[Jim Boden]

I finally got around to calibrating my 503CMX with ColorFacts and the results are outstanding. I first calibrated a 21†Viewsonic monitor to get some experience working with ColorFacts. The biggest difference is digital photos look much more realistic, especially colour rendition.


The equipment:


503CMX with about 350 hours on the meter.

Faroudja NRS connected to input 1 (RGB device).

McIntosh interlaced DVD player connected via S-Video to McIntosh pre/processor, with S-Video out to the NRS.

Laptop computer running Windows 2000 with ColorFacts installed.


The process:


Set the 503 and NRS to factory defaults, then set NRS noise reduction to 0 and detail to 4.

Set 503 to power control mode1.

All equipment warmed up for 30 minutes beforehand.

All 503 adjustments were made in Integrator Mode, none were made in normal menu mode.

Set brightness and contrast with Video Essentials.

Disconnected NRS from input 1 and connected laptop.

Plugged light meter into USB port, attached light meter to 503 with suction cups, booted laptop, and started ColorFacts.

Used all ColorFacts default exposure settings.

Performed Aim Meter function to verify everything was working.

Measured Device Primaries (ColorFacts sequences through all required test patterns and produces a CIE chart).

Measured Grey Scale (ColorFacts sequences through all required test patterns) using the default 10 step IRE measurements. ColorFacts will do 10, 20, 32, and 64 detail levels, but I didn’t try anything higher than 10.

Based on ColorFacts findings, adjusted 503 RGB settings and repeated measurements until it reported somewhere around 6480 degrees. It’s supposed to be 6500, but what the heck, close enough I say.

Disconnected laptop and reconnected NRS.


I watched parts of a few “reference†DVD’s, including Video Essentials, Gladiator, and the Fifth Element. An improvement in black level detail was the first most obvious difference. I now get much better detail in dark areas and colours are slightly toned down and more natural. I’m sure a qualified ISF person with a spectroradiometer would do a better job, but I’m quite pleased with the results.


So, what did I learn?


The 503’s contrast level is difficult to set up properly. I couldn’t find any patterns on VE which showed the effect of changing contrast levels. There’s supposed to be a white box in one of the Snell and Wilcox patterns which has a second box inside it, but I couldn’t see the inside box no matter what. The bending bar routine doesn’t work with plasmas, either. I set contrast to where I thought it looked “bestâ€.


The 503 displays the blacker than black pluge pattern and setting brightness with it was easy.


Calibration has absolutely no effect on colour impurities in the white mask screen. It still is slightly blue towards the centre and has a trace of red towards the top and bottom.


I had to turn down the brightness level much further than expected to get it right. Maybe this has something to do with the NRS being in the loop. Other people have shown brightness settings somewhere around 130 plus, whereas I ended up at 108.


ColorFacts is dead simple to use. I have no experience with this stuff and it produced good results. Am I going to put any ISF guys out of business? No! Besides, I can make better money keeping my “day jobâ€.


Final Integrator Mode settings are:


Contrast 127, brightness 108, V and H enhance 8 (default), R high 255, R low 124, G high 255, G low 129, B high 236, B low 128.


I don’t understand the difference between the RGB high and low settings and played with them both. If anyone reading this can educate me, I would appreciate it.


As an experiment, I tried these same settings on input 2 (component), and they don’t look good at all, so there’s a big difference between RGB and component.


Anyway, that’s it for now. Any comments, criticisms, suggestions, or questions are welcome. Please be gentle.

 

[cheridave]

Jim,


Are you saying that there is nothing in the "ColorFacts" program for measuring "White Level"(Contrast)? If there was you would put up the 100ire and measure that and then adjust accordingly.


But I'm sure that you will pass the "ColorFacts" around to all your Plasma friends here on the AVSForum- RIGHT?


Dave

 

[wrkn4thwknd]

Jim,


If you can get your hands on the Avia disk there is a great pattern on there to set contrast by (moving black bars in the white field). I just received my 503 a couple of days ago and the contrast was a breeze to calibrate with Avia. I found my display to be right on at the factory default, but you may still want to cut it back a bit if you are concerned about burn-in.


Scott

 

[cheridave]

Scott,


If the pattern that you are talking about is what I think it is- no, that is for checking the "Black Level" while you are making adjustments to the "White Level" on a CRT, etc. unit. The only thing one can do is turn down the "Contrast" to its lowest setting- (it will appear more Grey at this point) and then turn up the Contrast until the Grey looks White and stop there. But we don't have a hard measurement to go by, i.e. 23flts, 25flts, etc. It has been stated before and it will be stated again- these Test Disc don't provide a good test pattern for adjusting "White Level" on Plasma Displays.


Dave

 

[Jim Boden]

Thanks for all the contrast tips. I think it's very close as is at 127, which I believe is just under the default value.


I just watched Independence Day, which is a fabulous quality DVD. Everything is so much easier on the eyes and smoother than it was before. This includes scenes where there is harsh lighting. It no longer has the feeling that it's burning your eyeballs.

 

[Jim Boden]

cheridave:


ColorFacts allows you to measure contrast, but I don't understand its significance. I might take another look at it when I get a chance.

 

[hob]

Jim that was interesting we seem to be in the same neck of the woods, I liked mine so much I out I my signature. Second one down under the FCC URL scrool down to Master List. Jim Have a Happy Thanksgiving both you and your family.

 

[Ajlift]

Jim, thanks for your report.


What was your ambient light conditions when you set brightness? My setting using Avia in a dark/no light room is 134.


Were you able to flat line your "temperature histogram" below 30IRE? I was able to flat line between 30-100 but was unable to below 30. Every time I adjusted R/G/B levels to 100 at 20IRE, the levels above 30IRE were thrown way off.


Did you adjust hi/low colors randomly? I tried to adjust the hi color above 50IRE and the low below 50.


Did you adjust color or tint settings?


Alan

 

[cheridave]

Jim,


I am recommending that your baseline start with the White Level (Contrast) set at 25Ftls. Your "ColorFacts" program should tell you what the current measurement is with a 100ire test pattern. Then re-set your Black Level (Brightness) and go back to the White Level and re-check. Then go and re-check your Greyscale settings.


Dave

 

[DaveBoswell]

Jim,


Good to see you're getting great results with Colorfacts - I'm really happy with mine too.


The "correct" white level setting for any projector is actually not a set number persay but a suggested range. The white clipping point is the maximum it should be set at as mentioned above, but the suggested ideal range is likely significantly below that.


SMPTE recommendations are for a range of around 15 to 20 foot/lamberts base on a white (or 100 IRE) display showing. This (I assume) should be measured from quite close up to the face of the screen as the luminance will decrease as the sensor moves away from the screen per the inverse square law.


Colorfacts outputs in units of cd/m^2 (or candela per m squared), which can be converted to Ft.L by dividing by 3.4. Or alternatively you can just aim for a luminance reading of 51 to 68 cd/m2 while measuring the 100IRE pattern and you'll be there.


Note that this light level is actually quite low and really requires a completely dark room unless your screen is enormous. I only try for this level when I go into "full movie mode". In a daylit room it's way too dark for my taste.


This is using the 2.20 version of Colorfacts now available on the Milori site. I think the original version had unitless measurement using the Optimagery sensor. Mark Hunter, can you confirm that luminance is now NOT unitless?


To your other queries,


I believe the high and low RGB settings should be adjusted only when viewing light grey and dark grey test windows respectively. I alternate back and forth between the 30IRE and 80 or 90IRE patterns, adjusting only the low settings for the 30IRE window and the high settings for the 80/90IRE window. I usually don't go as low as 10 or 10 IRE as most projectors will tend to go significantly non linear. Once I get good results with these, I go back and check the other IRE levels and tweak accordingly.


I recommend not bypassing the scaler and DVD player by hooking up the laptop directly to the projector (I think this is what you are doing, right?). If you use the window patterns on the VE disk playing on the DVD player r through the NRS instead of using those generated by the Colorfacts software on your laptop, you will correct for any changes to the signal right from DVD source to Screen. It's a pain in the butt but might make a subtle but visible difference.


Hope this helps! Any suggestions from anyone else would be great. I think we are all having fun learning how to get the most out of the excellent Colorfacts product and our gear.


Dave

 

[Ajlift]

Dave:

Colorfacts gives a contrast reading as a ratio (ie 125:1) . What is Ftls and do you know what ratio would correspond to 25Ftls?

Alan

 

[ghibliss]

Adjusting the Contrast on a plasma display to 25 fL is not reccomended! You will be throwing out over 50% of the Contrast Ratio making the panel look dull and lifeless. The Contrast needs to be reduced typically on a 503 by about 7 clicks so that when the Brighntness level is properly adjusted the Grayscale tracking will perform in a very linear fashion. I can tell from the values Jim posted that the High Green value is close to 20 clicks too high as they all come adjusted at 255,255, 255 for the high value from the factory and 128,128,128 for the Low values. The High should be set against an 80 ire White window pattern and the low values with a 30 ire pattern.


The largest change will be to the High Green value and a slight Blue adjustment. The Low values will require very little change (only a few clicks).


Since you are using the NRS in the chain simply set the piece to F for flat rather than using a saved preset. This will provide the best method to calibrate the panel to the source that you are using without adding additional signal attenuation. You can achieve ruler flat calibration on a plasma panel (within 100 degrees from 20-100 ire at D65) quite easily.


You will need to apply the various signal types that you wish to calibrate to such as 480i, 480p, 1080i (1080i will be difficult without a generator), etc on the various inputs and perform the calibration routine for each of these signal types. A typical calibration performed for these three frequencies on both Component inputs (BNC and 15-pin) will take about 2 hours or so along with checking the results for each of these signal with video source material.

 

[DaveBoswell]

Alan,


Setting the Contrast and the Contrast Ratio are not the same thing, though I can see where why it's confusing.


Setting the Contrast, which is the main point of this thread, is actually setting the luminance of the "whitest" white possible on the display by adjusting the Contrast control on the monitor. Look at the "Luminance" dial in Colorfacts to see the effect of setting Contrast.


The Contrast Ratio on the other hand is in effect a measure of how far apart are the Contrast setting and the Brightness setting (or more correctly the White Level and Black Level - confusing isn't it?). It is the ratio of the luminance measurement of a white window (100IRE) pattern divided by the luminance of the black window pattern (0 or 7.5IRE).


For example:

White window -> luminance = 80 cd/m2 (by adjusting the "Contrast" control)

Black Window -> luminance = 0.5 cd/m2 (by adjusting the "Brightness" control)


therefore Contrast Ratio = 80 / 0.5 = 160:1


ghibliss:


I completely agree that setting the white level to SMPTE standards can make the picture dull and lifeless - that's what I think most of the time when I'm in a real movie theatre! Way too dark and no punch. The standard is based on the limitations of the brightness of pratical projector bulbs. I read somewhere that typical movie screens only have a 50:1 contrast ratio!


I say set the black level where it's supposed to be via Avia or VE and then set white level to whatever is most comfortable for you in your lighting environment, within the upper limits of your display. For me, I like watching movies in a completely dark environment, so I prefer keeping the contrast level at a lower than max setting, but somewhat higher than the SMPTE reccomendations.


Dave

 

[llogan]

Great post, this is something all of us Pio owners love to see. Will have to try your settings out to see, most interested in an improvement in the black levels.

 

[DaveBoswell]

Here's another great link to go with ghbliss' that I just came across which describes really well how all this works.

http://www.dk-audio.com/PDF_filer/colourtruth.pdf


Dave


PS

anyone know what the theoretical SMPTE-C xyY chromaticity coordinates are for the magenta, yellow, and cyan "secondaries" are? I'm playing around with the hue control seeing how it affects things and want to tweak these points.


http://www.dk-audio.com/PDF_filer/colourtruth.pdf

 

[Jim Boden]

Wow, thanks for all the input. Lots to think about and review. I'll get the version 2.20 update from Milori's web site and have another look.


I don't think I'm too far off, as the picture really is stunning. I've had an ISF calibration done before with an RPTV and the plasma PQ looks very similar, in terms of colour rendition and black level detail.


ghibliss:


I've seen some other posted calibration settings and agree that GH 255 doesn't seem right.


Ajlift:


I did the work in a completely dark room. I wasn't paying close attention to the histogram readings so guess I should go back and do that. I know the lines were reasonably flat when I took the last measurement, which showed a colour temp of about 6480.

 

[Gordon Fraser]

Can I also add that if you can't see the white box inthe whte box in the Zone plate on VE, even after adjusting contrast it may be becasue your scaler is clipping white detail. You might have to adjust contrast down on the NRS (or your dvd player if it as such an adjustment).


You can see effect of contrast in high ire image by putting a loop of the montage of images section up. Loop the bit where the fish jumps out of the weir. You'll see as you reduce and increase contrast the delineated line where the falling river meets the level river below will disappear and reappear. Try it and see. You obviously want to set contrast to a point where you can see the line where they hit each other.


Gordon

 

[Ajlift]

This is an exciting thread.


I am intersted in learning how to generate test patterns through the 503's various inputs. Since, according to GHIBLISS, calibration settings are not equal and need to be done separately.


I am using S-video connection for my cable box input for non-HD channels and component input for HD. How can I generate Avia test patterns on these inputs?

 

[Jim Boden]

Ajlift:


I have the same problem as you. I can only do it for the devices connected to my NRS on input 1. I could do it for inputs 3 and 4, I guess, but I never use them.


If you have a progressive DVD player, you should be able to calibrate input 2 at 480p and 480i, assuming the player will send 480i through its component out. For this, you can use Avia.


I have an HD satellite receiver on input 2 and no progressive DVD player. I understand from someone else who had an ISF calibration done that his settings are the same for both 480i and 1080i, which I found interesting. I thought they would have been slightly different.


Does anyone know if DirectTV ever broadcasts test patterns which could be used with a satellite receiver? I’ve never heard of it, but you never know unless you ask.

 

[Jim Boden]

DaveBoswell:


Thanks for all the great input.


I downloaded ColorFacts 2.20 and it has a problem, which I don’t understand. I sent an email to them about it but they’ve been very quiet for the past few days. Of course, I realize that today is a US holiday.


When I go through the same steps as with version 2.17 (i.e. measure primaries, measure grey scale), it only returns the illuminance history, the CIE chart, and the temperature histogram, showing only green. There is no RGB data to be found anywhere.


I went back to 2.17, and it’s working fine. Have you actually used 2.20?