interesting but in accord with what Cb and Cr signify: the degree of color mapped along either a blue-yellow axis or along a perpendicular red-green axis, independent of luminance. If you crank CBGN above its "partner" CRGN, there is considerable blue and yellow "push" added to the picture -- not one color or the other, but both. Same with CRGN, except the colors are cool-red and green that are exaggerated. Grays are not affected. If both are increased the same amount, all colors are increased together -- as though one has simply increased saturation.

Those two color-axis parameters are together paired with a Y parameter, which is the luminance of the substance of the picture: Increase it by cranking up YGN, and the picture gets brighter but less saturated; lower it and it gets dimmer but the colors get more intense. Raise and lower all three together, and you just increase and decrease the picture's contrast without affecting color.

I guess that if there is a characteristic in the color-processing chain that must be compensated, then Sony might go for the CBGN/CRGN unbalance. Or if one of the processing chips has a lower output for one of those color components. But my 36XS955 came with them at 4 - 4 - 4 out of the box. Go figure.

I note that the color-coding of the cells for these values in Sony's chart indicates that some are set "at F[actory] A[djustment]" and others "at C[ircuit] B[oard] A[djustment]." (My interpretation.) Perhaps if one of the circuit boards is replaced, service personnel have to attach an oscilloscope and use these settings to "trim" the color-component waveforms to the correct amplitudes.

If I see color aberrations that push blue *and* yellow or red *and* green, I'll now think about trying a correction in CXA2171 -- for 480p, HD video, or HDMI, that is. Anything that goes thru 480i DSC processing doesn't go anywhere near this chip.

KenTech
"We all get smart slowly."
KenTech is offline  
post #915 of 2969 Old 02-03-2006, 03:51 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
first -- Another quick note on the new charts Ken posted yesterday -- they also helped "remind me" that although I had made sure I'd set up everthing properly to "customize" the picture modes for all the inputs/sources/etc, I was using at the time .. When I added "MS" recently, I had missed changing the value to "0" for the seperate column for "ABLT" for "MS" -- and finally realized this last night when I first glanced over the charts. Don't know why I didn't notice the slight change when switching between "Pro" and other pic modes with MS, but I thought there was something that was a little "off" ... now I know what it was ... Thanks again Ken ...

Quote:
Originally Posted by KenTech View Post

I just did a little tinkering this evening with CXA2171/CBGN~YGN.

Thanks for giving those a little more of a look. Your findings on this match my observations on this to a T. And, you have done a much better job explaining it than I have explained my observations and tinkering with this.

Quote:
Originally Posted by KenTech View Post

But why did Sony "unbalance" the factory settings for the two components Cb and Cr in the service-manual chart -- and apparently also for Nitewatchman's 34XBR960? Mine were set alike: CBGN - CRGN - YGN at 4 - 4 - 4 or 3 - 3 - 3, in different columns.

Unless someone tinkered with it(or the set was serviced) between the factory and the time I got the set(and for various reasons described below, that may be a possibility), the factory settings for CBGN~YGN on my set were 4-5-5.

I don't think I have seen much possible sign of any such tinkering or "add'l" servicing being a possibility elsewhere, although its hard to say. For instance, all 2170P4 SHUE/SCOL set defauts match the servicecode listing from the service manual, as do most other values you'd expect. I don't know, but, the "factory" SPIO setting of "10" along with most "SPIC=7" - (the big exception being DRC SPIC=4), might perhaps seem a bit unusual.

[update] Almost forgot : Other possible unusual "clues" that perhaps might suggest there was some "tinkering" or servicing going on with my set between the factory and the time I received the set : #

1). The "factory" default on this set for CBOF/CROF = 31/31 for 720p - although all other YOF/CBOF/CROF factory defaults match servicecode listing. So does perhaps(as I just happened to notice) the CBOP/CROP "factory" default of 9/36(36/37 in Servicecode listing).

2). Also, unusual perhaps especially as noone else has reported seeing these issues : a). the unusual, "sporadic" incidences I'm occasionally getting when only working in SM concerning values for specific columns of either 2170D-4 ASPW/ASPT or ZOOM occasionally "switching" on their own to values in the "next column"(and sticking) while I'm switching channels or screen modes at times when I'm working in SM(never outside SM). b). As well as a couple of odd things which have occured in SM concerning signals from internal NTSC or ATSC tuners. More detail avaialable concerning what I'm experiencing with these issues (a+b) as I reported in a more accurate and detailed fashion in Post #876 [end update].

It was "supposed" to be a new set -- but, nevertheless I suppose it is possible it wasn't entirely "new". My KD34XBR960 is a warranty replacement for another set. On that other set, long story, but essentially a bad diode in an aux power supply led to a 3 month wait and a set replacement rather than a repair. I know, it sounds crazy, and there is more to it than that, but that is the only info on it I was able to verify as 100% accurate.

I do know we took delivery of the Sony set in July 05, date of manufacture on the sticker says Jan 05. Also, I did notice when at the store "picking out" which set to replace my RCA F38310 with(of course, I had already decided there was only one choice, given the choices I had), they had no fewer than 5 open box KD34XBR960's sitting on the shelf. I also noticed the owner's manual said I was supposed to get some sort of "welcome" screen when I turned the set on for the first time, but I didn't get that.

Quote:
Originally Posted by KenTech View Post

I guess that if there is a characteristic in the color-processing chain that must be compensated, then Sony might go for the CBGN/CRGN unbalance. Or if one of the processing chips has a lower output for one of those color components. But my 36XS955 came with them at 4 - 4 - 4 out of the box. Go figure.

I note that the color-coding of the cells for these values in Sony's chart indicates that some are set "at F[actory] A[djustment]" and others "at C[ircuit] B[oard] A[djustment]." (My interpretation.) Perhaps if one of the circuit boards is replaced, service personnel have to attach an oscilloscope and use these settings to "trim" the color-component waveforms to the correct amplitudes.

For whatever "reasons", on my set there seems to be slight mismatchs for "color" between #1). those scan rates effected by CBGN~YGN/CXA2171 chip(480p/720p/1080i) and 2.) everything else(DRC). As well as a very slight mismatch between a)480p and b)720p/1080i. And, for whatever reasons, My best guesses are either #1). is that whoever(factory or someone else? I don't know with 100% certianity see above) made any "CBA" or "FA" adustments managed to perhaps not find quite the "best" settings for their adjustments, or for CBGN~YGN(if they even adjusted those at all) to best compensate for these slight mismatches I'm getting or #2). This set is on and used by various family members on average about 12 hours a day. It just doesn't make much sense to turn it off much of the time, as someone will just come in 5 minutes later and turn it back on. Perhaps it's possible some "drifiting" has already occured, although I'm fairly sure this issue has actually been there all along.

Follows is just some of the evidence I have for this -- -- note that in one of my previous posts on this which might suggest somewhat different results -- at that time I was also "experimenting" with 2170P-4 SLO - In some cases changing it to "7", from the "factory" default of "6". This time, I left SCLO at the default "6" for everything.

#1). Using the Blue+white QM PATN 6, 46 and 66 for 1080i, 480p and 720p respectively, and turning on blue only gun via RGBS = 1 -- With the "factory" defaults for CBGN~YGN -- While I found I could adjust SCOL+SHUE for the "most even" blue(to where I can't even see a "boundry" between top+bottom 1/2 of pattern) --- For 720p AND 1080i ... I need different SCOL/SHUE(Or UCOF/UHOF) settings in order to get the same results from 480p ATSC(QM PATN=46). I think this may partly explain why finding out exactly how the "MTRX" control in SM worked and how to use it was important for me.

#2). Using the "factory" defaults for "CBGN~YGN" and adjusting SCOL/SHUE for best results with QM PATN= 6 (1080i), or PATN=66(720p) -- I do need to set SHUE=32(default 30), and SCOL=33(default 32) for 720p/1080i ATSC for best results. However, that change in SHUE from default doesn't really "work" well for Red gun only, or Green gun only tests/adjustments - I can get it to "work" to somewhat of a degree, but I have to use quite different, and unusual values for RYR~GYB than I need for everything else( in my case 480i DRC, basically) - which is 13-15-5-3. And again, the results for 480p(QM PATN #46) are somewhat different.

#3). I've played around with trying all the slightly different values for CBGN~CRGN I could think of to try, to see what works best to match 480p with 720p/1080i with the QM PATN x6 series of patterns, while adjusting SCOL/SHUE as necessary as well as trying to find the values which will allow me to use the same, or nearly the same values I need for RYR~GYB for 480i/DRC.

There is nothing I've found that "works" perfectly -- However -- I think the best I can find is CBGN~YGN at "4-4-4" -- It's still not 100% perfect solution, and CBGN~YGN "5-5-4" might be the better 'compromise" -- I'm not sure at this point. With CBGN~YGN at 4-4-4 - 480p and 720p/1080i with internal QM PATN #6 "series" match the closest, with "SCOL=34" for ATSC 480p/720p/1080i - best with 720p+1080i -- This also turns out to be the best "match"(more or less) I can find for 480p via component, but with SCOL=33. Note that SCOL=33 works best for 480p ATSC as well, but I can't set it differently - I can of course set SCOL=32 for 480p/720p/1080i ATSC, and then raising 2170P3 UCOF values for 720p and 1080i - before messing with anything like that though I want to find the permanant values for CGBN~YGN I'm going to use). No change in SHUE from "30" default is required for any scan rate.

Also note that concerning all the slightly different CBGN~YGN values I've tried, the factory defaults appear to be about the worst choice of what to use out of all of them concerning these slight "mismatches".

With CBGN~YGN at 4-4-4, The "best" settings for RYR~GYB for 720p/1080i ATSC are still not the same as the 13-15-5-3 I need for 480i/DRC -- Instead, after a little more "fine tuning" and checking last night -- the best settings for RYR~GYB for 720p/1080i using CBGN~YGN at 4-4-4 turn out to be "14-15-6-4" (and I was about to detirmine this time that is slightly noticably better than 14-14-6-4) -- Which Is perhaps interesting that Ken had posted he had came up with those same values for his set for 1080i with a recent recalibration using the QM patterns/etc. But, on my set for 480P component V5/V6, it turns out the best settings for RYR~GYB are the same as everything else -- "13-15-5-3". This provides for pretty much "perfect" color balance for all scan rates involved(only with 2 slightly different RYR~GYB settings - luckily which I can choose from user menu per input and pic mode), and I can't seem to find any "better" settings - not only concerning CBGN~YGN, but also the P21704 SHUE/SCOL offsets/etc.

#4). What I thought was also interesting was -- When I was experimenting a bit with setting "SCLO=7"(I know it says set it with a scope, but for one thing, I don't have one currently) several days ago instead of the "factory" default of "6" -- In which case, I found CGBN~YGN setting of 5-4-4 (servicecode listing default BTW) worked perfectly with 720p/1080i WITH RYR~GYB at 13-15-5-3. However, that wasn't the case for 480p -- including 480p via component. I also noticed raising SCLO to 7 also made the issue mentioned in #5 below involving S-Video input to become "worse". So, even though at first glance it seemed SCLO was a global setting that would have the same effect as turning up the "Color" slider, or adjusting 2170P4 SCOL, or 2170P-3 UCOL offsets -- that didn't seem to quite turn out to be the case.

#5). As you say, and I also have observed and reported-- 480i/DRC or NTSC isn't effected by these settings or the CXA2171 chip -- but -- neveertheless, I do wonder if perhaps the following issue might still in some way be "Related", but only in the sense that on my particular set I've also found a slight bit of a color balance "mismatch" between inputs in one other place, which doesn't involve CXA2171 CBGN~YGN :

Even though Color tests show "right on" results for 480i component via DVD(or 480i ATSC via internal test patterns), for All three guns used indivdually with :

#1). "SCOL=33" (factory default was SCOL=32 for everything, SHUE=30 for everything - again note that I have left "SCLO"=6, although I did experiement a bit with "SCLO=7") ---

#2). RYR~GYB 13-15-5-3 ...

Via CV/YC (S-video) (AVIA or DVE color tests played from DVD) -- I'm getting excellent results with these settings(and centered Hue/Color sliders) and 2103-1 SCOL=8, SHUE=10 -- with "blue gun", and "Red Gun" -- BUT, with "green" gun, for best results Hue slider needs to be set to "G1" - (or 2103-1 SHUE adjustment at SHUE=11/etc) Trying to change RYR~GYB instead doesn't help, GYR~GYB at 5-3 are still the best values.

Again, since I'm not getting this from 480i component or internal ATSC pattern at 480i, I don't know what is going on with this. Could be something with the DVD player via S-video, or perhaps the s-video input itself and any "CBA" adjustment/etc -- Although, I'm not getting this with this DVD player via s-video vs. component hooked up to a another set - so that would seem to take the DVE player out of the "mix".

Quote:
Originally Posted by KenTech View Post

If I see color aberrations that push blue *and* yellow or red *and* green, I'll now think about trying a correction in CXA2171 -- for 480p, HD video, or HDMI, that is. Anything that goes thru 480i DSC processing doesn't go anywhere near this chip.

In my case, I allways knew something looked a bit "wrong" about the color with 720p/1080i. I noticed it most being an issue with "green" - seemed to be too much of it, or "unnatural" looking Green. But, it wasn't really apparent enough to be able to "quantify it" as Red+Green push(or probably more like "not enough" Blue+Yellow), as opposed to perhaps issues with the broadcast material/etc. But, I'm quite sure now that's exactly what was happening.

After changing CBGN~YGN from 4-5-5 to 4-4-4 (or 5-5-5 - I haven't decided on that yet or for that matter SPIC=0 or 1 for 720p/1080i ATSC) I think I can be reasonably sure just a "tinge" of Red+green push is what I had been getting with the "factory" CBGN~YGN settings of 4-5-5, as, in addition to what I've seen with the color bars/etc, --- those "unnatural" greens(it also effected flesh tones slightly) are now gone.

If there is anything going on now with CBGN~YGN at "4-4-4", and "SCOL=34" for 720p/1080i ATSC -- it may be a slight "electric blue" "push" you'd mentioned you had an issue with ... . I'm not sure yet, and hopefully I was just seeing things --- There was a lot of "blue" in this week's "Smart Gardening" episode and it looked just wonderful -- but Last night the fellow's blue shirt on the PBS HD Mexico cuisine show was awfully "electric blue" looking ... Even though of course in my case my QM #6 pattern with blue gun is perfectly even across the screen with the current settings, and tests with Red or green only are showing excellent results as well. Hopefully, it was just his shirt and the lighting involved ....

If not, perhaps it may be turn out to be the case that "5-5-4" is my better "compromise" setting for CBGN~YGN - In which case It works best to lower "SCOL" for 720p/1080i to 32 (note that SCOL=32 also works best for 480p in this case) for best results, and use RYR~GYB=13-14-6-4 .. which in this case, also works best for 480p - unlike the differences in RYR~GYB needed between 480p and 720p/1080i with CBGN~YGN at 4-4-4.

I've tried using both "5-5-4" or "4-4-4" "options" and the differing required SCOL settings for each for CBGN~YGN now(but only for a few hours each), and at this point I'm not sure which offers the better "compromise". I think it's "4-4-4" but may require a little more (and longer) closer observation of actual programming. I know in this case I'm just changing the color saturation(gain) by lowering or raising both CBGN+CRGN with the sam values - but the question is I think - Is it a tad " more" or a tad "less" less saturation that is the best "compromise" here --- 2 "ticks" up for SCOL, or one "tick" up for one value for Cb/Cr.

So, In any case and in conclusion, I've been able to make improvement, but it seems there are still some slight color "mismatches" going on between inputs/scan rates that it seems I have no way of completely "fixing" with SM (or user menu for that matter) adjustments. It is slight enough that at least so far with current settings(with perhaps exception of the "electric blue thing at 720p/1080i), It's not noticable when viewing programming. I may be able to find a slightly better "compromise", and I'll continue to look, but it is difficult to do as It the mismatches are not very significant.

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

2/4 Update : Last night, went through a couple of my different options again, and have finally decided CBGN~YGN = 4-4-4 turns out to be best for my set.

One perhaps interesting thing I noted was that in SM and using RGBS 2/4 and the internal QM PATN #4/6, #24/26, #44/46, #64/66 -- as I mentioned earlier RYR~GYB at 14-15-6-4 is best for 720p/1080i, but 13-15-5-3 (confirmed with AVIA/DVE via DVD) is best for 480p, just as it is for everything else (480i component, 480i ATSC/i.e. DRC). However -- I left the QM patterns up for 1080i and went outside of SM, and noticed that with the DVE green or Red color filters, RYR~GYB=13-15-5-3 appeared to be the better choice -- Now, I know there can be errors in these filters(and that seemed easy to spot for the DVE Blue filter), but what was odd was that using RYR-GYB at 13-15-5-3 outside of SM with the DVE Red+blue filters perfectly matched my results with RYR~GYB 14-15-6-4 + RGBS 2/4 inside of SM. Probably just a "coincidence", but it just seemed odd and thought I should mention it ...

Jeff
Nitewatchman is offline  
post #916 of 2969 Old 02-04-2006, 01:44 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

RYR~GYB at 14-15-6-4 is best for 720p/1080i, but 13-15-5-3 (confirmed with AVIA/DVE via DVD) is best for 480p, just as it is for everything else (480i component, 480i ATSC/i.e. DRC)

My results are similar. East to determine whether it matters. Just set up one for Default and the other for Monitor color axes. Now you can observe a good subject pucture and switch between them.

I fond the difference is a tiny change in green intensity -- totally insignificant. SInce broadcast TV is poorly color controlled anyway, I'll happily "sacrifice" that for accurate DVDs. So 14-15-6-4 is my Default. I have almost no incentive to ever switch.

If skin tone had changed noticably, I would have an issue. But a tiny green intensity change? Hardly!

Ken

KenTech
"We all get smart slowly."
KenTech is offline  
post #917 of 2969 Old 02-04-2006, 02:28 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

My results are similar.

That's good to hear. Not unexpected, but now I know besides the "factory" CBGN~YGN setting on my set, the very slight color mismatches I'm getting on my set are not unusual.

Quote:
Originally Posted by KenTech View Post

Just set up one for Default and the other for Monitor color axes. Now you can observe a good subject pucture and switch between them.

I fond the difference is a tiny change in green intensity totally insignificant.

Same here. I've also already checked broadcast sources(even HD) with both "default" and "monitor" choices I have set up, and I really, really have to look closely to see any differences, so either choice will suffice in any circumstance, really.

Quote:
Originally Posted by KenTech View Post

SInce broadcast TV is poorly color controlled anyway, I'll happily "sacrifice" that for accurate DVDs. So 14-15-6-4 is my Default. I have almost no incentive to ever switch.

If skin tone had changed noticably, I would have an issue. But a tiny green intensity change? Hardly!

Ken

Yes, what I did is set up RYR~GYB "13-15-5-3" for the "default" color axis choice avialable from user menu, and RYR~GYB "14-15-6-4" for "monitor". Only really because those are the very slightly "different" RYR~GYB settings that worked out for me, and because in the 34XBR960 for "default" color axis choice it's labled in the user menu as "emphasize red colors", and the "monitor" color axis choice is labeled as "de-emphacize" red colors - which just seems to make more sense in my particular circumstance.

On 34XBR960 with the "mode memory" function in user menu set to "on", I can use different "color axis"(either default or monitor) per pic mode, and per "input". So, since in my case, "13-15-5-3" is "most accurate" for everything (including DVD at 480i/DRC or 480p) except 720p/1080i HD -- I am using "default" color axis choice for all inputs except from internal tuner and MS, which I have set to "monitor", currently.

If there is anything that might cause me to change my setting for broadcast sources(from internal tuner), from previous experience with this sort of thing, it might be for the orange portions of Cincinnati "bengals" graphics or uniform as received from the "poorly controlled color" from some broadcast sources - mainly that of course involves SD or upconverted local/syndicated SD programming for various reasons, not HD from the network feeds.

I just find it extremely annoying when the "orange" portion of those graphics turns out being more of a weird "sort of" pinkish+reddish+Magnenta color ... And, I've found that easily occurs with a slight tad less "green intensity"(any "red push" doesn't help either in a slightly different sort of way) - even from a network feed during a game. And, As it so turns out, my "monitor" setting will probably end up being the best choice for this, anyway.

Jeff
Nitewatchman is offline  
post #918 of 2969 Old 02-05-2006, 11:04 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

On 34XBR960 with the "mode memory" function in user menu set to "on", I can use different "color axis"(either default or monitor) per pic mode, and per "input".

I thought of this, too, but then one would be sacrificing one of the (only) four Picture Modes for a hairsplitting difference.

Consider how easy it is to "use up" the four modes for more important things. (Remember that Sony's mode-naming has no inherent significance. You just can't change it.) For example, my choices are:

Standard = My best overall settings at the current time. Used for all viewing unless something "special" is needed.

Movie = Exactly the same as Standard, except for gamma. For this mode, GAMM = 1 and the gamma amounts for that settings are set to 3 and GAMS to 0. Used to correct intolerable "gloom" in some programming.

Pro = More conservative settings than Standard, generally based on SYSM = 3, not 2, and tuned for smoothness. I try this when I think I am overdoing "enhancements" in Standard. 90% of the time, I go back to Standard.

Vivid = Ad-hoc experimentation. I make it identical to Standard, then change one or more settings to see if I have learned something new that's of value. I'm currently experimenting (again) with the VM-shaping parameters in 2170P-3. Another example: If I think I have spotted an issue with Standard, I "fix" it in Vivid, then live with it for a couple of days, switching back and forth to confirm results. If I have discovered a legitimate improvement, I transfer it to Standard, and then Vivid becomes available for another experiment. Vivid has become my "scratchpad" for settings.

Nowhere in this scheme is there room to assign an insignificant color-matrix shift to specific inputs. Mode Memory is OFF, since I might use any mode for any input. Obviously, I wish there were, oh, about a dozen or more possible preset "modes" some of which could be made input-specific.

Think of the problems folks have who simply want to balance black level among 5 discrepant inputs and have no inclination to do it in service mode! They're always playing with Brightness of leaving it set and complaining about the picture quality. (Or they don't see any difference. Ignorance is bliss!)

KenTech
"We all get smart slowly."
KenTech is offline  
post #919 of 2969 Old 02-05-2006, 11:59 AM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

I thought of this, too, but then one would be sacrificing one of the (only) four Picture Modes for a hairsplitting difference.

Consider how easy it is to "use up" the four modes for more important things. (Remember that Sony's mode-naming has no inherent significance. You just can't change it.) For example, my choices are:

Yes -- I agree. I don't use different "Color axis" settings for different pic modes, and wouldn't want to for the reasons you mention. Which is part of the reason why I want "color" to match as much as possible from all inputs/sources/scan rates, was just posting that it is possible.

I actually brought it up to address the capability with "mode memory on" to use different RYR~GYB Color axis settings in an "input specific" manner, but whenever possible I usually like to be as complete and accurate as is possible so also mentioned it's possible to have different "color axis" settings for the pic modes as well. As I explained, for instance, I'm using "default" color axis choice for everything except sources from the internal tuner(NTSC/QAM or ATSC) and Memory stick for which I'm currenlty using "monitor" color axis choice. That's the only "difference" between Inputs I'm actually using "mode memory= on" for. [update : oops. I lied I'm also having to use user menu Hue setting of "G1" or "G2" for my Dish network receiver via S-video, and Hue = "0" for a S-VHS deck via S-Video. [end update]

I use the pic modes pretty much exactly as you do, well not just pretty much --- I have them set up+use them exactly the same way as you are. I do keep trying to come up with some sort of other good use for "vivid" for when (if ever, LOL) I get done "experimenting".

I have been experimenting a bit with ABL settings, and if it were possible(which doesn't seem to be the case for YMLT), and If I ever get done "experiementing" with "vivid" it might be nice for instance to be able to set up say, "YMLT=2 or YMLT=1" with "vivid", as opposed to "YMLT=3" for the other pic modes.

Jeff
Nitewatchman is offline  
post #920 of 2969 Old 02-06-2006, 08:39 AM
Advanced Member
 
Napoleon D's Avatar
 
Join Date: Oct 2005
Posts: 589
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 8 Post(s)
Liked: 10
Ken - I had asked this question on another thread here but this one might seem more appropriate.

For your Pro mode, where you have SYSM = 3, how are you working with that smoothness (which SYSM-3 creates) to bring out sharpness? SYSM-3 i think is a much more attractive picture that the 2 setting. Although SYSM-3 requires some kind of sharpneing, more definition to make the details more visible. You mentioned you used a MIDE 60 setting with the appropriate sub-category settings. I tried this, but it still did not sharpen up the picture enough. My calibrator has it all set this way for Pro (with a light MIDE setting). All edge enhancemnts (except for MIDE) are turned off. He also has sharpness slider in middle, which is convenient place to keep it. I was curious how you were sharpening the SYSM-3 image to the point where fine detail is easily visible, but not overshooting/ringing etc.

Thanks Ken!
Napoleon D is offline  
post #921 of 2969 Old 02-06-2006, 02:41 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 Napoleon D View Post

Ken - For your Pro mode, where you have SYSM = 3, how are you working with that smoothness (which SYSM-3 creates) to bring out sharpness?

First, I think the "smoothness" is an illusion. If you reduce Sharpness for the SYSM=2 setting, you get smoothness, too -- it's just a different look. And the reason is probably that SYSM=2 really does reach into the very high video frequencies for its sharpening, and too much looks "scratchy."

To address the several questions you ask . . . My working model (in my head) is that these controls and filters, taken together, act as sort of a "graphic equalizer" for the video spectrum, analogous to the same for the audio spectrum in a stereo system. One additional factor, hard to find an "equalizer" analogy for, is the matter of visual unsharp-masking, which, for a scanning TV, is a matter in the time domain, not the frequency domain. The effects are positioned on-screen relative to the details on which they are acting. Example: a sharp vertical edge from 20% to 80% gray (dark on the left) would have a slight dark line added just before the transition to 80% and a slight bright line added just after it. That's if you turn the effect up too high to prove the point; normally one would add just enough of this effect to cancel an inappropriate softening of that transition becuase of losses in the video chain somewhere. So it's a great conpensator. I think that is SYSM=2's greatest strength, and SYSM=3 can't compete.

Here's sort of how I see these effects:

SYSM=2 = Great unsharp masking, especially with PROV=3. Affects mid-detail and up, the effect increasing with finer detail, like a video frequency-response boost-ramp. Overdone, it looks scratchy. My Sharpness is generally set about 25-31 for this. No effect is NOT at Sharpness slider=0 but more like at 10-12. Below that, Sharpness has a softening effect!

SYSM=3 = Effect occurs almost entirely to the right of the detail affected, and it appears to me more like a simple upper-frequency boost-plateau or peak in the small-detail frequencies. Overdone, it produces a bright line following a dark-light transition -- overshoot, but not much multiple-cycle ringing. It does not reach into the highest video frequencies, and so the finest detail remains unaffected. That's likely why it looks so smooth. (But I don't want Bruce Willis' face in Fifth Element to look smooth!) I choose to build my enhancement base on SYSM=2, unless the video source is way ovesharpened to begin with or has lots of film grain. (But that's what the DNR in my DVD player is for, no?) No effect is 0 on the Sharpness slider, unlike SYSM=2.

Now for the filters. The coarsest for DRC-processed video (480i) is 2103-1/SHAP. Above 2 or 3, it gives that cheaply-sharpened "TV" look to good video. I hate it, and for now it is set to 0, with PREO set to 0 or 1 to avoid a slight ringing side-effect. Maybe SHAP=1 would be helpful, but I haven't tried that yet.

Slightly finer is the boost in 3D-COMB #18, YPFG, for SD from RF and CV sources. I don't want it either, and so it is set to 8, which Sony indicates is zero effect. (Below 8, the image is softened.)

Medium-fine detail is enhanced by 2170P-3/F1LV settings higher than zero. Good for SD sources.

Only the very finest detail is enhanced by setting MID5/MHYE from 1-7, providing MYHO=0 and MHYL=3. If MHYO is set to 1, the affected video frequencies are lowered, and you can again choose MHYE from 1-7 for increasing effect. in other words, MHYO=1 makes the enhancement coarser -- which is just what is needed for SDTV, but not HDTV or fine DVDs. I find it hard to see the difference between the MHYO=1 group of MHYE settings and the same thing applied by 2170P-3/F1LV. Maybe the affected video frequencies are very similar.

For fine sources from DVD, I do all fixed sharpening in MID5 with MHYL=3, MHYO=0, and MHYE at 3 or 4. For HD sources, roughly the same, with MHYE at 2 or 3.

**************
So -- what to add to SYSM=3 to try for increased detail? I would use a column in MID5, set MHYL=3 and MHYO=0, and crank up MHYE from 1-7 to see if you can accomplish what you seek. That's the only filter that affects the finest details and textures. All others affect coarser detail, contributing to that "scratchy" thing.

I am baffled by the purpose of 2170P-3 nos. 10-12, LTLV thru CTLV. Sony calls it "LTI" level. I know it's zero at __LV = 0, and LTMD positions the effect such that LTMD=1 fattens vertical black lines, and LTMD=0 fattens white lines. If you look at frequency-sweep and resolution patterns on AVIA, you can see that the "LTI" effect is mostly with large details in the picture -- "midrange" rather than high frequencies. I have never found it to complement any image, so I never use it. Wish I knew what its intended purpose was; it's not just a filter but appears to operate in the time domain.

An aside: Filters that operate in the time domain give themselves away by shifting the picture to the right -- a delay. Compare 2170P-3/SHF0=0 with 1 (off). Or SYSM=2 with SYSM=3. Not a bad thing, if something important is accomplished. But it's a giveaway that there is not just simple frequency filtering going on!

2170P-3/SHF0 is another puzzle. Sometimes I have convinced myself that setting to 0 enhances SD video, but then I notice that fine textures are slightly attenuated by it, even on SD material -- and so I want nothing ot do with it. I have used my Vivid mode to experiment to see if I can live with something like this. (See my recent post #920.) I always keep returning to SHF0=1 becuase SHF0=0 does something ugly to the picture. The whole "LTI" group is even worse, attenuating fine detail a bit. No thanks!

The bottom line is that it is a balance among these choices that give your display a certain "look." With most real-world video sources, especially 480i, a "flat" video frequency response is a pretty soft or dull look, and some enhancements nicely fool the eye at a decent viewing distance into thinking the picture is better than it really is. That's the "art" of this and the reason to experiment and share results.

BTW, VM plays its part! Used discreetly, it does nothing negative, helps to compensate for the finite resolution of a CRT, and suppresses some sharpening overshoot. I use it on everything, even HD. I have spent a week at a time living with certain VM settings in the 2170P-3 group. Current favorites are:

VMCR~VMDL = 0-3-1-4 with actual VM level (Medium) of 6 (confirmed with VMLV).

Sorry to have been so long-winded, but getting the image right is a pet endeavor of mine, being a Photoshop junkie and a photographer. I think this series of TVs is absolutely amazing in how well they're engineered to get it right! If they were only adjusted this way from the factory . . .

KenTech
"We all get smart slowly."
KenTech is offline  
post #922 of 2969 Old 02-06-2006, 03:54 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Napoleon D,

Not meaning to "butt in" here -- just in case it is of any assistance, per some of the comments in your post in other thread --- if you haven't seen it you might want also want to check out Ken Tech's more recent(from Jan 06) posts and findings concerning "image processing". Most significantly, post Post 707 in this thread and the attached file "IpChart05tall.pdf".

Quote:
Originally Posted by KenTech View Post

First, I think the "smoothness" is an illusion. If you reduce Sharpness for the SYSM=2 setting, you get smoothness, too -- it's just a different look. And the reason is probably that SYSM=2 really does reach into the very high video frequencies for its sharpening, and too much looks "scratchy."

I think so too at this point. One recent experiment I tried involved using your findings from Post 707 .

Utilizing those findings, I set up both "vivid" and "standard" pic modes so everything was the same EXCEPT I set SYSM=3 for "vivid" and SYSM=2 for "standard". I then observed the results with various programming(including HD/SD from ATSC OTA sources) via switching between those 2 pic modes, in some cases changing the sharpness slider a bit.

Much to my surprise, there was little, (nothing of consequence really) I liked more about the SYSM=3 setting when compared to SYSM=2 in this way, although that had never been the case for me when I'd tried similar experiments before while using (mostly) your earlier image processing findings.

What is also puzzling, I think, are not only the SYSM "pro" mode set defaults in some cases, but also the description from Sony in the XBR2 service manual for SYSM. It says :

SYSM Bandwidth [0: NTSC, 1: FF, 2: HD, 3: DTV]

Ok, so "bandwidth" makes sense ... However --- why #2 for "HD" and #3 for "DTV?" In other words, what is that supposed to mean ?

"DTV"(for digital terrestrial Television(ATSC) I assume - Includes various 480i, 480p, as well as 720p and 1080i formats, the HD formats being 720p and 1080i) ... Just a WAG -- Could SYSM=3 perhaps be a setting which is there as a means for reducing visable MPEG2 compression artifacts when too high compression ratio is utilized ? Which of course, is an issue that is going to affect "HD" or SD MPEG2(or probably other codecs as well) sources other than via "DTV" ...

If there is any one setting I'd like to hear about in detail from the engineers at Sony who worked on this, it would be SYSM. If we could get a Sony Engineer as a "Genie in a bottle" sort of thing, what would be your 3 questions

Quote:
Originally Posted by KenTech View Post

Current favorites are:

VMCR~VMDL = 0-3-1-4 with actual VM level (Medium) of 6 (confirmed with VMLV).

Allways something new to try -- Just kidding! Looking forward to it, as I've noticed VMCR~VMDL 0-3-2-8 with VM level 6 (set up for "medium" VM choice for user menu - which is where I generally most like it) seems to perhaps be doing a little "too much", to the point of it perhaps involving a bit of a "noticable" VM look to it on say, small, bright white text.

Jeff
Nitewatchman is offline  
post #923 of 2969 Old 02-06-2006, 04:40 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

SYSM Bandwidth [0: NTSC, 1: FF, 2: HD, 3: DTV]

Ok, so "bandwidth" makes sense to a degree ... However --- why #2 for "HD" and #3 for "DTV?" In other words, what is that supposed to mean ?

Another Sony description: 0 & 1 = narrow peak, 2 = broad peak, 3 = flat. I don't believe a word of this!
Quote:


I've noticed VMCR~VMDL 0-3-2-8 with VM level 6 (set up for "medium" VM choice for user menu - which is where I generally most like it) seems to perhaps be doing a little "too much", to the point of it perhaps involving a bit of a "noticable" VM look to it on say, small, bright white text.

I was experimenting there with using VM-delay (VMDL) for the first time to keep the "center" of hot, bright objects from shifting to the left with the addition of VM. I now think VMDL=8 was a bit much, and 4-6 is better. Further, the higher you set VMF0, the more low video frequencies are excluded from VM. If all you wanted is more detailing, this might work, but VM is too crude for mere sharpening. The effect is far more subtle than that, and the ability to reduce overshoot increases with the lowering of VMF0. There's a big jump in VMF0 from 1 to 0; maybe the high-pass filter is removed entirely, who can tell. But VM effects get a little "fat" with VMF0 at 0. (Notwithstanding Sony's choice of VMF0=0 for HD!) I am trying this other compromise for a while: 0-3-1-4 instead of 0-3-2-8. Last night's viewing of DVD and ordinary SD cable broadcast was very fine. Is it me, or is it VM?

BTW, in the case of "tuning" VM, I believe the effects are subtle and not readily appreciated when quick-switching between otherwise-identical modes. I.e. if I set up both VM groups above for two modes, then switch back and forth during a live local newscast, I'm hard-put to see any difference. But when (how to say this?) my eyes have "locked in" on the picture, I eventually notice something very right or wrong about it -- which is the reason I have canned 2170P-3/SHF0=0: looks okay on a quick-switch, but after a while I notice that something isn't right. So I really believe that, on these subtle changes, one should live with an evening's programming to see if your recent changes "fit."

KenTech
"We all get smart slowly."
KenTech is offline  
post #924 of 2969 Old 02-06-2006, 05:32 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

SYSM Bandwidth [0: NTSC, 1: FF, 2: HD, 3: DTV]

Ok, so "bandwidth" makes sense ... However --- why #2 for "HD" and #3 for "DTV?" In other words, what is that supposed to mean ?

"DTV"(for digital terrestrial Television(ATSC) I assume - Includes various 480i, 480p, as well as 720p and 1080i formats, the HD formats being 720p and 1080i) ... Just a WAG

Please don't take these descriptions too literally! I have begun to believe they are (1) badly-translated from Japanese and (2) full of errors. In the last charts I published, I copied the column labels from the service charts, but in many cases they make little sense, confusing ATSC, "V5, V6," YUV, etc. Sometimes OTA signals are referred to as "RF," in other places "ATSC." QAM (digital over cable) is never mentioned; AFAIK, QAM is *not* a subset of ATSC.

Eventually I hope to make corrections in the charts, but not just yet. I'm a little weary of testing!

KenTech
"We all get smart slowly."
KenTech is offline  
post #925 of 2969 Old 02-06-2006, 07:11 PM
AVS Special Member
 
Nitewatchman's Avatar
 
Join Date: Dec 2001
Location: Middletown, Ohio
Posts: 6,286
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 11 Post(s)
Liked: 16
Quote:
Originally Posted by KenTech View Post

I eventually notice something very right or wrong about it -- which is the reason I have canned 2170P-3/SHF0=0: looks okay on a quick-switch, but after a while I notice that something isn't right. So I really believe that, on these subtle changes, one should live with an evening's programming to see if your recent changes "fit."

I had much the same experience with SHFO. In short, I had it set at "0" for quite some time until I decided I didn't like it.

Quote:
Originally Posted by KenTech View Post

BTW, in the case of "tuning" VM, I believe the effects are subtle and not readily appreciated when quick-switching between otherwise-identical modes.

Absolutely. I'll often go weeks before deciding on implementing a change, especially as I said earlier, I don't spend a lot of time "watching TV" on this set. Which is part of why I *very* much appreciate your work on the image processing settings, and, rather than spending a LOT of time on it experimenting(although it has been fun to check things here and there and do a bit of experimenting) I've been very happy using what you've found - especially your latest findings.

Anyhow, In addition to programming material(which we are most interested in of course), I also have a few things I look at closely when making or deciding on changes.

Such as : #1) various Local news graphics.

#2) A small ID bug one of my local PBS stations uses.-- It's a small circle, inside is white background with very small black text which Says "Think TV" (the "TV" is an even smaller "superscript"). If, for example I can't read that clearly from say 10 feet away, I know there's something wrong, either with my eyes+eyeglass perscription, or something has "changed" for the worse with the "image processing" settings.

#3). Another example(this one having to do solely with color) involving -- The EPG info from Dish network 311 receiver. For the channels I'm not subscribed to, they're "shown" in Red in the channel listing. If that red is a "cool" red, or undersaturated or oversaturated I immediately notice it.

#4). Also, concerning the 311's EPG info -- If I press the "info" button once on the remote, the programming description/guide/etc. pops up with small, bright, white "hot" lettering on a "clear" background". For instance, Some even Subtle changes in VM settings can especially be visable here - but we aren't talking big differences, and it might take a while before I notice a subtle change has changed things for the worse, or better for that matter. Press the info button a 2nd time, and the text turns black with a fairly bright "eggshell white" background(it's a rather large "box" which fills about 2/3 or more of a 4x3 frame). Any "Ghosting" issues can be especially apparent around that black text.

Quote:
Originally Posted by KenTech View Post

Please don't take these descriptions too literally! I have begun to believe they are (1) badly-translated from Japanese and (2) full of errors.
."

That was one of the exact points I was trying to make, when I said "what is that supposed to mean?" -- Perhaps another clarification that might help --" WAG = Wild A** Guess. "

It's sort of like those "super green motherboard" manuals I used to seem to come across a lot in the Mid-90's that were written in Taiwan/china and translated to English. Stuff like "Jumper the M54 setting PIO mode 4 "/etc ...

Just especially gets very confusing at times in this case when your going through and looking at the charts at all the "columns ... "

Quote:
Originally Posted by KenTech View Post

AFAIK, QAM is *not* a subset of ATSC.

It's not, QAM is the signal modulation that is most used by cableco's for "digital" cable. 8VSB(Trellis coded 8-level vestigal sideband) is signal modulation used for OTA "digital"(ATSC) signals in U.S. There is also a little known signal modulation in the ATSC specs for use by digital cable that is "16VSB", and requires less FEC(forward error correction) than the OTA version. I don't think anyone is using it in U.S. as I think Cableco's(and decoders/"digital cable ready sets") have pretty much "adopted" QAM as the standard they want to use, but not sure.

16VSB has about the same data payload in a 6MHZ RF channel as QAM256 -- about 38Mb/s . Because of the extra error correction needed for "robust" OTA transmission/reception, with 8VSB, the data payload rate is 19.38Mb/s for 6MHZ wide RF channel. Thus, cableco's can "fit" 2 entire 19.38mb/s datastreams from 2 seperate digital ATSC stations into a single 6MHZ channel. Thus, there are often "two" slots per QAM channel. I've even read that it might be possible it can be 3 stations and 3 slots in one 6MHZ channel with the use of "rate shaping"(which doesn't require an add'l decoding/reencoding process).

With these sets which are "digital cable ready", the receiver "front end" has ATSC+QAM "tuners". In most cases, (in all cases I know of in my area), when a cableco distributes a station's digital "signal", the Transport stream(MPEG2 Video/AC-3 audio/etc) remains the "same" as what OTA viewers get(even the PSIP info), the only difference is the cableco doesn't Use the ATSC 8VSB "RF"(signal modulation) portion of the signal, and instead distributes the datastreams via QAM 256 on their system.

I suppose MPEG2 could in a way be thought of as usedas an important "subset" of ATSC - --- and you'll find a LOT of "MPEG2" info in ATSC white papers, but of course MPEG2 is not a "ATSC only" thing(DVD's use MPEG2 for example), and other codecs besides MPEG2 CAN be used with ATSC OTA as well -- although no current receivers I know of will work with anything else other than MPEG2.

Quote:
Originally Posted by KenTech View Post

Sometimes OTA signals are referred to as "RF," in other places "ATSC."
."

Without going into great detail or accuracy here -- in the charts, I think it's working out that "RF" is refering to the NTSC (analog) "tuner" -- including for either OTA or analog cable(which uses the same "tuner", and channel numbers/frequencies on VHF as OTA, but different channel #s along with in part different frequencies on UHF cable bands). I think ATSC on the charts is referring to both ATSC(OTA) and QAM -- as, even though the tuners and RF signal de-mods that need to occur are "different" for 8VSB+QAM -- after that, For instance, I think whether you're getting a datastream originating from a station's MPEG2 encoeder via OTA via the OTA 8VSB RF signal or via cable via the QAM RF signal the cableco is sending via the "wire", the MPEG2/AC-3 audio/PSIP/etc. streams involved are going to be the same, and are also going to use the same MPEG2 decoder in the set+ probably the same signal paths(at least mostly)/etc.

Update :

I can actually hook an antenna up to the "cable" RF input on KD34XBR960, and it will work just fine with 8VSB ATSC signals on VHF(I have a local station transmitting on VHF 10) via the "cable" input - even though it isn't QAM - shows up as C9.1 with the PSIP virtual remapping. It Will NOT work for UHF ATSC stations.

end update

Quote:
Originally Posted by KenTech View Post

I'm a little weary of testing!

Me too! I do know I had the chance to watch quite a bit of TV on the XBR960 yesterday, Including the ABC HD superbowl coverage and everything looks so good at this point -- I don't feel all that motivated to do too much more of anything .. Of course, I said that about 6 months ago as well ...

Jeff
Nitewatchman is offline  
post #926 of 2969 Old 02-07-2006, 07:53 AM
Member
 
loadams's Avatar
 
Join Date: Apr 2003
Location: Mount Pleasant,SC
Posts: 158
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Ken, don't mean to ask a silly ? here, but............

I've got the older, bigger brother 40" gorilla. Now with 9 of 10 dvd rentals being in 16x9, why would I want to continue using 480p (component) when the set upconverts those dvds to 960i anyway? Pros.....cons ?

And that being said..... I have balanced my color decoder (13,14,6,3) with 480i and tweaked 480p with CXA2151, however, there is NO WAY to have a happy medium with 480p and 1080i on my set !!! Which is giving me a good reason to dump 480p altogether for the sake of decent color decoding in 1080i.

I guess I'm asking............ if 960i doesn't touch CXA2151, dump 480p ? Your opinion ?
loadams is offline  
post #927 of 2969 Old 02-07-2006, 08:09 AM
AVS Special Member
 
SurfingMatt27's Avatar
 
Join Date: Dec 2005
Location: Massachusetts
Posts: 2,770
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by Napoleon D View Post

Ken - I had asked this question on another thread here but this one might seem more appropriate.

For your Pro mode, where you have SYSM = 3, how are you working with that smoothness (which SYSM-3 creates) to bring out sharpness? SYSM-3 i think is a much more attractive picture that the 2 setting. Although SYSM-3 requires some kind of sharpneing, more definition to make the details more visible. You mentioned you used a MIDE 60 setting with the appropriate sub-category settings. I tried this, but it still did not sharpen up the picture enough. My calibrator has it all set this way for Pro (with a light MIDE setting). All edge enhancemnts (except for MIDE) are turned off. He also has sharpness slider in middle, which is convenient place to keep it. I was curious how you were sharpening the SYSM-3 image to the point where fine detail is easily visible, but not overshooting/ringing etc.

Thanks Ken!

You want a sharp picture, turn all the MIDE5 settings to 0, your PQ will be ultra sharp with no enhancements neccessary.Just look at text and tell me how perfect it is.
SurfingMatt27 is offline  
post #928 of 2969 Old 02-07-2006, 09:00 AM
AVS Special Member
 
justsc's Avatar
 
Join Date: Jun 2005
Location: San Jose, CA
Posts: 2,689
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by SurfingMatt27 View Post

You want a sharp picture, turn all the MIDE5 settings to 0, your PQ will be ultra sharp with no enhancements neccessary.Just look at text and tell me how perfect it is.

Matt,

I'm with you on almost all of the settings you have suggested - and have a very nice picture - thanks! I realize now that "adding" to the original picture (signal) can actually take away from the overall fidelity.

But, what if there's elements in the signal that I'm not seeing - elements that could make it even sharper? This is the direction my thinking is going. Like when raising the black level on a overly dark "dusk" scene and seeing how much detail I was missing (just an example).

I'm just thinking there may well be more "sharpness" to be enjoyed by "bringing it out" rather than "adding it in."

So, I'm focusing on finding ways to "expose" levels of fidelity that are already there but I haven't seen - yet.

This is why I find the latest "back and forth" between KenTech, Nitewatchman and Napoleon D so interesting. This is especially important since my 34HS420 doesn't have the SFP tube.
justsc is offline  
post #929 of 2969 Old 02-07-2006, 09:07 AM
Advanced Member
 
tlniec's Avatar
 
Join Date: Jan 2006
Posts: 742
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 13
Great thread, lots of information here! I have a question regarding a standard-def Sony CRT (KV-32FS100). I've tweaked the user menu settings and gotten some good results. However, I think I'm ready to take the next step and delve into the service menu. I have logged all the existing service menu settings (but have not yet changed any). Do you guys have any advice for a Sony CRT that only sees 480i (maybe 480p for DVD) input from DVD over component, VCR over composite, and Digital Cable over composite? I suspect I might be able to get by tweaking a relatively small subset of the parameters KenTech and Nitewatchman have posted, but I want to make sure. Thanks!
tlniec is offline  
post #930 of 2969 Old 02-07-2006, 04:40 PM
AVS Special Member
 
SurfingMatt27's Avatar
 
Join Date: Dec 2005
Location: Massachusetts
Posts: 2,770
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:
Originally Posted by justsc View Post

Matt,

I'm with you on almost all of the settings you have suggested - and have a very nice picture - thanks! I realize now that "adding" to the original picture (signal) can actually take away from the overall fidelity.

But, what if there's elements in the signal that I'm not seeing - elements that could make it even sharper? This is the direction my thinking is going. Like when raising the black level on a overly dark "dusk" scene and seeing how much detail I was missing (just an example).

I'm just thinking there may well be more "sharpness" to be enjoyed by "bringing it out" rather than "adding it in."

So, I'm focusing on finding ways to "expose" levels of fidelity that are already there but I haven't seen - yet.

This is why I find the latest "back and forth" between KenTech, Nitewatchman and Napoleon D so interesting. This is especially important since my 34HS420 doesn't have the SFP tube.

Well Steve,..You can't really add detail that is'nt there,all you will be doing is introducing stuff into the picture that does'nt belong resulting in noise, or artifacts.My advice is to keep the signal clean and you will thank yourself later on trust me

I was skeptical as you a year back and found out that it's not really neccessary at all, since i noticed all your doing is adding false information that is'nt there.

I just keep it simple and feed the signal to my tv without any processing, easy breezey

As a matter of fact i'm thinking of recalibrating my tv, mainly only touchup.Brightness and color mainly is my only issues, since i think i may have a low setting right now and it could be higher,And color although good, i think i can tweak it a bit.It may not be D6500k but it will look good none the less by simple refinements.

Beback in a few.
SurfingMatt27 is offline  
Reply Direct View (single tube) CRT Displays

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


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