[philaitman]

Hi,
I've read the article at GREYSCALE & COLOUR CALIBRATION FOR DUMMIES and given a new Samsung 37" B650 a greyscale calibration with a Colovision Spyder and ColoHCFR.

Now the Delta Errors are under 3 (Under 1 in some cases). Contrast set to about 40ftl and Brightness with a Pluge test from the AVS HD test disk which was used for the whole calibration.

The Gamma has came out at a ridiculous 1.18 which I would expect to produce a very washed out image which it Doesn't! This was with the TV's Gamma Control at 0 reducing this to -3 brought this up to 1.6.
Neither image appears washed out I do however see a green tinge at the very low end which isn't apparent in the numbers (This could just be perception to the change from the defaults!)

Could any of the more experienced experts on here give me some pointer as to where I may have gone wrong.
Nothing is hurt of course as I never needed to go past the user menu's thanks to the Samsung's excellent user colour controls.

Thanks in advance guys.

Forgot to add backlight is set to 3 on the Set as it usually viewed in a dimmed room with 1 lamp, calibration was performed under these conditions (Not sure if that matters to the numbers!)

[Avi]

Originally Posted by philaitman

Hi,
I've read the article at GREYSCALE & COLOUR CALIBRATION FOR DUMMIES and given a new Samsung 37" B650 a greyscale calibration with a Colovision Spyder and ColoHCFR.

Now the Delta Errors are under 3 (Under 1 in some cases). Contrast set to about 40ftl and Brightness with a Pluge test from the AVS HD test disk which was used for the whole calibration.

The Gamma has came out at a ridiculous 1.18 which I would expect to produce a very washed out image which it Doesn't! This was with the TV's Gamma Control at 0 reducing this to -3 brought this up to 1.6.
Neither image appears washed out I do however see a green tinge at the very low end which isn't apparent in the numbers (This could just be perception to the change from the defaults!)

Could any of the more experienced experts on here give me some pointer as to where I may have gone wrong.
Nothing is hurt of course as I never needed to go past the user menu's thanks to the Samsung's excellent user colour controls.

Thanks in advance guys.

Forgot to add backlight is set to 3 on the Set as it usually viewed in a dimmed room with 1 lamp, calibration was performed under these conditions (Not sure if that matters to the numbers!)

Can you confirm what boxes are ticked in the HCFR menu >advanced>preferences>advanced>Delta E section ?

AVI

[philaitman]

Originally Posted by Avi

Can you confirm what boxes are ticked in the HCFR menu >advanced>preferences>advanced>Delta E section ?

AVI

Hi Thanks for replying neither Boxes were/are ticked.

[KMO]

That's really, really wrong. You can see from the xyY numbers that the response is nearly linear throughout most of the range - every 10 IRE increment is giving you about 15 cd/m^2.

The Y difference between 90 and 100 IRE should be maybe triple the step between 30 and 40 IRE.

It's so wrong, that I find it hard to believe it's not a measurement error. But I m not sure how.

Maybe it's what you were measuring. Where were you getting your 10% IRE test bars from? What disc, and what player?

[philaitman]

Originally Posted by KMO

That's really, really wrong. You can see from the xyY numbers that the response is nearly linear throughout most of the range - every 10 IRE increment is giving you about 15 cd/m^2.

The Y difference between 90 and 100 IRE should be maybe triple the step between 30 and 40 IRE.

It's so wrong, that I find it hard to believe it's not a measurement error. But I m not sure how.

Maybe it's what you were measuring. Where were you getting your 10% IRE test bars from? What disc, and what player?

Thanks,
I'm using the AVSForum AVSHD709 disk played from a SonyBDP-S350 Via a STRDG820 Amp with a Colorvision Spyder colormeter.

This is the first time I've tried this so as they say you never learn without doing. I wonder if my Spyder is bust although it calibrates my screens for photography lovely!

I'll try a reset of everything again and have another attempt.

[philaitman]

Solved!
User error, I used windowed patterns and not full screen. The thick black borders on the Windows are effected by the Dynamic dimming on the set.
Just finished a greyscale pass and Gamma is bang on 2.2 and all Delta errors are less than 5.

The image looks stunning now

[Avi]

Originally Posted by philaitman

Hi Thanks for replying neither Boxes were/are ticked.

If you tick the box maked "Us gamma reference to compute grayscale delta E" you will see how changes you make to the RGB ratio at each point affects luminance as well as grayscale. Without this your grayscale may show low delta E but have luminance issues i.e. gamma. If you check the box you can compare the grayscale delta E with and without contribution.

AVI

[philaitman]

Originally Posted by Avi

If you tick the box maked "Us gamma reference to compute grayscale delta E" you will see how changes you make to the RGB ratio at each point affects luminance as well as grayscale. Without this your grayscale may show low delta E but have luminance issues i.e. gamma. If you check the box you can compare the grayscale delta E with and without contribution.

AVI

Thanks that's a useful tip

[KelvinS1965]

I did another greyscale run last night on my eye-one LT, HFCR, HD350 and Lumagen HDQ. I got the greyscale really flat from 20 to 100 IRE (using the parametric 11 point adjustments in the HDQ to fine tune some bumps). I also had delta E of 2 or less, before applying the setting Avi recommended above.

My issue is that the gamma is too low at an average of 2.0. I measured off the screen, in a fully darkened room with 'Bat tent' and various pieces of black material to reduce room reflections. The blue gamma in particular seems worse (see attached pic), so I'm after hints as to how to correct this. I should point out that I seem to be running out of blue at 100 IRE, so I reduced the main contrast to find the point where Blue was just starting to drop at 100 IRE...perhaps I need to lower it further?

I didn't use the HD350's custom gamma control, only had it set to 'default' for 2.3 which is usually recommended by other owners as measured gamma is usually 0.1 lower, so should be 2.2. If I changed the custom gamma to 2.4 or 2.5 this totally altered my greyscale (though it did increase the gamma a bit). I'm wondering if I should reset everything (again ) and try to get a gamma of 2.2 first, then adjust the greyscale? It may be a chicken and egg situation, but my initial delight at the RGB chart was soon changed when I looked at the gamma one.....

EDIT: Just to add I'm using windows rather than full fied test patterns, but my HD350 doesn't have any kind of dynamic iris/gamma control so I thought this is correct?

[Avi]

Originally Posted by KelvinS1965

I did another greyscale run last night on my eye-one LT, HFCR, HD350 and Lumagen HDQ. I got the greyscale really flat from 20 to 100 IRE (using the parametric 11 point adjustments in the HDQ to fine tune some bumps). I also had delta E of 2 or less, before applying the setting Avi recommended above.

My issue is that the gamma is too low at an average of 2.0. I measured off the screen, in a fully darkened room with 'Bat tent' and various pieces of black material to reduce room reflections. The blue gamma in particular seems worse (see attached pic), so I'm after hints as to how to correct this. I should point out that I seem to be running out of blue at 100 IRE, so I reduced the main contrast to find the point where Blue was just starting to drop at 100 IRE...perhaps I need to lower it further?

Do you calibrate the target grayscale and luminance (based on gamma target) at the the same time using the Lumagen ? Also can you post your measured luminance chart including breakdown as the variance between measured and target luminance at the various stim points is affecting the gamma result.

AVI

[KelvinS1965]

Hi Avi, no I trimed the greyscale first using the the HD350's high and low end RGBs, then I targeted a few peaks between 40 and 70 IRE with the HDQ. I did try adjusting the HDQ's 'Lux' (?) setting so that at each 10 IRE point the 'output' would be lower, but this only made a tiny difference to the gamma and I'm not sure this is the correct thing to do anyway? I set the 50 IRE to 48.0 for example.

I've attached the luminance chart (and my HFCR file if it helps). Just let me know if there is anything else you need. In general should I try to get the gamma right first using the HD350's controls, then use the HDQ if needed to trim any bumps in the greyscale?

[Avi]

Originally Posted by KelvinS1965

Hi Avi, no I trimed the greyscale first using the the HD350's high and low end RGBs, then I targeted a few peaks between 40 and 70 IRE with the HDQ. I did try adjusting the HDQ's 'Lux' (?) setting so that at each 10 IRE point the 'output' would be lower, but this only made a tiny difference to the gamma and I'm not sure this is the correct thing to do anyway? I set the 50 IRE to 48.0 for example.

I've attached the luminance chart (and my HFCR file if it helps). Just let me know if there is anything else you need. In general should I try to get the gamma right first using the HD350's controls, then use the HDQ if needed to trim any bumps in the greyscale?

Assuming your measure is accurate the luminance chart shows that you're not tracking target hence why your gamma isn't tracking target(Gamma determines luminance target) . So either your measurement is inaccurate of the luminance levels on the PJ preset you're using are wrong based on the grayscale RGB mix you've set. Gamma and grayscale interact hence setting one then making adjustments to the other often has an impact and luminance is governed by the relative mix of RGB.

It's a very iterative process to balance the RGB to achieve good grayscale whilst also achieving accurate luminance settings for your target gamma at each stim level. As your charts show it's possible to have good grayscale but this mix isn't producing the correct luminance at each stim level. Thus you need to remix RGB to hit the correct level at each stim point.

One option is to work back from 100% stim using the HDQ's RGB settings whilst also monitoring luminance at each stim level at the same time you're calibrating grayscale. What you should find is changing the mix of "RGB" values changes the HDQ's "luma" value and changing the "luma" value changes the "RGB" mix values. You can use the RGB mix to raise or lower the "luma" value for a given stim rather than adjusting the "luma" setting directly. Once you get luminance correct and grayscale is still good you've solved your gamma issue.

The other option is trying different settings on the PJ to establish if a more accurate grayscale/luminance balance can be achieved.

All this assumes your measuring device is accurate.

AVI

[KelvinS1965]

Originally Posted by Avi

All this assumes your measuring device is accurate.

AVI

Thanks for your comments Avi, I'm going to give it another try soon due to a hardware fault: Last night I found my HDMI 4 in 2 out box had died. I don't know if this had any effect on the readings, but it was working fine (as far as I can tell) two nights ago when I did the calibration. I don't know if I'm imagining it, but last night's viewing seemed to look really good after I bypassed the failed HDMI box.

I think I'll work on just using the HD350 controls initially (custom colour temp and custom gamma), then work down from 100 IRE with the HDQ like you suggested.

I realise I don't have the most accurate sensor, but it's better than my eyes alone I hope. At least my results seem repeatable with it now that I've learnt how to use it properly and with regular black level calibration between greyscale runs.

[Avi]

Originally Posted by KelvinS1965

I realise I don't have the most accurate sensor, but it's better than my eyes alone I hope. At least my results seem repeatable with it now that I've learnt how to use it properly and with regular black level calibration between greyscale runs.

Don't get me wrong it's just the sensor is one area that can obviously affect what you're measuring. Not suggesting your sensor has a problem or is inaccurate.

Originally Posted by KelvinS1965

I think I'll work on just using the HD350 controls initially (custom colour temp and custom gamma), then work down from 100 IRE with the HDQ like you suggested.

Sounds like a sensible plan. Get the display as close as possible using the internal controls and then fine tune with the HDQ. I find it easier to use the luminance type chart to set gamma.

AVI

[KelvinS1965]

Wasn't meaning to sound defensive...it's always in the back of my mind that the sensor may not be as accurate as I'd like. I figure that if I get the results as close to delta of 0 then I'm only out by whatever the sensor error is (hopefully less than a delta of 5 in the middle IREs which is better than by eye?).

What do you think of my failed HDMI switch/splitter? I tend to take the view that 'it's digital and either works of doesn't' and it was 'working' during the calibration.

Another thing I realised last night was that I was suffering reduced ANSI contrast due to having the lens dead centre. As I posted in the HD350/750 owners thread, I could see a 'ghost' image of the end credits, that I could 'tune out' by adding extra vertical shift. I now need to move my HD350 so that I can have some lens shift and still line up with the screen. I wonder if this might also effect the gamma as I was using windows not full fields?

[Avi]

Originally Posted by KelvinS1965

What do you think of my failed HDMI switch/splitter? I tend to take the view that 'it's digital and either works of doesn't' and it was 'working' during the calibration.

I suspect an issue here would cause far more visible problems that just specifically affecting luminance data.

Originally Posted by KelvinS1965

I wonder if this might also effect the gamma as I was using windows not full fields?

It's possible that changes to the lens zoom and shift setting can affect output characteristics especially at extremes of range. My guess is not achieving a gamma target is more likely to be a calibration or sensor related issue. That said it's worth investigating the ghosting especially if it's causing visible artefacts with real world material.

AVI

[KelvinS1965]

I've just updated the firmware (for better remote sensivity mostly) and moved the HD350 about 1 foot lower so I can use a bit of lens shift to check for 'ghosting' tonight. I'll have to wait for another chance to rerun the calibration due to work commitments (rather just watch it tonight ), but I'll be taking your advice on board when I get round to it and I'll post back my results.

Thanks again Avi.

[Avi]

Originally Posted by KelvinS1965

I've just updated the firmware (for better remote sensivity mostly) and moved the HD350 about 1 foot lower so I can use a bit of lens shift to check for 'ghosting' tonight. I'll have to wait for another chance to rerun the calibration due to work commitments (rather just watch it tonight ), but I'll be taking your advice on board when I get round to it and I'll post back my results.

Thanks again Avi.

I think calibration is a bit like a Rubik's Cube puzzle. It's easy complete one side at time in isolation but completing them all at the same time is a lot harder.

Happy viewing !

AVI

[KelvinS1965]

I deceided to treat myself to the forums power buy of Chromapure software. I've only had a brief play while I was stuck in a hotel away on bussiness this week, but so far I'm impressed. It looks like it will really make setting the gamma easier.

I used some 'test patterns' created using MS Paint, just to try out the software and I particularly like the gamma adjustment screen: You go down from an initial 100 IRE reading and can see your gamma in real time (continuous reading mode) at each point. Using the HD350's gamma control (and some info I dug up over on AVS forums on how to use it best, by adjusting one point down from the one you're measuring seems to be the main tip) I should be able to tune to 2.2 more easily.

I'll be interested to see what this brings to actually watching the image, but it'll be nice to know that it is adjusted to standards.

Of less use to HD350 owners is the colour gamut screen, at least it will help to optimise my HDQ settings and any future external CMS device I may buy.