Bias Lighting - Page 4

That was accomplished a while back.

The monitor's black level and peak white should be adjusted for dark room conditions. Then the bias lighting can be adjusted for about 10% of peak white. Black level may need to be tweaked a bit after this due to the change in black level perception with the ambient illumination.

Bias lighting is recommended for providing 'Critical Viewing Conditions for Evaluation of Color Television Pictures'(SMPTE RP 166-1995). It is also specified for use in the more recent ITU-R BT.710-4 'Subjective Assessment Methods For Image Quality In High-Definition Television' document. Acknowledging this should lead to the logical conclusion that its use would be desirable during calibration and/or picture analysis.

Best regards and beautiful pictures,
G. Alan Brown, President
CinemaQuest, Inc.
A Lion AV Consultants affiliate

"Advancing the art and science of electronic imaging"

Thanks, Thanks and more Thanks! Your the Man!

What's a good way to do this?

AVSHD has a test pattern for backlight evaluation, but its a full field pattern and so I'm not sure how suitable it would be for plasma?

This, and many other details about viewing environment issues, are explained more fully on my company's web site. Here is an example: http://www.cinemaquestinc.com/pdf/T5AdjustmentTips.pdf .

I read you short Adjustment tips article. There are a couple of areas that bother me regarding the process.

First, as quoted from the article (above) - "Once your TV is adjusted for a darkened room, you can simply compare the amount of illumination on the wall by eye to that in the test pattern." This paper references the typical 5-10% gray pattern found on calibration discs. However, with a colorimeter available for measurement and comparison, how can this be used to more accurately set the brightness level?

I have an i2Display meter with translucent cover for measuring ambient light. Do I sit in my normal seating position - bias light off. Measure from this position the 100% white window coming from the TV and room. Then turn off the TV and compare or adjust the amount of light with bias light on from the same viewing position until a 5-10% luminance level is achieved. This method would seem to be more accurate at getting the level correct.

Second, assuming you have a perfect D65 colored light. Doesn't the wall color greatly affect the color that you would see? If I remember correctly the light you see is that which is reflected (not absorbed) by the surface. Therefore if the surface is gray, then no problem. But I have honey-mustard colored wall for the room and therefore the bias light to reflect off of. So for these conditions, the perfect light is no longer giving perfect color to a viewing position. It is now giving a yellow-brown color without blues and reds. How do you correct or compensate for these lighting and room conditions? Assume changing the wall color is not an option - like in most non-theater mult-use TV rooms.

My feeling is that you would need a complementary color filter based on the wall color. Any ideas?

Most lower cost instruments would have difficulty reading the reflected light on the wall with much precision. It's much cheaper, quicker, simpler, and sufficiently precise to use the referenced test patterns. If you own a Minolta LS-100/110 or similar spot luminance meter, that would be suitable.

Wall color is discussed elsewhere on my company's web site. Most consumers who don't have a neutral wall color have one that is nearly neutral. Many rental properties insist on white walls. In such cases, other lamps that could be used for a bias light are much farther away from D65 than the typical daylight fluorescent tubes. This would accentuate, or further distort, the coloration of the wall in many cases.

Using color filters on the bias light to compensate for wall color would be very challenging. My company has experimented with filtering techniques over the years. Typical polyester theatrical gel sheets are not available in very fine steps of color difference. Combining them reduces light output substantially. Then the surrounding wall is still colored during viewing under normal room lighting. Providing a neutral wall behind the display solves that problem. Neutral goes with any other conceivable color design scheme.

Would you consider an off-white wall color (our particular wall color was called Navajo White for example), one that has just a hint of gray in it, neutral enough?

White and gray are both neutral, unless some additional color is added. Technically, a truly neutral color (Munsell Neutral Value, Photo Gray Card, etc.) will reflect all parts of the color spectrum evenly.

So that should be sufficient, for now, for the use of a CFL light source with a temp rating of 5500k (CRI unknown), understanding that there is no way to verify the mfrs package claim?

I used a CT 1/8th Blue filter on my bias light to make a correction for my chalk-beige wall (reddish cream). It favourably trimmed the excessive red and brought it down inline with green, but the result is still blue deficient. Filters block spectrum, they don't add colour.

If I really wanted a neutral result, I'd need a custom made filter that had slightly more cyan to trim both the red and green to bring them down to meet the blue. The expense outweighed the benefits in my particular case.
As GeorgeAB has said, you wouldn't want to mix filters as the light output would be seriously capped.

any reason why you chose a 5500K-rated bulb instead of a 6500K-rated one?

What's a CT Blue Filter? I agree the filter only selectively subtracts light and never adds light. How did you arrive at that color filter selection? Did you use a colorimeter to determine the required filter or just eye-ball-it?

From post #53:
"You may encounter either compact fluorescent or LED screw in lamps on the market, claiming to be 6500K. My tests on a variety of these solutions has repeatedly found them to measure far from the claimed color temperature. Both technologies also have poor color rendering indices (CRI) as a rule. If you lack the means to verify the color performance, all you have to go on is the manufacturer claims. Look for 6500K and a minimum 90 CRI if possible. Recent white LED products I've seen claim a CRI of around 70. The compact fluorescent products have typically claimed a CRI of around 75 to 80.

Each viewer gets to decide how far from the ideal they are willing to settle for."

CT filters are used by photographers primarily when they want to change lighting conditions from one standard to another (say daylight to tungsen or vice versa).

I used my colorimeter to reference the tube (which was pretty damned close to 65D with a small push towards green) and measured the reflection of the tube off the wall. From there, I could calculate the difference.

I actually found a really handy website that could calculate the perfect filter for me, but I seem to have lost it now. The closest standard gel that didn't block too much light was the 1/8, so I got some and made a new hood for my light.



Without the filter, the red in the paint was very noticeable to my eye. I had run the bias light for over a year before I got my meter, but I hadn't really known how far out it was from ideal. With the filter, I'm certainly not aware that it is deficient in blue. The excessive red was far greater than the lack of blue. Its measuring around 6000K and I'm very happy with the result.

I couldn't find one locally at 15W that fit the screw-in base that I have. I'm always on the look-out for one though with a CRI rating close to, if not over, 90. I even tried an LED but at 7.5W (450 lumens) but it just didn't seem to give off enough light. Expensive, even on sale.

I've been reading through the posts and comments here in this thread since I will be buying a new TV soon, but also have used an indirect "bias" light behind my TV for a few years now. I didn't think it was all that important to make sure of the bulb color temperature though. For one, the wall colors many people have may not be ideal as others have stated. But also, what about daytime viewing of the TV with some widow light entering the room? I've noticed that in summer, the incoming light actually imparts a greenish tint to the walls from sunlight reflecting off of the lawn and trees. Or it may be cloudy and the light is more bluish. So how is a person to deal with these different colors of light?

It seems to me that as long as the light isn't too bright or flooding in and striking the screen it isn't that detrimental. Also, at night, it seems that as long as the light behind the TV , and wall color, isn't some terrible harsh color on the eyes it should work fairly well, although not ideal. But neither is day time light entering the room very ideal. My wife and family think I am very fussy about settings on the TV, the sound system for the TV, and even the indirect lighting. It would seem that choosing neutral colors and the proper bias light would be ideal, but not practical for most.

"Most" aren't that interested in image fidelity. You are in the display calibration section of the AV Science forum. Please review this article about viewing environment principles: http://www.cinemaquestinc.com/ive.htm . Bias lighting is a regular feature from video industry best practices. The technique is founded upon a century of human perceptual factors studies and color science. The technique is entirely "practical" for every single viewer who has reference image quality as their goal. Too many consumers have the mistakenly presumed expectation that all video is "plug and play." It is not, and never has been.

It sounds like you may have photography interests. I liked your custom bracket and light hood. I have thought about doing this with my CCFL light tube. The bracket system using the TV's wall mounting holes is clever. I assume the bracket extention upwards from the light is for a center speaker? Your wall seems quite white or neutral in the photo. Did the filter sheets go inside your clear tube? How did you arrive at the correct amount of light luminance? Did you use the colorimeter to measure or eye-ball it?

I have tried some preliminary i2D2 colorimeter measurements on my TV/bias light/wall system. I am measuring with the translucent diffusor on the colorimeter from about 4 feet in front of the TV even with the top edge. So that from a single position, I can alternately measure the TV bias light output test pattern and then the reflected wall light from the bias light.

I don't understand the process that you used for selecting the filter? Measure reflected wall color then if too much Green get a green filter? I'm stumped.

Correct.


That is because the filter is doing its job. Here is a reference showing the colour of the paint with sunlight.




Hope this shot better explains it.


The filter is attached to the top of the visor and then slides inside the hood. I can open or close the apeture of the light by adjusting the visor with a twist of the end-cap.


Eyeballed it.
I think that is the real benefit of using a fluorescent tube with a hood + a little bit of distance between the display and the wall. You get to control how the light spreads. The hotspot (spectral highlight) that the tube creates is far much hotter than the 10% target we are going for. The hood allows you to contain the hotspot within the area of the wall that the TV is obscuring, so from the viewing position, you are only seeing the drop-off, which gets progressively dimmer as it spreads out.
The filter did cut down some of the light's intensity, but it wasn't enough to drop below the minimum I required.


You have to measure the light directly, then measure wall being illuminated by that light. When you compare these two measurements, you'll see how much deviation the light is from D65 and how much the wall reflection is from it.
Once you know the difference between these two, you can calculate a filter that will correct it.

This is the direct measurement of my tube


6418 is fairly close to the target of 6500 that we are looking for.

This measurement is the reflection from the wall


Not great. So I have a colour temperture of 5375 that I want to make into 6500. The perfect filter would be one that cuts red and green by the right amount to bring them both down to meet the blue. But the perfect filter (which would be a cyan colour) doesn't exist commercially.
So the best compromise ended up being the CT 1/8 blue which I calculated would reduce the excessive red.

And this is the end result.



So its not 6500, but its an improvement over 5375, which was visibly too warm.
I could layer a green filter over the blue one (to create cyan), but I calculated that it would kill my luminance and the light output would no longer be effective. Not worth the payoff. I'm more than happy with it. I ran it for over a year without the filter and I could always tell it was too red. Now, its very neutral and I don't stare at it thinking to myself that it needs more blue.