or Connect
AVS › AVS Forum › Display Devices › Display Calibration › The VideoEq -- A low cost external grayscale/gamma/cms tool
New Posts  All Forums:Forum Nav:

The VideoEq -- A low cost external grayscale/gamma/cms tool - Page 11

post #301 of 714
My VideoEQ Pro turned up yesterday, after sending in my EAP unit. With the EAP, the CMS was there, but there were no tools to make use of it. (and it was not stored in the simple HSL format it is now) I normally use CalMAN for everything, but as I am currently without a laptop, I need to calibrate the screen that the software is running on, and it seems CalMAN is not suited for that (details here) so I have had to use HCFR.

Rather than redo the greyscale with HCFR, I just loaded up the old LUT I had previously created, and then used the CMS controls on top of that. My display doesn't need much in the way of hue/saturation adjustment, just luminance really.

Unfortunately, it looks like HSL are not independent of each other with the VideoEQ CMS. The adjustments here were -2 Red, -12 Blue and -4 Magenta luminance. (the others were within +/- 0.5% according to HCFR and any adjustment made them worse)

Pre-CMS:



CMS with luminance adjustments only:


The change is most obvious with blue, as it had the biggest adjustment, but reducing luminance is clearly reducing saturation as well.


Also, if anyone is experiencing posterisation, it may be the GUI application that is causing it.

Here is my hand-adjusted LUT:


And how the same LUT looked after making use of the CMS controls in the GUI:


It is not the CMS that is causing this, as I was able to fix it by downloading the LUT with the CMS adjustments from the VideoEQ, and copying/pasting the CMS adjustments into my hand-adjusted LUT.

(note: those photographs are probably not indicative of how the image looks on my display - they are only to show the posterisation the VideoEQ GUI is introducing)
post #302 of 714
Thread Starter 
Quote:
Originally Posted by andrewfee View Post

It is not the CMS that is causing this

Are you running b25?
post #303 of 714
Quote:
Originally Posted by bodosom View Post

Are you running b25?

Yes, sorry I forgot to mention that.
post #304 of 714
Interesting, I haven't tried the same test in my system. I'd be curious to see if I have similar effects. I thought build 25 was supposed to get all the CMS stuff working properly in the GUI...maybe not.
post #305 of 714
What hand adjustments are you making to the LUT?
Can you share how you do that (methodology) ?

Also what calibration source (Disc, pattern generator) are you using with HCFR?
I have the VideoEQ Pro and any insight would be greatly appreciated.

Cheers,
Dan
post #306 of 714
Thread Starter 
Quote:
Originally Posted by Dan2112 View Post

I have the VideoEQ Pro and any insight would be greatly appreciated.

Are you following the SpectraCal forums?
post #307 of 714
Quote:
Originally Posted by bodosom View Post

Are you following the SpectraCal forums?

Yes. Of course
I was interested in the hand editing aspects of the LUT (what tweaks he was making to it) and what he was using to get the multi-point (3d) cie chart.
post #308 of 714
Thread Starter 
Quote:
Originally Posted by Dan2112 View Post

what he was using to get the multi-point (3d) cie chart.

That's a C-HCFR saturation plot.
post #309 of 714
Quote:
Originally Posted by Dan2112 View Post

What hand adjustments are you making to the LUT?
Can you share how you do that (methodology) ?

Also what calibration source (Disc, pattern generator) are you using with HCFR?
I have the VideoEQ Pro and any insight would be greatly appreciated.

Cheers,
Dan

Well when I got my VideoEQ EAP (beta) there was basically no software for creating/adjusting LUTs. There was a web-based tool, but it did not seem to be producing good results, so I didn't even bother trying it.

Instead, I took readings of greyscale points (10, 20, 30% etc.) and manually adjusted the numbers in the LUT (nothing more than a text file) until they were as close to measuring D65 as possible.

I then created a spreadsheet that expanded from those 10 or so points and created a full 1024-point LUT as required by the VideoEQ.

I suspect that my requirements are quite different from some though, being a CRT user. I need a LUT that has everything below video black at 0, and is then elevated quite a bit at 1% rather than it being a gradual decline to zero.

I was seeing quite a lot of posterisation from my initial LUT, and then moved to a 33-point LUT to see if that would help. When doing this, I found that if I interpolated the full LUT differently it drastically reduced posterisation. With that spreadsheet, I also added a graph to visualise it, and to my surprise I saw that it was close to being a straight line-I suspect this is due to the analogue nature of a CRT vs most digital displays, and that any irregularities were measurement errors:


So I then moved to a 3-point LUT, which has minimised posterisation. One point at 109%stim to ensure the highest value in the LUT is at D65 and contrast is maximised. (one point is at 1023) Luminance at that point is set with the display's contrast control to achieve my desired target value. (100cd/m2)

The next point is at 100%stim which is used to make sure luminance is at reference levels for that point (80cd/m2) and it is still D65.

The final point was at 4%stim, which was the lowest point I could get a stable, repeatable reading with my meter. (a DTP-94 profiled off my i1Pro) This was set to achieve the correct values for D65 with a 2.35 gamma.

I then interpolated between those points down to 1%, and cut off everything else at 0. This gets me D65 with a flat 2.35 gamma all the way to zero with a minimal amount of posterisation introduced and the maximum contrast that a CRT allows. (black is 'off' without the loss in shadow detail you would see without a LUT)



Because I was calibrating the same screen the software was running on, and I use the PC for playing Blu-ray, I used HCFR's own generator as it allows you to set PC/Video levels and the pattern size. (CalMAN's forces you to use either full field or 25% patterns which are unsuitable for a CRT)
post #310 of 714
Quote:
Originally Posted by andrewfee View Post

My VideoEQ Pro turned up yesterday, after sending in my EAP unit. With the EAP, the CMS was there, but there were no tools to make use of it. (and it was not stored in the simple HSL format it is now) I normally use CalMAN for everything, but as I am currently without a laptop, I need to calibrate the screen that the software is running on, and it seems CalMAN is not suited for that (details here) so I have had to use HCFR.

Rather than redo the greyscale with HCFR, I just loaded up the old LUT I had previously created, and then used the CMS controls on top of that. My display doesn't need much in the way of hue/saturation adjustment, just luminance really.

Unfortunately, it looks like HSL are not independent of each other with the VideoEQ CMS. The adjustments here were -2 Red, -12 Blue and -4 Magenta luminance. (the others were within +/- 0.5% according to HCFR and any adjustment made them worse)

Pre-CMS:



CMS with luminance adjustments only:


The change is most obvious with blue, as it had the biggest adjustment, but reducing luminance is clearly reducing saturation as well.

[...]

Hi,

IMHO, it's a real problem if setting are not independant. I remember it was a same problem on the first CMS version on the JVC HD750 (saturation and hue setting had an effect on luminance level)

I presume you know that but about luminance you can use this sheet for level control :



It's a recent result with the HD950 CMS and we can see there are a (real) problem on RED at 75 IRE.

Do you have any measure on "your" blue ?

About saturation Plot you can use another presentation on this sheet :



The below part is delta E calculation between reference on each color at different saturation level.

In this case, it appears a problem on red Ã* 50 % saturation level (delta E = 6) and at 100 % blue saturation level (delta E = 6).

I'm interested in this product (I'm an HD1 owner) but you're point is a real problem (IMHO).
post #311 of 714
I received my VEQ Pro last week and tried a first calibration yesterday. I was very pleased to get an accurate gamma curve and D65 within 3dE from 20-100% (my meter is not very accurate at 0 and 10% so they are more of an eyeballed result).

I then tried a gamut adjustment to rein in my oversaturated primaries, but while I can get xy accurate, they are all below luminance. As brightness is default 100% in the VEQ, and driving > 100% seems to make a mess of things, do I need to increase my TV 'colour' control to add luminance (and then redo saturation due to shift from increasing colour?

Also, whereas my greyscale was visually perfect from 0-100%, WTW signals are very pink. This is not overdriving contrast as they are fine with same contrast when VEQ bypassed. How can I adjust WTW to remedy this?

Thanks in advance, Carl
post #312 of 714
I did my first run last night, but I'm feeding my VEQ with 10 bit 4:4:4 colourspace output from my DVDO Edge, so I don't know if this makes the difference. I didn't notice any strange effects having watched a few test clips afterwards (though it was very late I have to admit). I use Chrompure, which made the CMS work very quick (the greyscale and gamma seemed to take me an age, but I kept reworking the gamma and this in turn knocked the greyscale out a little). It would be good to have access to the luma control in the Custom RGB adjustment page, like a Lumagen HDQ does.

I'm using an i1-LT which isn't ideal for CMS work, with it trained to use facing the projector using the offset feature in Chromapure. I'll revisit the measurements another night and check at different saturations and maybe try using the projector's colour/tint controls as well to see if I can reduce the lightness adjustments I made in the VEQ, seeing the comments above (not that I saw anything 'weird').

I have found that sometimes I get a strange 'pop art' type effect when first turning the projector on, but reloading the custom setting seems to clear this: Select any preset, then reselect the custom setting you wish to use and it then clears.
post #313 of 714
Quote:
Originally Posted by Carl Bradshaw View Post

I received my VEQ Pro last week and tried a first calibration yesterday. I was very pleased to get an accurate gamma curve and D65 within 3dE from 20-100% (my meter is not very accurate at 0 and 10% so they are more of an eyeballed result).

I then tried a gamut adjustment to rein in my oversaturated primaries, but while I can get xy accurate, they are all below luminance. As brightness is default 100% in the VEQ, and driving > 100% seems to make a mess of things, do I need to increase my TV 'colour' control to add luminance (and then redo saturation due to shift from increasing colour?

Also, whereas my greyscale was visually perfect from 0-100%, WTW signals are very pink. This is not overdriving contrast as they are fine with same contrast when VEQ bypassed. How can I adjust WTW to remedy this?

Thanks in advance, Carl

Yes we allow you to push more than 100% luminance and saturation but unless the display has some head room they may not work right. What we recommend is setup the display to be in wide gamut mode so you can de-saturate which should give you some head room to run luminance over 100%.

Even thought we can control every aspect of the signal going to the display we can't make the display to do more than what it is capable of for luminance and saturation we can only take away. Its a physics thing.
post #314 of 714
Quote:
Originally Posted by andrewfee View Post

My VideoEQ Pro turned up yesterday, after sending in my EAP unit. With the EAP, the CMS was there, but there were no tools to make use of it. (and it was not stored in the simple HSL format it is now) I normally use CalMAN for everything, but as I am currently without a laptop, I need to calibrate the screen that the software is running on, and it seems CalMAN is not suited for that (details so I have had to use HCFR.

Rather than redo the greyscale with HCFR, I just loaded up the old LUT I had previously created, and then used the CMS controls on top of that. My display doesn't need much in the way of hue/saturation adjustment, just luminance really.


Unfortunately, it looks like HSL are not independent of each other with the VideoEQ CMS. The adjustments here were -2 Red, -12 Blue and -4 Magenta luminance. (the others were within +/- 0.5% according to HCFR and any adjustment made them worse)

The change is most obvious with blue, as it had the biggest adjustment, but reducing luminance is clearly reducing saturation as well.

A couple of points to keep in mind when calibrating the VideoEq.

1) The LUT and CMS are not exactly independent, in that the LUT is applied, then the CMS adjustment. Meaning that, for example, you can zero out the Red component in the LUT, and never get it back with CMS.

2) The GUI (right now) won't preserve hand-edited LUTs. When the GUI reads in a LUT, it preserves the 11 control points, and then (re)interpolates the values in between. This works well for those that are not hand-editing, but doesn't work so well for those that use other tools for LUTs and not CMS.

3) The main (conceptual) difference between saturation and brightness is that reducing saturation pulls a color toward white. Reducing brightness pulls it toward black. So either control will pull the chromaticity toward the "white" point on the CIE diagram, just that brightness also reduces the intensity. If you are not looking at the big-Y axis, the brightness will seem to desaturate as well.


-Eric
post #315 of 714
Quote:
Originally Posted by elhernes View Post


2) The GUI (right now) won't preserve hand-edited LUTs. When the GUI reads in a LUT, it preserves the 11 control points, and then (re)interpolates the values in between. This works well for those that are not hand-editing, but doesn't work so well for those that use other tools for LUTs and not CMS.

Is it correct to assume a future version of the GUI will support loading/preserving the static values in between (for those with their own tools)?
post #316 of 714
Quote:
Originally Posted by elhernes View Post

2) The GUI (right now) won't preserve hand-edited LUTs. When the GUI reads in a LUT, it preserves the 11 control points, and then (re)interpolates the values in between. This works well for those that are not hand-editing, but doesn't work so well for those that use other tools for LUTs and not CMS.

I suspected that might be the case, though I didn't have time to test it. I'm curious about why it introduced so much posterisation though - the LUT was basically a straight line:


With that data, I would have expected to see a straight line from the same 11 points anyway.

Quote:
Originally Posted by elhernes View Post

3) The main (conceptual) difference between saturation and brightness is that reducing saturation pulls a color toward white. Reducing brightness pulls it toward black. So either control will pull the chromaticity toward the "white" point on the CIE diagram, just that brightness also reduces the intensity. If you are not looking at the big-Y axis, the brightness will seem to desaturate as well.

I'm not sure I follow. To reduce the luminance of red, you reduce the amount of red.
To reduce the saturation of red, you add green & blue.
To change hue, you add green or blue.

Should they not be independent of each other? (because luminance is the only one that changes the red value, and hue/sat are adding green/blue)

I don't see why reducing the luminance of a colour should also be desaturating it - that's the same thing a display's 'color' control does, not what I would expect from a 3D CMS.
post #317 of 714
I wanted to share my results after a late night of tweaking my Pioneer 6020.

I wound up with a Gama around 2.3, perfect primaries (CalMAN says they are perfect, but HCFR says they are off a bit ) and secondaries (except for Blue which is a known issue with the Kuros). I tweaked the grayscale in on the TV first using ControlCal (Thanks D-Nice). This yielded excellent grayscale from 40-100, but blue was high in the lower grayscale (10-30).

I started with a gamma of about 2.2 but after adjustmenting the primaries and secondaries the resulting gamma was 2.3. I am not sure why that is, perhaps someone can explain that. I then played with the LUT to tweak in the grayscale down low. Once done the image was in a word - Stunning. (I will try and post the same Ratatouille before and after as above)

One thing that confused the heck out of me is the LUT/RGB screen. It is not very intuitive to use. This is my take on it:
1) Take a base measurement gamma and get an average number say 2.2.
2) Enter 2.2 in the "measured gamma" box.
3) Enter the desired gamma - say 2.3 and push the magic "Initialize Table" to recalculate the table.
4) Hit "Save to VideoEQ"
5) Repeat to fine tune.
Is this the correct procedure? I really hope CalMAN 4 automates this process to make it simple.

I used an eyeOne Monitor for all my measurments. I have an Oppo BDP-80 (configured as source direct) and DirecTV HD DVR running into a DVDO Edge into the VideoEQ and then finally into the Pioneer. Input and output are set to 1080p/24.

Pre CMS & LUT Adjustments (Saturation):


Post CMS & LUT Adjustments (Saturation):


Post CMS & LUT (Saturation Shift):
post #318 of 714
Quote:
Originally Posted by andrewfee View Post

I suspected that might be the case, though I didn't have time to test it. I'm curious about why it introduced so much posterisation though - the LUT was basically a straight line:


With that data, I would have expected to see a straight line from the same 11 points anyway.

If it was a straight line, then it should be pretty well preserved. If you get a chance, I'd be curious to see if it did change, or if the differences you are seeing are from the CMS.

Quote:


I'm not sure I follow. To reduce the luminance of red, you reduce the amount of red.
To reduce the saturation of red, you add green & blue.
To change hue, you add green or blue.

Should they not be independent of each other? (because luminance is the only one that changes the red value, and hue/sat are adding green/blue)

I don't see why reducing the luminance of a colour should also be desaturating it - that's the same thing a display's 'color' control does, not what I would expect from a 3D CMS.

It all depends on what colorspace you're using to make the adjustments (or visualize what they're doing). You are correct, if the adjustments were being made in RGB space. But they're being done in HLS space.

For the HLS space, I like the picture on this page http://escience.anu.edu.au/lecture/c...SV_HLS.en.html The "double cone" describes what I'm talking about pretty clearly. At the midpoint, you can make any fully saturated color possible, or desaturate the color to get to the (center) white point. Decreasing or increasing luminance from this midpoint restricts how much you will be able to saturate the color, to the point that at max luminance you can only make white and at minimum luminance you can only make black.

From what I understand, a display's color control works in the YCbCr space and adjusts the matrix that turns YCbCr back into RGB. Which is yet another colorspace that is used for some adjustments.
post #319 of 714
Quote:
Originally Posted by turbe View Post

Is it correct to assume a future version of the GUI will support loading/preserving the static values in between (for those with their own tools)?

Yes, that is a correct assumption. I've got several things that will be added to the GUI in the near future. It is definitely taking on a life of it's own.

Just out of curiosity, there are several ways that this can work, what would be your preference?

1) Have the user manually select "Write LUT" and "Write CMS".
2) Have the software not do anything to the CMS/LUT if the user doesn't make any changes to the controls.
3) For the LUT, only recalculate the changed points. For example, if you make a change to 30% Red, the only values that get recalculated are Red values between 20% and 40%.

It would be nice to add more control points, but the screen real-estate and UI controls beyond 11 get pretty messy. One option would be to add some kind of a "parametric" control that lets you pick the control point and the width of the effect.

Yet another option would be to change the spacing of the 11 points, so that at the low end they are closer together, using something like 0 5 10 15 20 30 40 55 70 85 100 for the points.

What are you all seeing on your displays, and what kind of control would be useful? Or is the 11 evenly spaced points good enough?

thanks,
-Eric
post #320 of 714
I'm one who would like to have the ability to control various points, not only every 10% (though that's an improvement over what most of us have to begin with). I'd like to have the ability to "spot-touch" some points below 10 and maybe a couple of others at other places in the grayscale. I don't know how best to accomplish that in an easy way though. Would it be possible to add a third screen (i.e. Grayscale, CMS, and a fine-tune screen or something) that would allow you to to fine-tune your grayscale at several definable points? Some sort of parametric control in that situation would probably be very helpful. I'm not a programmer, but that seems really complex to me, but I don't know. It's just a thought.

On my tv, any points that are not dead-on appear very noisy, and even though I can get pretty linear 10-point grayscale, there's spots in between that still show noise--that's why I'd like to be able to fine-tune in between. The same is true in the low end. My lower 10% fluctuates from too blue to too red from point to point, so that's why it would be nice to have control over those.

Anyway, that's kind of my ideal situation. Hope explaining what my needs are is helpful--whether you can actually act on it or not!
post #321 of 714
A detailed review of the VEQ Pro.

http://www.hippotechsolutions.com/?p=1905
post #322 of 714
Quote:
Originally Posted by Thebes View Post

IMHO, it's a real problem if setting are not independant. I remember it was a same problem on the first CMS version on the JVC HD750 (saturation and hue setting had an effect on luminance level.

Eric had a pretty concise answer to the issue, below:

Quote:
Originally Posted by elhernes View Post

For the HLS space, I like the picture on this page http://escience.anu.edu.au/lecture/c...SV_HLS.en.html The "double cone" describes what I'm talking about pretty clearly. At the midpoint, you can make any fully saturated color possible, or desaturate the color to get to the (center) white point. Decreasing or increasing luminance from this midpoint restricts how much you will be able to saturate the color, to the point that at max luminance you can only make white and at minimum luminance you can only make black.

Let me try to expand on it a bit, though, for the detailed folks. There are two issues involved here. One is the change from the relevant color models (e.g., YCbCr, RGB, HSL) and the other is the dependency on integer math to do it. Let's hit the latter first...

Grassman's Law dictates that what we perceive of as light is an algebraic combination of red, green and blue (simplifying here). As a result, if we had infinitesimally small "steps" that we could take, we would adjust colors by adding or subtracting the opposing primaries to move a primary anywhere within the physical gamut of the display. However, since we are limited to integer math, we have to use somewhat large discrete steps when making adjustments. As a result, unless you happen to need a mix of primaries that works cleanly with integer math, you will have a slight deviation from your target luminance due simply to rounding issues alone. This never goes away with ANY CMS until you can get away from relying on integer math in the signal chain. You can throw more and more bits at the problem to minimize it, but you ultimately never get away from it. For the VEQ, I suspect that this is relatively minor, unless you are putting in an extreme correction.

Now let's tackle the first issue...

Take a look at Bruce Lindbloom's site, and in particular, his wire frame rendering of the sRGB gamut. Note that the maximum "swept" area is at the bottom of the construct, and it tapers ultimately to a single point (white) at the pinnacle.
http://brucelindbloom.com/index.html...ityGamuts.html


This is what it would look like if the color space were rendered by a display with infinite dynamic range. Since we don't yet have such a thing, you end up with a similar "conical" view at the bottom of the "pyramid". Look at the attached JPEG. It is a single primary that has 100% saturation. Without looking at the underlying properties, can you tell which primary it is visually? How it will measure depends almost entirely on the dynamic range (on/off contrast in this case) on the display (and the meter) and the actual gamut of the display is almost completely irrelevant (within limits). When I talk about "gamut decay", this is an extreme example. However, as Derek pointed out, above, you can't get around physics, so while it would be great to keep HSL completely independent of each other, the real world just doesn't let us.

Bill
LL
post #323 of 714
Quote:
Originally Posted by elhernes View Post
If it was a straight line, then it should be pretty well preserved. If you get a chance, I'd be curious to see if it did change, or if the differences you are seeing are from the CMS.
According to the VideoEQ GUI, the points are:
  0 10 20 30 40 50 60 70 80 90 100
Red 000 262 336 410 485 559 533 707 781 855 929
Green 000 302 376 450 524 598 671 745 818 892 969
Blue 000 296 366 436 507 577 647 718 788 858 932

When saving the LUT, the options were: Measured Gamma: 2.20, Target Gamma 2.20 (I assumed this would leave the data alone) and Video Levels checked.

I have downloaded the LUT this creates from the VideoEQ (no CMS adjustments applied) and this is the result:


Compared with the LUT I uploaded:


Note: downloading from the VideoEQ in TAB format seems to be broken - the LUT was 0000 at all points when I did this. I had to save it as CSV and convert it myself.

Quote:
Originally Posted by elhernes View Post
Yes, that is a correct assumption. I've got several things that will be added to the GUI in the near future. It is definitely taking on a life of it's own.

Just out of curiosity, there are several ways that this can work, what would be your preference?

1) Have the user manually select "Write LUT" and "Write CMS".
2) Have the software not do anything to the CMS/LUT if the user doesn't make any changes to the controls.
3) For the LUT, only recalculate the changed points. For example, if you make a change to 30% Red, the only values that get recalculated are Red values between 20% and 40%.
I assumed it would only write changes. That is, if I only made adjustments to the CMS tab, it would only write CMS changes to the LUT.

I think with the current GUI the best option would be to have the "Save to VideoEq" button on the Config/Info tab write the whole LUT.

On the RGB-LUT tab, you would have a "Write LUT" option that only writes to the greyscale section.

And on the CMS tab, you would have a "Write CMS" button that only writes the CMS to the currently selected LUT.



Regarding a GUI for adding/removing control points, I did have an idea for something like that a while back, that would be manageable with any number of points. (well, within reasonable limits)

I'll see if I can do a rough sketch for you and post it here later. I only have access to Paint.net at the moment though, so it might be pretty crude.

Quote:
Originally Posted by drrick View Post
I'm one who would like to have the ability to control various points, not only every 10% (though that's an improvement over what most of us have to begin with). I'd like to have the ability to "spot-touch" some points below 10 and maybe a couple of others at other places in the grayscale. I don't know how best to accomplish that in an easy way though. Would it be possible to add a third screen (i.e. Grayscale, CMS, and a fine-tune screen or something) that would allow you to to fine-tune your grayscale at several definable points? Some sort of parametric control in that situation would probably be very helpful. I'm not a programmer, but that seems really complex to me, but I don't know. It's just a thought.

On my tv, any points that are not dead-on appear very noisy, and even though I can get pretty linear 10-point grayscale, there's spots in between that still show noise--that's why I'd like to be able to fine-tune in between. The same is true in the low end. My lower 10% fluctuates from too blue to too red from point to point, so that's why it would be nice to have control over those.

Anyway, that's kind of my ideal situation. Hope explaining what my needs are is helpful--whether you can actually act on it or not!
I don't know if this will be of any use to you or not. Attached is the spreadsheet (OpenOffice format) I used for calculating a 33-point LUT. That is Video levels with 0-10 in 1% steps, 10-100 in 5% steps, and 109%. I have updated it to include the fix for posterisation I saw between the lowest measured point and 0.

To use this, you need to manually adjust the LUT at each point. So rather than using the GUI to set the RGB values, you are changing the text file the LUT is saved as and uploading that to the VideoEQ.

Once you have the points adjusted, paste the values into the red/green/blue sections, the spreadsheet fills in the gaps, and you can copy/paste columns G-L into a LUT that is stored using the TAB format.

 

LUT33.zip 197.6533203125k . file
post #324 of 714
No image comes up(just black) when trying to view the .jpg in post #322.
Thanks.
post #325 of 714
Quote:
Originally Posted by andrewfee View Post

Regarding a GUI for adding/removing control points, I did have an idea for something like that a while back, that would be manageable with any number of points. (well, within reasonable limits)

I'll see if I can do a rough sketch for you and post it here later. I only have access to Paint.net at the moment though, so it might be pretty crude.

This is the sort of thing I had in mind:


I have no idea how difficult something like this would be to implement though.

It would allow for as many or as few points as you want if you could type in what each point was as you added/removed them, and gives you a visualisation of the LUT.
post #326 of 714
Quote:
Originally Posted by andrewfee View Post

I don't know if this will be of any use to you or not. Attached is the spreadsheet (OpenOffice format) I used for calculating a 33-point LUT. That is Video levels with 0-10 in 1% steps, 10-100 in 5% steps, and 109%. I have updated it to include the fix for posterisation I saw between the lowest measured point and 0.

To use this, you need to manually adjust the LUT at each point. So rather than using the GUI to set the RGB values, you are changing the text file the LUT is saved as and uploading that to the VideoEQ.

Once you have the points adjusted, paste the values into the red/green/blue sections, the spreadsheet fills in the gaps, and you can copy/paste columns G-L into a LUT that is stored using the TAB format.

Thanks! I may give that a try sometime when I have the time to mess with it. I was thinking about trying to mess with doing something like that by hand--your spreadsheet will make things easier.
post #327 of 714
Quote:
Originally Posted by wfrick View Post

No image comes up(just black) when trying to view the .jpg in post #322.
Thanks.

I don't know whether you are just being funny or not, but the image is supposed to be a black screen. I think what Bear5k was pointing out was that hue, saturation, and lightness are related to one another in the real world--in other words, the image he showed had 100% saturation, but the lightness was too low for you to be able to discern what color it was...thus saturation and lightness are related. Hope that helps.
post #328 of 714
Quote:
Originally Posted by drrick View Post

I think what Bear5k was pointing out was that hue, saturation, and lightness are related to one another in the real world

In multiple, real, meaningful and unavoidable ways. Whether we come at it from the standpoint of the sheer physics of the displays, the limitations of the underlying math system or features/functions of the encoding system, you can never truly get full independence.

Some more examples for those playing the home game (nothing too fun since I mocked these up quickly -- no subliminal messages or anything, but I did think about it!). I assume most people are viewing these on a PC, but to keep the saturation "pure", I'm using 0 for the off primaries and beginning at 17 here (so technically 100% saturated and just above video black). On a PC, this should become more obvious, but on a calibrated display -- even a contrast monster -- I doubt you will be able to tell which primary is invoked in example 1 and probably example 2.
LL
LL
LL
post #329 of 714
Quote:
Originally Posted by elhernes View Post

Yes, that is a correct assumption. I've got several things that will be added to the GUI in the near future. It is definitely taking on a life of it's own.

Just out of curiosity, there are several ways that this can work, what would be your preference?

1) Have the user manually select "Write LUT" and "Write CMS".
2) Have the software not do anything to the CMS/LUT if the user doesn't make any changes to the controls.
3) For the LUT, only recalculate the changed points. For example, if you make a change to 30% Red, the only values that get recalculated are Red values between 20% and 40%.

It would be nice to add more control points, but the screen real-estate and UI controls beyond 11 get pretty messy. One option would be to add some kind of a "parametric" control that lets you pick the control point and the width of the effect.

Yet another option would be to change the spacing of the 11 points, so that at the low end they are closer together, using something like 0 5 10 15 20 30 40 55 70 85 100 for the points.

What are you all seeing on your displays, and what kind of control would be useful? Or is the 11 evenly spaced points good enough?

thanks,
-Eric

Hi Eric ,

Do you expect to be updating the on-line PDF documentation as this Product continues its development ? It would be nice to see some of the new features and capabilities reflected in the PDF .

Scott......................
post #330 of 714
Quote:
Originally Posted by elhernes View Post

I've got several things that will be added to the GUI in the near future. It is definitely taking on a life of it's own.

Just out of curiosity, there are several ways that this can work, what would be your preference?

What are you all seeing on your displays, and what kind of control would be useful? Or is the 11 evenly spaced points good enough?

thanks,
-Eric

I agree with all of andrewfee's recent posts (#323 and #325) regarding what should happen when working with the GUI, what buttons should be added, and how to implememt more than 11 points. I hope you can add these soon, especially the first two so that the program acts in a predictable way.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › The VideoEq -- A low cost external grayscale/gamma/cms tool