WinDAS White Point Balance guide for Sony Trinitron CRTs

Yep I'm cool with tinkering. Only shitty part is I can only do this at night when I have time and when there's zero ambient light.
 
NP.

Worked all night on calibrating... but after all that, it didn't save properly and the screen was messed up.
as in the colors were off? that happened to me the first two times I did it on my fw900. no idea why, but I've never had the problem since.
 
as in the colors were off? that happened to me the first two times I did it on my fw900. no idea why, but I've never had the problem since.

What I do now is stop the WPB procedure before it goes to 6300k, select "final setting" and okay that. No problems at all now.

The trickiest portion is determining a good G2 that won't kill all the details.
 
What I do now is stop the WPB procedure before it goes to 6300k, select "final setting" and okay that. No problems at all now.

there is no 6300K, the first is 9300K, then 6500K, then 5000K, then the sRGB contrast adjustment. Which one are you stopping it before? And probably not a good idea to stop it prematurely unless nothing else works.

The trickiest portion is determining a good G2 that won't kill all the details.

did you read the last portion of the guide? This fine tunes the LUT so you don't kill the details even while having a low G2.
 
did you read the last portion of the guide? This fine tunes the LUT so you don't kill the details even while having a low G2.

I think that is a big flaw in your guide. Gamma above 2.4 will lead to a lot of detail loss in darker areas, something which can't be fixed with a LUT since many games overwrite it and prevent forced restoration. We should aim for 2.2-2.4 from the monitor itself.
 
I think that is a big flaw in your guide. Gamma above 2.4 will lead to a lot of detail loss in darker areas, something which can't be fixed with a LUT since many games overwrite it and prevent forced restoration. We should aim for 2.2-2.4 from the monitor itself.

Many games do overwrite the LUT, but not all do. And those that do usually have controls that adjust the LUT from within the game, so detail can be uncrushed. Also, there are tools like cpkeeper that attempt to maintain the current LUT (although this may not work with some games).

For anything other than gaming (movies, desktop, viewing photography, web surfing), the Argyll LUT adjustment works perfectly.

You can certainly adjust G2 so that you end up with 2.2 to 2.4, but your blacks won't be as deep.
 
there is no 6300K, the first is 9300K, then 6500K, then 5000K, then the sRGB contrast adjustment. Which one are you stopping it before? And probably not a good idea to stop it prematurely unless nothing else works.



did you read the last portion of the guide? This fine tunes the LUT so you don't kill the details even while having a low G2.

Sorry I meant 9300k, then 6500k.

I didn't have enough time to do the entire WPB procedure.

I read the final bit about LUT. Will try it out once I get a chance.
 
I think you'll love the image quality once you've implemented the Argyll LUT. It's remarkable to see all that detail recovered while maintaining the incredible image depth afforded by the super deep blacks.

Another thing that I haven't yet included in the guide is the notion that deeper blacks also means that the primaries reach their full saturation potential (i.e. you end up with a larger color gamut). With even slightly raised blacks, a pure red signal (RGB [255 0 0]) will actually have a bit of residual luminance in the blue and green channels, and this will desaturate the red primary. Same for each of the other primaries. This is just one of the many reasons that having deep blacks is important.
 
I think you'll love the image quality once you've implemented the Argyll LUT. It's remarkable to see all that detail recovered while maintaining the incredible image depth afforded by the super deep blacks.

Another thing that I haven't yet included in the guide is the notion that deeper blacks also means that the primaries reach their full saturation potential (i.e. you end up with a larger color gamut). With even slightly raised blacks, a pure red signal (RGB [255 0 0]) will actually have a bit of residual luminance in the blue and green channels, and this will desaturate the red primary. Same for each of the other primaries. This is just one of the many reasons that having deep blacks is important.

Yep. I just need more time lol.

Also am questioning how hard/easy it will be to use LUT for different apps and games.

One question: the unit I got from Vito had its G2 either on 162 or 165. Is this normal?
 
Also am questioning how hard/easy it will be to use LUT for different apps and games.

Which apps and games specifically?

One question: the unit I got from Vito had its G2 either on 162 or 165. Is this normal?

Hard to say what normal is. I imagine it depends on the life of the tube and how the calibrator chose to set the G2. I think after adjusting my G2 it was at 138.
 
Which apps and games specifically?



Hard to say what normal is. I imagine it depends on the life of the tube and how the calibrator chose to set the G2. I think after adjusting my G2 it was at 138.

No specific apps or games -- just that I read some games won't let you change the LUT easily.

I heard that the older a tube gets, the lower the G2 needs to be for it to get good black levels.
 
No specific apps or games -- just that I read some games won't let you change the LUT easily.

As mentioned earlier, some games will allow you to respect the pre-existing LUT, and those that won't can often be forced to respect it with tools such as cpkeeper. And for those that overwrite the LUT no matter what can still be adjusted within the game to uncrush detail (e.g. gamma sliders).


I heard that the older a tube gets, the lower the G2 needs to be for it to get good black levels.

Yes I think this is true, although I think a better indication of tube life may be as described on posts 9 and 10 in this thread.
 
As mentioned earlier, some games will allow you to respect the pre-existing LUT, and those that won't can often be forced to respect it with tools such as cpkeeper. And for those that overwrite the LUT no matter what can still be adjusted within the game to uncrush detail (e.g. gamma sliders).




Yes I think this is true, although I think a better indication of tube life may be as described on posts 9 and 10 in this thread.

Ah, read 9 and 10... So the C B Max setting can help determine the life of the tube?

What's the default value of C B Max?
 
What do you mean default? Every tube is different, and even when brand new will probably have different values. The way to interpret the CB max, with respect to tube life, is to take note of how much you have to slide the value up in order to reach the luminance target.
 
What do you mean default? Every tube is different, and even when brand new will probably have different values. The way to interpret the CB max, with respect to tube life, is to take note of how much you have to slide the value up in order to reach the luminance target.

Okay I see. I'll check it out soon (not at home at the moment).
 
As mentioned earlier, some games will allow you to respect the pre-existing LUT, and those that won't can often be forced to respect it with tools such as cpkeeper. And for those that overwrite the LUT no matter what can still be adjusted within the game to uncrush detail (e.g. gamma sliders).

Game's gamma slider will lower the whole gamma, not just the bottom part. Do you have any measurements with better default gamma to see how much black levels are raised?
 
Game's gamma slider will lower the whole gamma, not just the bottom part. Do you have any measurements with better default gamma to see how much black levels are raised?

Nope, once I learned how to set deep blacks, there was no turning back for me :p

flod may have those measurements though.

As for the whole gamma being raised, this may not be a problem if setting the G2 really low actually raises the entire gamma. I do remember that before I learned how to use Argyll, I used the nvidia control panel gamma slider to adjust gamma (which is similar if not identical to what in game sliders do), and after doing so, the curve was a fairly clean fit to 2.4, which suggests that the gamma was uniformly higher across the luminance range. (edit: turns out I was wrong here, see post below).

I'll take some measurements right now without the argyll LUT, and post the results. I'll also take measurements after using the nvidia slider to give an idea of what results turning up in game sliders will do.
 
Last edited:
Here are measurements from my display without any LUT adjustment:

nnpdts.png


Here are measurements after Nvidia slider adjustment:

2lthelj.png


Your concern was well founded after all. In order to uncrush detail using a simple gamma slider, you have to raise the slider to a level where things start to look washed out.

So if there is a particular game that is important to you that makes it impossible to preserve a custom LUT, then it may be a better choice to raise your G2 level. I'd appreciate it if someone could post the gamma curves after such an adjustment. That will inform me better and I'll update the guide.

I do want to emphasize that if there if you are in a situation where your games do respect custom LUTs, deep blacks + argyll is the way to go - the precision is unbelievable. The above graphs are based on 21 measurements across the gray level. I've taken all 256 measurements after an Argyll adjustment, and the fit to the target curve is virtually pinpoint on all the levels. This is really a bonus if you want to watch well mastered HD content.
 
Thank you for the measurements! Yeah, it confirms what I've feared. Would be great if someone had measurements before LUT with some "conservative" G2 settings. Is there really such a difference in perceived contrast? Maybe we should PM Vito if he has some measurements from his calibrations.


I've found this on my computer. I think this was measured after WinDAS procedure, before LUT calibration, but I can't be 100% sure since it wasn't properly labeled and was done a couple of years ago.

AyuJPGz.png



I hate games messing with LUTs. Why in the world was that "feature" even implemented?
 
Last edited:
Is there a way to see what G2 was set to from the dat file?
 
I've opened the dat in Hex editor, but not sure which exact value is it and don't feel like manually counting to 110 :) (EDIT: It seems that G2 is the first value after REG at the beginning, in which case mine is 161).

Tried to open it with WinDAS, but it is no longer working. I've done the following:

Win7 x64

- ran patchDas.bat
- copied MSFLXGRD.OCX to C:\Windows\SysWOW64\msflxgrd.ocx
- ran regsvr32 C:\Windows\SysWOW64\msflxgrd.ocx
- set windas.exe compatibility to "Windows 98 / Windows ME"
- ran as Admin

It errors out with Astro SG communication problems (I think this was normal), but then shows "No mdl Files" and crashes.

Any ideas?
 
Last edited:
no idea, and I've only tried running the program on xp, although I understand it can be made to function in win 7.

When was last time you successfully ran it? And was it on the same OS you're running now?

Try downloading it from the link in that youtube video, and reinstalling it from scratch.
 
You need to place the msflxgrd.ocx file in all "system" folders (sys32, syswow64, system), not just the default System32 and Sys (I'm assuming you're using win x64).
Then run the .bat registration file as admin. Should work.
 
Ok, I'll try that. Though, I thought that by registering it, its location would be stored in the registry.

EDIT: Didn't work. Still the same error. Tried the one from the Youtube video, the same. Could you guys upload your version that you have working on Win7 x64?

EDIT2: Figured it out! It was the Windows DEP, doh! I was hoping that WinDAS could load dat files and show the stored settings, but it seems it doesn't support that.
 
Last edited:
Is there a way to see what G2 was set to from the dat file?

not sure if this is what's shown in spacediver's link, but basically use Help > Expert > Viewer and click around on the left hand side until you see a G2 parameter in the right side.
 
question:

do monitors show any hysteresis when adjusting settings? suppose I manage to do a very good wpb, save the settings to a dat file and redo the wpb with, for example, a lower G2. if I reload the saved dat file and wait the for monitor to re-warmup, will the monitor behave the same way as when the first wpb was done?
 
It's an interesting question - I'd be more tempted to say that if the monitor didn't behave the same way, it's not due to hysteresis, but that it's due to wonky EEPROM behaviour (which the FW900 has certainly shown to be vulnerable to).

Have u had any strange experiences in this respect?
 
a while ago in the fw900 thread I was complaining about how my dE increased from <1 to >2 overnight (and I did ensure it was warmed up). then I did some stuff including reloading my previous dat files, changing color temp from 5k to 9.3k to 6.5k, and when I returned to the original dat file, the dE went back to <1.

on the other hand my g520p seems to have very good stability. average dE without argyll was still 0.4 as of yesterday, which is the same as what it was after my incredibly lucky wpb (screenshots from pg 1).
 
Yea I think I remember, and I asked if it was perhaps due to inconsistent placement of your instrument, and you said it wasn't.

One thing that might be behind it is if you changed your brightness or contrast (OSD), and then didn't turn them back to what they were when you first measured. It's easy to fool yourself here - I think there are actually 300 brightness values (not just 100), as you have to click three times to move up a number. I remember discovering that if I clicked the brightness up once (even though it didn't change number), it had an impact on my readings. I'll have to re-do that experiment to be sure, but it's one possible explanation for your findings (assuming you even messed around with the brightness/contrast controls).
 
i'm pretty sure each osd setting has exactly 256 values because some values require 2 steps to get over whereas others require 3 steps, and 256 is the only reasonable number between 200 and 300 for the total number of steps.
 
not sure if this is what's shown in spacediver's link, but basically use Help > Expert > Viewer and click around on the left hand side until you see a G2 parameter in the right side.

It complains about monitor not being connected.
 
Are you even able to get WinDAS to function normally otherwise? Try loading up the WPB procedure (without actually starting it). Does it give you an error?
 
Oh, I really don't have my FW900 connected at the moment, new rig. I was just hoping to compare some of my saved dat files from different points of adjustment. Reconnecting it and redoing the WB procedure is something I'm planning to do next month, so I'm brushing up.
 
I think you need the cable connected to perform any operations in WinDAS, even expert viewer mode (too lazy to turn on my laptop to confirm).
 
my eyes + my program tell me that 194 is about half of 254 on my screen.

194 measures 26.73
254 measures 52.4

following up on this, here are my results from a brief test:

I created full screen image that had horizontal gratings on the left, and a homogenous patch on the right. I set the gratings to spatially alternate between video levels 20 and 60, and adjusted the video level of the homogenous test patch until the grating appeared to match the test patch.

I then measured the luminance for each of the three levels.

Here's what I found:

darker grating luminance: 0.26 cd/m2
lighter grating luminance: 3.18 cd/m2

(the average of these is 1.72 cd/m2)

The luminance of the test patch turned out to be 1.68 cd/m2.

Very close indeed!

I'm attaching the image I used to illustrate, although obviously I don't expect the two halves to match as well as they do on my own display, since this particular image uses values that matched on my particular display.

Now this was done by adjusting the patch with 8 bit precision. Using direct LUT adjustments, I should be able to achieve at least 10 bit precision.

All this is to say that it may be possible to actually measure the luminance of the really dark levels :)

Image is below:

39532_TestGrating.png
 
Back
Top