Mark Hunter from Milori Colorfacts 14-18 Januari.

Started by J.A.F._Doorhof, January 3, 2003, 21:28:04

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

J.A.F._Doorhof

For this week our special guest Mark Hunter from Milori Color Facts.
They have introduced a very payable coloranalyzer solution on the market which has raised some eyebrows from the competition and from some calibrators, I love to have Mark over to answer some question I have and maybe some other people.

Keep it civilised and keep it English for Mark.

Mark Welcome,
And I will shoot first.

Gamma
One of the remarks I had on the Gamma issue was the following.
When you look at the diagram you can see the Gamma in a curve.

My remark was the following.
To have correct Gamma, you need a straight line from 10-100IRE.
So the difference between 10-20-30-40-50-60-70-80-90-100 IRE must be of the same ammount.

I understand that you have shown a curve for the desired Gamma response of a display, but why not the straight line ?

Measurements
According to Frank all measurements must be taken straight before the lens, why is that.
For example when I calibrate a projector on D6500 and move the analyzer (Sencore CP288) towards the lens I get a whole different reading, can be 500-700 degrees off.

Contrast ratio
Same thing here.
When measuring before the lens you will get a VERY high reading which is not what you see on the screen. As you will know the same projector on a 110" screen or on a 150" screen will yield different ratio's. Measuring before the screen + the gain seems to be the only correct way to REALLY know what you get.

Also on the contrast ratio, how did you make it work out.
When using for example the CP288 on a Marantz VP-12S2 it drops of in the blacklevel (below 0.050 ftlmb) when measuring the HT300+ it also drops of in the 0.50 ftlmb, still the Marantz had MUCH blacker blacks.

PLEASE don't think I don't like the product, I would love to play arround with it next to the Sencore, it would be a lot cheaper than the CP5000, but I also have some question marks, I calibrate for a living, people expect the best results from me and pay for that.

And to be honest, whatever analyzer you have, the calibrator does the work, and experienced calibrator can do more with an optical comparitor than a hobbist with the most expensive analyzer.

Hope you can shed some light.

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

FrankL

Frank, please correct:
Mark Hunter will answer our questions from Januari 14th to Januari 18th.
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

J.A.F._Doorhof

www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

Mark Hunter

Thank you, Frank!

It is my honor and pleasure to be here. Unfortunately, I will only have a chance to say hello and answer a few questions at the moment. I would be happy to participate more fully the week of Jan. 14th, but am preparing to leave for CES in Las Vegas and will be at the show all next week.

I also appreciate your flexibility in allowing this special guest thread to be in English. It is very much appreciated.

On to your first couple of questions...

Gamma

The relationship between input voltage and output luminance of a display device is non-linear, and it is this curve that is "drawn" by ColorFacts when analyzing that aspect of a display. I won't go into the CRT history or the non-linearities of the human eye, but if you are really interested, do an Internet search for "Charles Poynton" and "Gamma". Charles is widely regarded as the Gamma guru, and you will find out more than you want to know on that topic. :D

Gamma is most often expressed as an equation of the form:

y = x^g

Possibly with an offset (for bulb displays that have a non-zero black level). It is the "g" in that equation that we are solving for. A Gamma of 2.2 is the NTSC spec in North America, and when a display is calibrated to have a 2.2 output in a brightly lit room, it will appear realistic. However, a 2.2 gamma display in an otherwise dark room (known as "dark surround") will appear washed out. Unfortunately, the "dark surround" environment is exactly what most home theaters have. Our automated calibration systems all default to a gamma of 2.5, which is the recommend dark surround gamma, and this is also what is recommended in the ColorFacts documentation.

A linear output gamma (1.0) would look very washed out, with bright mid-tones and a lack of depth and shadow detail.

Just to be clear, there are several "parts" of gamma, and we are ONLY talking about the output ("system") gamma, which means the way that the light exits the display. This is not gamma source encoding, degamma or gamma correction within the display. We are measuring output gamma, which should resemble a 2.5 exponent function.



Measurements

QuoteAccording to Frank all measurements must be taken straight before the lens, why is that.

We recommend positioning the sensor in such a way as to maximize the accuracy of the data.

With single-lens projectors, the sensor can be placed close to the lens to see further down the gray scale. The colorimetric accuracy is not sensitive to position, and it doesn't matter if the sensor is close or further away (and the data does not change!), but the grayscale can be calibrated more accurately much lower when the sensor is receiving a lot of light.

With CRT projectors, the technique is different, of course, and the sensor is positioned further away from the projector as you do not want to introduce errors from the three guns into the measurements.

We will be introducing a new product at CES next week that takes readings directly off the SCREEN, so positioning will be quite simple, and will take all screen criteria into account.


Contrast Ratio

In theory (only), it shouldn't matter if the sensor is in front of the screen or in front of the lens. It shouldn't matter what gain the screen has or the size of the screen. If the instrumentation will perfect, with an infinite range, the contrast ratio would be correct under all of these conditions.

Of course, that is only in theory. ;D

In practice, all instrumentation will have it's valid measurement range, and the above statement is true only to the extent that both the numerator and denominator of the contrast ratio calculation are within that valid range.

Finally, I agree with you that the system is only the tool, and that the person supplies the talent to make it work.


Sorry for the brief stay, but I will be back on the 14th. Thanks again for the invitation. I am looking forward to returning shortly...

Mark

J.A.F._Doorhof

#4
Ok,

On gamma we have a different way of looking at it, but I think we maybe mean the same thing. Only in a different stage of the measurements.
we calibratie Gamma only with HT-PC's and offcourse when using CRT you need different settings than on Digital, so maybe we don't drift apart as much as I thought. ;D

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

FrankL

Hi Mark,

It's great to have you here! It's not the 14th yet, so I'll be patient ;D

I think the gamma-measurements are very helpful. Using my Philips 963SA DVD player, which has a custom gamma option, I was able to calibrate a gamma of 2.49. I'm very happy with the detail and punch I get with this setting. It's a lot better than just having to guess between the film/video/graphics setting of my Piano.

I do have some questions...

1. You have probably compared Colorfacts with the Colorvision Optimagery Sensor and the (much more expensive) products of your competitors. If you compare the resulting greyscale/temperature, would they differ much, say more than 100 IRE?

2. I believe Frank Doorhof uses the Sencore CP-288 sensor to calibrate.
On you specsheet (http://www.colorfacts.com/colorfacts/hardware/default.asp) you write:
Quotethe Sencore CP-288 will not yield accurate data from non CRT devices and can not be used to calibrate digital display devices.
Could you comment on this? What makes this product suitable for CRT, and not suitable for digital displays?
Would you advise FrankD to upgrade to another sensor?

3. Basic video calibrations.
Using AVIA or another calibration disc, I find it pretty easy to calibrate Black Level / Brightness with the black bars pattern.
However, setting the correct white level / contrast seems to be a problem, since my DLP projector does not bloom.
Is there a way, using Colorfacts measurements and a test pattern, to set the correct white level?

I also have a problem with calibrating color/chroma/saturation, using a blue filter. There is rather large range wherein my 'eye' doesn't see any difference in the test pattern.
Is there a way, using Colorfacts measurements and the test pattern, to set the correct value?

Thanks in advance,
Frank
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

J.A.F._Doorhof

Frank,

For contrast check clipping, you must see the two bars in the white area, than you are safe.

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

FrankL

I can see the 2 bars, so I am safe, but I would like to have the optimal setting  :)
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

J.A.F._Doorhof

#8
Drive it al the way up untill one bar disappears and than back it up one level.
Remember that contrast & brightness interact, so go between them untill you see four bars, two black and two white.

Use the Black/White pluge pattern for contrast and in a totally controlled room the 2 blackbars for the brightness.

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

Raf1

Mark,

I hope you surived CES?
And can cope wiht yet another round of questions about the little thingy you make?

Right,
It is very easy to go ahead and take meseaurements.
But, in order to realy get some where, is the setting of brightnes/contrast even G2 not part off the equesion?
Not the I am willing ore promoting any off this.
But Its my understanding, that having set these at the correct level, you have the base fore getting a good grayscale trekking.
If so, howe does Colorfacts can help in this task.

Second question:

I am very buzy with getting to the piont where optimal viewing conditions exicst.
Screen height, blacknes(I win), viewing angle, al are extremly close to ideal.
But, I hear that A figure between 11 and 13 footlambert is the thing needed.
I have follewed the disscussion's on AVS,and read the manual of the Colorfacts many times, and I d'ont get it?
Some clue's, but insight, no way.
So, Mark, how does(ore d'osend) the Colorfacts helps me?

Greetings,

Raf.

Mark Hunter

#10
Hello! It's good to be back from CES! Wow, what a show that was. Too much to see in only a few days.

Anyway, on to your questions!




Hi Frank L,

The Optimagery Sensor is a decent meter. Not reference quality, but not bad, either. Certainly excellent for it's price point. It's somewhere between the Philips Color Analyzer and a Spectroradiometer (it's a low-resolution spectroradiometer of sorts of its own).

The new ColorFacts CF-6000 spectroradiometer that we introduced at CES is an extremely accurate device. In fact, the CF-6000 won the accuracy shoot-out against the new Sencore and Progressive Labs offerings.

The Sencore CP-288 is a device with silicon photodetectors under colored gel filters. The colors of the gel filters are specifically matched to SMPTE-C, and will work great with devices that conform to that specification (read: CRT monitors).

However, fixed panel display devices do not match this spec (see SGHT or WideScreen Review magazines to see how far off they measure). For these devices, the CP-288 will produce "unknown" results.

For this reason, Sencore introduced the CP-5000, which they say is the "All Display Calibrator" and will work with any display device (not just CRT monitors). They brought this device to the accuracy shoot-out. Their new product is very expensive for what it is ($5000 USD), and our CF-6000 was shown to be more accurate (see here: http://www.colorfacts.com/colorfacts/news/20030114.asp), and less expensive ($1995 USD).



Hi Raf,

Yes, brightness/contrast are certainly part of the equation. To set brightness/contrast correctly, there are a couple of ways to do it.

Since I am most interested in making sure that shadow detail and highlight detail are fully resolved with no crushing, I use a technique where I do a grayscale analysis and then see how the luminance graph appears.



The goal is to make sure that the curve goes directly into the corner on the right-hand side, without going "flat" on top (crushing), and to make sure that it doesn't flatten out completely on the bottom end, either. Of course, brightness and contrast will help here.

Then, one moves on to grayscale tracking. Yes, there are other ways, and they work well. This is one technique that I use.

I will answer other questions as soon as I can.

Thanks!

SiR-ROUND

Quoteknown as "dark surround")

Nice name for a home theater...

;D

Mark, thanks for now... I don't have any questions but I enjoy reading this thread.


Manfred

FrankL

Hi Mark,

Do you have some input on my third question in the above post?

Quote from: FrankL on January  7, 2003, 19:39:45
3. Basic video calibrations.
Using AVIA or another calibration disc, I find it pretty easy to calibrate Black Level / Brightness with the black bars pattern.
However, setting the correct white level / contrast seems to be a problem, since my DLP projector does not bloom.
Is there a way, using Colorfacts measurements and a test pattern, to set the correct white level?

I also have a problem with calibrating color/chroma/saturation, using a blue filter. There is rather large range wherein my 'eye' doesn't see any difference in the test pattern.
Is there a way, using Colorfacts measurements and the test pattern, to set the correct value?

Thanks!
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

Mark Hunter

QuoteUsing AVIA or another calibration disc, I find it pretty easy to calibrate Black Level / Brightness with the black bars pattern.
However, setting the correct white level / contrast seems to be a problem, since my DLP projector does not bloom.
Is there a way, using Colorfacts measurements and a test pattern, to set the correct white level?

Hi Frank,

Here is how you would find the correct white level with ColorFacts that maximizes luminance without clipping:

  • Set all gains to the same value.  Start at maybe 50%, or wherever you are sure that the gain setting is not "clipping" or "blooming"
  • Take a reading and look at the RGB window in Colorfacts
  • The reading with the *lowest* % will be your limiting colour.  This is probably Blue with a CRT or Xenon-based digital projector, and probably Red with a UHP-based digital.
  • Go to Edit-Options and set Lock Primary Color pull down to that limiting colour
  • Click on "Measure" and select "Continuous Readings".  Now, increase the gains of all three primaries until you see the RGB values for the non-limiting colours start to climb quickly above 100%. This is the point where the limiting colour simply can not get any brighter.  You have found the maximum gain setting for the limiting color of the display.
  • Leaving the limiting color alone, set the other two colors to reach the target (probably selected to be D65).  You can just leave the RGB bars on the screen and adjust so that they read 100%, 100%, 100%.
  • You're done!  This is the brightest white point you can achieve without clipping/blooming any of the colors.
Note: Of course, you may not want to leave the controls at the absolute brightest non-clipping/non-blooming white point with a CRT projector, depending on how hard you want to drive the machine.

You can set color/chroma/saturation using the "Device Primaries" wizard in ColorFacts.  You will want to match up the primaries with the video standard you use (NTSC, PAL/SECAM).  Note that this can only be done with an instrument that is accurate not just for grayscale, but for the entire human color gamut.  For example, this can be done with our new CF-6000 spectroradiometer, but will not work (very well) with the CF-100.

Thanks!

Mark

FrankL

Hi Mark,

Thanks, great suggestion. Maybe you can put it in an advanced part of the manual  :)
I will try it out and compare the resulting setting with the 'simple' way using FrankD's suggestion.

About the saturation setting:
QuoteYou can set color/chroma/saturation using the "Device Primaries" wizard in ColorFacts.  <...>
For example, this can be done with our new CF-6000 spectroradiometer, but will not work (very well) with the CF-100.
I was thinking of a very straightforward way to do it, but I have no idea if it would result in measurements with any meaning using the Optimagery sensor. Maybe you can comment?

When you use the Avia test pattern and calibrate the saturation by eye, you look through the blue filter at a 'rectangle switching 2 colors' until through the filter the 2 colors are 'the same'.
I find this very hard to do, there is always a range in which it is very hard to decide which setting seems accurate.

What if I measure the 2 RGB values of the switching rectangle through the filter using Colorfacts, to decide the best setting. Would that make sense? And how should I interpret the RGB values?
Or could it be done without the filter, using some transformation comparing the RGB values?
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

Mark Hunter

Hi Frank,

I have to be honest with you: I've never tried combining filters with ColorFacts.  In that respect, some of our customers are actually much more "advanced" at using it than we are!

I know that some of the serious tweakers have used filters with ColorFacts to do things that we never really anticipated (two mentions of this are here and here).  We really need to catch up on these 'cutting-edge' ideas...after all, we might be able to help make it easier.

I like your idea of using the filter and substituting the sensor for your eye in the Avia test.  I have no idea how (or IF) that will work, but it is certainly worth a try.

I will be doing some testing this weekend, and will try to work this into it to give you a better answer.

Thanks!

Mark

RuudG

Hi Mark,

Great to see your input on this forum.

The gamma response of a CRT is caused by the fact that it is a phosphor using device. Now I have two questions.

1. What causes the gamma response in a digital projector?

2. I wonder why the gamma FrankL needs for his Piano is the same as for an CRT, since the causes must be different. Is this pure coincedence or is there a physical explanation?

I hope you can shed some light on this.

Thanks,

Ruud

FrankL

Quote from: Mark Hunter on January 17, 2003, 14:24:30
I know that some of the serious tweakers have used filters with ColorFacts to do things that we never really anticipated (two mentions of this are here and here).  We really need to catch up on these 'cutting-edge' ideas...after all, we might be able to help make it easier.
Hi Mark,

Actually, the first time I read about Colorfacts was in a similar thread, where someone was using Colorfacts and a FL-D filter to tweak his Piano:
http://www.avsforum.com/avs-vb/showthread.php?s=&threadid=165779
I've already bought the FL-D filter to experiment  ;D
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

J.A.F._Doorhof

IF an analyzer is working correctly (meaning not using cal files) the use of filtering before the lenses will NOT influence the working of the analyzer, only the outcome of the readings.

What you could do is het following.

Take the SMPTE standards for Red, Green and Blue.
I believe that these are:
SMPTE C RED  X-0.635 Y-0.340
SMPTE C Green X-0.305 Y-0.595
SMPTE C Blue X-0.155 Y-0.07

Now first measure the Green, Red, Blue fields without a filter and note the values. Than use a filter and do the same, experiment with filtering untill you are the closest to the SMPTE standard, on digital projectors this will "never" be accurate all the way. If you have to choose always choose AWAY from green, in other words you can better lean towards blue than towards Green, in lame man terms, choose a green value that is closest to the Standard (not too much).

the only thing you have to do is recalibrate after finding the correct filter. When using a analyzer with cal files I believe this method will not work completly to get an accurate grayscale after filtering, using a standard NO CAL file will help better I guess (no experience there).

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

FrankL

Hi FrankD,

I'm not sure what you're trying to say...
My plan with the FL-D filter is pretty straightforward: just run the complete greyscale calibration with Colorfacts, starting with the Device Primaries reading.
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

J.A.F._Doorhof

You could go further than just enhancing the picture.
With the above mentioned procedure you can SERIOUSLY tweak your picture into a whole different class.

My 808 has this type of colorfiltering and the result is amazing, the colors are accurate and very saturated, while standard the colors are plain dull and not so accurate.

When using filtering with digital projectors you can offcourse just buy a filter and put it for the lens, but you are better off by buying alot of filters and using the above mentioned procedure to try to get your colorgaumut better.

Grayscaling must be done after that.

I would not just put a filter before the lens without first checking if the coordinates are more accurate than the original.

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"

Mark Hunter

Frank D. and Frank F.,

Very good points!

Ruud,

As digital displays do not have the same inherrent response as CRT-based displays, the "gamma" response is artificially recreated by "look-up tables" in the electronics of these displays. These look-up tables allow a great degree of flexibility in styling the image, and the quality of the output of the device depends to a great deal on how well the values of these look-up tables are adjusted.

In the end, the light emitted from two calibrated displays should have the same characteristics, even though the techniques that arrive at that end may be different. Purely digital devices (DLP) have a direct relationship between the look-up table values and the output. Liquid Crystal displays have a non-linear correction to produce the correct output, and CRT displays coincidently natively possess almost the exact correct gamma transfer function, and do not normally have a look-up table of values.

Thanks!

Mark

FrankL

#22
Yesterday my new component cable arrived, so I spent part of the evening calibrating :)
Among other things, I tried out both methods of setting the contrast.

1. First I tried out FrankD's method.
QuoteDrive it al the way up untill one bar disappears and than back it up one level.
Remember that contrast & brightness interact, so go between them untill you see four bars, two black and two white.

Use the Black/White pluge pattern for contrast and in a totally controlled room the 2 blackbars for the brightness.
I'm not sure what the b/w pluge pattern is. I used the 'standard' contrast calibration pattern on Avia, which has 2 white bars in a white area.
However: even with the contrast on max, both bars stayed visible. So I was not able to use this method to set the optimal value of contrast.

2. Then I tried out Mark's method.
Quote from: Mark Hunter on January 17, 2003, 00:44:40
Here is how you would find the correct white level with ColorFacts that maximizes luminance without clipping:

  • Set all gains to the same value. Start at maybe 50%, or wherever you are sure that the gain setting is not "clipping" or "blooming"
  • Take a reading and look at the RGB window in Colorfacts
  • The reading with the *lowest* % will be your limiting colour. This is probably Blue with a CRT or Xenon-based digital projector, and probably Red with a UHP-based digital.
  • Go to Edit-Options and set Lock Primary Color pull down to that limiting colour
  • Click on "Measure" and select "Continuous Readings". Now, increase the gains of all three primaries until you see the RGB values for the non-limiting colours start to climb quickly above 100%. This is the point where the limiting colour simply can not get any brighter. You have found the maximum gain setting for the limiting color of the display.
  • Leaving the limiting color alone, set the other two colors to reach the target (probably selected to be D65). You can just leave the RGB bars on the screen and adjust so that they read 100%, 100%, 100%.
  • You're done! This is the brightest white point you can achieve without clipping/blooming any of the colors.
I did not completely understand 'Set all gains to the same value. Start at maybe 50%', so I interpreted that as starting out with setting contrast in the middle of the scale. I also changed the procedure a bit, since the 'lowest' color at my setting was green, which was not the limiting color.
So this is what I did:
1. I used the 100 IRE pattern on AVIA. I set the contrast in the middle of the scale.
2. I changed the custom color temp on my Piano, so that the RGB meter was at 100%/100%/100%.
3. I locked the Primary Color to red. (This keeps red at 100%, and green/blue is shown relative to red).
4. I increased contrast. I saw that increasing it to the max setting did indeed raise the green and blue level relative to red, but not by very much. I think blue ended up at 107% or so.
I chose the contrast setting just before a 'high' increase of green/blue, this was a setting where blue was at 102% and green at 101%.
5. I recalibrated brightness.
6. I repeated 4+5 twice.
7. Then I did a complete greyscale calibration with the new settings of contrast+brightness.

So was that a correct way to do it?
Marantz VP-15S1
Denon AVC-A1SR-A
Oppo 95

Rob_Dingen

Hello FrankL

I think you maybe can get some more lightoutput
What are your gain settings from red RGB and what are the maximum
As a quick test crank up the rgb gain settings and adjust contrast and see if the bars dissapear with the avia pattern you mentioned before
You also can check invidual colors on the avia pattern and see when they dissapear

Rob

J.A.F._Doorhof

Little explanation.

When you change the R,G,B values GAIN, you actualy give the projector more contrast. You have to change them all in the same ammounts however.

A recalibration should be done to double check after this.

Greetings,
Frank
www.hometheater.nl   /   ISF & HAA certified
Custom installer

Volledige ISF calibraties inclusief HDR en 4K.

"Omdat je je iets niet kan voorstellen betekent dat niet dat het niet kan gebeuren"