I was intrigued when I learned that Lumagen had ported the CMS technology originally developed for the Radiance, their flagship processor, over to the more affordable VisionHDP/HDQ processors. I have been clamoring for an ISF-in-a-box solution for some time now. The single biggest problem facing anyone who wants to improve a display through calibration is the often pitiful lack of control the display offers. You can't calibrate a device that has few or no calibration controls. In addition to the standard color, tint, brightness, contrast, and sharpness controls, most displays have some way of adjusting the grayscale. Typically this includes controls for adjusting RGB at the high end of the grayscale and the low end, though some displays offer only a single set of RGB controls that operate across the entire range. Much less common are color decoding controls, and even more rare is a properly functioning Color Management System (CMS) that allows a calibrator to bring the display's color performance into line with SMPTE-C or Rec. 709 specs, the standards by which all commercially available video material is mastered. This is an enormously important feature because manufacturers have trended in recent years towards increasingly oversaturated gamuts in an effort to win over undiscriminating consumers viewing displays on brightly-lit showroom floors. ISF fought and has now largely won the battle to encourage manufactures to stop setting displays at color temperatures that are grossly bluish, so accurate color reproduction is the next front on which this battle will continue.

What made the Vision so attractive is cost. The Radiance is a $4000 processor available only to a relatively few prosperous videophiles. The VisionHDP is about 1/4 the cost, bringing the ability to get a nearly perfect calibration within the reach of a much larger audience.

First, the HDP, which I tested at length, is not just for calibration. It is also a scaler, deinterlacer, switcher, and transcoder. It has a very large feature set the limits of which I have only begun to explore. The calibration tools were my main interest and these are fairly straighforward.

1. An 11-point grayscale/gamma adjustment tool.
2. A color decoding adjustment tool.
3. A color gamut adjustment tool.

Although the HDP has had the first 2 items for some time, it was the first time I had used them, so let me say a few words about each.

The 11-point grayscale/gamma tool works beautifully. It offers independent RGB and luma controls not for just the top and bottom end of the grayscale, but for each point throughout. Forgetting black, that is 40 separate adjustments! The result is a dead-on perfect grayscale and gamma (of your choosing). It works perfectly.

The color decoding tool provides what are essentially color and tint controls, but which operate on only red and green respectively. So, if you set the main color/tint control against blue you can theoretically get perfectly color decoding performance using the HDP tools. Color decoding is generally measured as proper lightness for the primaries and proper hue for the secondaries. This tool operates in the YPbBr color space. I was able to use it to adjust the lightness of red using the Red Color control, but via the DVI output, I didn't have any luck adjusting the hue of magenta using the Red Hue control. I would have tried it over component but I could not get the BNC connection to work with the component inputs on my Sony CRT. I have had this problem with other devices with BNC outputs. There is obviously some issue with my Sony's sync or timing requirements that I can't fathom. The HDP has a bewildering number of adjustments and I could probably resolve this given enough time, but the DVI output was fine.

One aspect of the color decoding controls that I did NOT like was that when activated it automatically puts the display into a red-only or green-only mode. This is for using gel filters and SMPTE color bars, but I much prefer to set the decoding controls using direct measurement, which I could not do in the 1-color-only mode. I wish that Lumagen would make this an option rather than a forced mode.

Next, I tested the CMS, the feature that I was most eager to try. As it happens my Sony CRT gamut is pretty good, except the green primary is measurably bluish. I was able to dial in green using the CMS very quickly. The test of this feature is not just in its ability to adjust the xy point of a primary color. It is critical that the CMS allows the user to do this WITHOUT messing up some other parameter of performance. When I had finished my adjustments, I was pleased to see that the gamma and grayscale were left unaffected, which is good. However, I also noticed that the yellow secondary was affected, which is not good. Fortunately, yellow was made worse by a much smaller margin that green was made better and with the color decoding improvements included I was able to substantially improve overall performance.

If you look at the numbers below you can see that I was able to achieve about a 45% reduction in color error. I measure color error using the CIE94 formula for dE. dE is a numeric representation of color error. The larger the dE number the further a color measures from a known standard, in this case SMPTE-C. 0 dE indicates perfect color, but errors of 1.0 or below are essentially invisible. I aim for a target dE of 2.0 or less. (The 1994 standard scales smaller than the more commonly-used 1976 standard, which results in much higher numbers.) As you can see, though yellow got much worse, the improvement in green was even larger. Red and cyan were also improved to a smaller amount, mostly because of better decoding.

http://home.comcast.net/%7Etlhuffman/lumagen/crt.gif

Next, I tried the HDP on my Sharp XV-Z20000 DLP projector. The 20K has a built-in CMS so it doesn't need the correction, but I wanted to use it as a test case because it offers a more typical example (in its uncorrected state) of the color performance of modern a digital display: oversaturated red, green, and to a lesser extent blue primaries. So, I turned off the Sharp's internal CMS and used the HDP's controls to adjust.

The results were disappointing. Although I was able to dial in the primaries to very near the SMPTE-C standard, this time the effect on the secondaries was more profound. The net result was slightly WORSE total color performance than when I started, as the numbers below indicate. The first set of numbers shows the color performance with the Sharp's internal CMS engaged. In the second set of numbers the CMS is turned off. In the third set of numbers the HDP's CMS is engaged.

http://home.comcast.net/%7Etlhuffman/lumagen/dlp.gif

The accompanying CIE chart tells much, though not all, of the story.

http://home.comcast.net/%7Etlhuffman/lumagen/cie_dlp.jpg

See how substantially under saturated the secondaries are with the HDP's CMS activated. Furthermore, what the chart does not show is that their lightness was also substantially raised, which greatly exacerbated the problem. I had seen a similar, though less profound, effect when using an early version of the Radiance's CMS on an InFocus SP7200 DLP projector.

Finally, I also thought that I would test the Lumagen's HD deinterlacing capabilities. I had tested its SD capabilities some time ago and was unimpressed with what I saw, at least with video-based material (film mode was much better). Using the Silicon Optics HD Benchmark Blu-ray disc, I set my Sony S1 Blu-ray player to output 1080i and connected it to the HDP outputting 1080P into the Sharp projector, which was set to dot-by-dot mode.

The results were quite good. Although it failed the video resolution test (the squares indicating the highest resolution were flashing and strobing), it passed all of the jaggies tests and did much better on the film resolution test, though there was some banding in the high resolution squares. It also properly applied 3/2 pulldown. These tests indicate that its inverse telecine performance with high-definition film-based material is better than that its motion adaptive performance with HD video-based material.

In addition to using artificial test patterns, I also tested the HDP with program material. I used the best example I know of a film-based deinterlacing torture test: the staircase scene at the beginning of Chapter 8 in Mission Impossible III. The HDP passed this test with flying colors with performance nearly as good as I saw with the Blu-ray player set to output 1080P directly off the disc. This scene will trip up most deinterlacers.

To sum up, as it currently stands I cannot recommend the HDP as an external CMS. Although it can properly adjust primary colors, doing so seems to undermine the performance of the secondaries as much or more. The more you adjust the primaries, the worse the effect on the secondaries. I understand that the Radiance CMS includes adjustments for secondaries as well. This is something the Vision series CMS needs to be of much value in this regard. On the other hand, if you only need to adjust a single primary, you may succeed with relatively few problems as I did with the CRT. Also, the color decoding controls are nice and the grayscale/gamma tool is world-class. Finally, the HDP offers quite good HD deinterlacing, especially with film-based sources.

Lumagen has a reputation for offering useful and timely firmware updates to its products. Perhaps they will enhance the Vision's CMS or, even better, offer a new calibration-only solution that provides ONLY a full-featured CMS, color decoding, and grayscale/gamma adjustments. A product like that at an even lower price with no scaling, deinterlacing, switching, or transcoding would be a revolutionary product. As it stands, the HDP is merely an evolutionary upgrade to an existing line.

Great write up Tom!
I was surprised Lumagen even added that feature to its older products, so I guess we can't complain.
I really want that Radiance, but its slightly out of my range. Here's hoping the RadianceXS gets here soon!

Tom: If you are not in the service mode On option of the scaler then you see all the colours when you adjust the colour decoder. This has always been the case. When you have service option ON it mutes the primaries not being corrected. It's a simple 1 second MENU 0910 command to toggle between states.


Gordon

Tom:

The problem which you encountered re: changing the primaries changes the secondaries is a function of the Viion series...let me explain a bit more:

-- As you are well aware the seondaries are a combination of the opposing primares thus changing a primary will impact the seondary...in the event that the primaries are spot on so *SHOULD* the secondaries but we all know that this is not the case.

-- As a result of harware limitations the Vision's implemenatation of CMS is a *light* version of the Radiance's...since you first saw the Radiance at my house things have evolved wuite a bit in that the Radiance now contains -- for CMS purposes -- a 7 point (R,G,B,C,M,Y,White) colour cube with each point being independent which allows one to precisely dial in xyY for each colour -- the one exception being undersaturated colors...the effctiveness is unbelievable as I know have -- EVEN ON MY COLOUR CHALLENGED PROJECTOR -- the luminance of all primaries and secondaries essentially spot on and the xy position of all primaries nd secondaries spot on (other than, of course, the two undersaturated primaries)

The point being that one needs to step up -- because of design / harware limitations -- to get the fully functional verions...I GUESS ANOTHER TRIP TO SEE BOTH THE RADIANCE AND THE NEW PROJECTOR ARE IN THE FUTURE!

I would also like to add that Lumagen should be commended for developing the Vision series to the limit of its physical capabilities...even offering a *LIGHT* CMS version put the Vision series ahead of many of its competitors...

Over to you...

Tom. I will have a new Radiance in a couple of weeks. I lent my own unit to a friend and he was so impressed he purchased it rather than shipping it back tome for a new one. When I get it, come on over and play with it to your heart's content and earn a few sheckles recalibrating my stuff --- CRT, SXRD and dila. Hell with the Radiance, CRT gray scale yada yada can all be done through the Radiance.

Re your comments re motion adaptive deinterlacing of video, it is much more difficult than deinterlacing of film. In motion adaptive a designer of the algs must make choices favoring certain artifact elimination over others. Depending on the test tract one uses, one machine will win and one will lose. use asnother track and the reverse results may occur. This is nothing new, you know this. At least with a CRT, I can watch HD video native and avoid deinterlacing artiacts except those contained in source commercials and the like.

Hi Tom. Thanks for your hard work on this. Some questions.

You used SMPTE-C color gaumt. Is this smaller than Rec. 709? If so, would the results be better (ie the secondaries not as off) if you used Rec. 709?

Dazzer had limited success with his attempts, but he did say he was able to strike a balance where the secondaries were on and the primaries were slightly oversaturated. Did you try this? If so, would it result in a smaller dE than what you get when the primaries are spot on?

Were you able to tame the secondaries via the HUE and HUE offset controls recommended by Jim Peterson?

I've ordered an HDP from AVS and Jason is going to calibrate it for me. I'm now wondering if I should just shell out and get the Radiance. The only issue I have is that this will be used almost exclusively for calibration - I will not be using it as a switching device and, aside from the rare SD DVD, will be feeding my pj with blu-ray and HD DVD at 1080p24. So, if the HDP can reign in the colors enough on my RS-1, it will be worth the $ savings vs the Radiance.

Thanks,
Noah

You used SMPTE-C color gaumt. Is this smaller than Rec. 709? If so, would the results be better (ie the secondaries not as off) if you used Rec. 709?The less correction you make, the smaller the effect on the secondaries, and since Rec. 709 is a wider color space calibrating to that standard would require less correction. I'll check tonight to see if it is enough less to make a meaningful difference. I doubt it.

Dazzer had limited success with his attempts, but he did say he was able to strike a balance where the secondaries were on and the primaries were slightly oversaturated. Did you try this? If so, would it result in a smaller dE than what you get when the primaries are spot on?Same answer. Less primary correction = less trouble with the secondaries..

Were you able to tame the secondaries via the HUE and HUE offset controls recommended by Jim Peterson?No. This is not how these controls work. They will adjust the lightness of the primaries and the hue of the secondaries. Unfortunately, the problem that the CMS causes for the secondaries is unrelated to hue. They become desaturated with increased lightness. The color decoding controls have no effect on this.

I've ordered an HDP from AVS and Jason is going to calibrate it for me. I'm now wondering if I should just shell out and get the Radiance. The only issue I have is that this will be used almost exclusively for calibration - I will not be using it as a switching device and, aside from the rare SD DVD, will be feeding my pj with blu-ray and HD DVD at 1080p24. So, if the HDP can reign in the colors enough on my RS-1, it will be worth the $ savings vs the Radiance.It is simply a question of cost. Are you willing to pay that much for accurate color? If you are, then it looks like it would be a good solution. I haven't tested the new version Lumagen recently implemented, but Joel Cohen swears by it and I trust his opinion. I would NOT recommend using the HDP for the RS1, whose primaries are considerably more oversaturated than the Sharp DLP. You are likely to make the color worse, not better.

Tom: If you are not in the service mode On option of the scaler then you see all the colours when you adjust the colour decoder.Thanks for the tip. This one of the many, many options on this device that I had yet to explore.

Thanks Tom. I wonder what the effect of the primaries would be if you added equal amounts of the colors to the primaries. For example, if blue was the least oversaturated, adding 3 "units" of Red and 3 "units" of green. Then for green, add 3 of red and 3 of blue, etc. I wonder if doing everything in equal increments would allow the secondaries to remain on the triangle formed by the incorrect but closer primaries. Can you recommend a decent colorimeter for use by a hobbyist that doesn't need pinpoint accuracy?

I would also like to add that Lumagen should be commended for developing the Vision series to the limit of its physical capabilities...even offering a *LIGHT* CMS version put the Vision series ahead of many of its competitors...Joel, I have a less sanguine view of this. For me it is not really about the effort or the credit, it's about the results. I'm afraid people will buy this hoping to make their color better, when in fact it can very easily make it worse. It is not so much a "LIGHT" version as a crippled one.

On the other hand, the device has many other virtues--the grayscale/gamma tool being at the top of my list--that make it easily worth the price. I just don't think the CMS provides any added value, except perhaps if used VERY sparingly.

Looking forward to your post-Jason visit post.

I wonder what the effect of the primaries would be if you added equal amounts of the colors to the primaries. For example, if blue was the least oversaturated, adding 3 "units" of Red and 3 "units" of green. Then for green, add 3 of red and 3 of blue, etc. I wonder if doing everything in equal increments would allow the secondaries to remain on the triangle formed by the incorrect but closer primaries. Can you recommend a decent colorimeter for use by a hobbyist that doesn't need pinpoint accuracy?Adding equal amounts of opposing colors will just desaturate the primary without affecting its hue. You will still have the problems with the secondaries.

The X-Rite Display 2 is a low-cost option that works reasonably well.

Although it can properly adjust primary colors, doing so seems to undermine the performance of the secondaries as much or more. The more you adjust the primaries, the worse the effect on the secondaries. I understand that the Radiance CMS includes adjustments for secondaries as well. This is something the Vision series CMS needs to be of much value in this regard.

This is very similar to my own observations. It would be great if Lumagen could somehow find a bit more room to fit the secondary adjustments. :o

D

Thanks for posting your findings Tom!
Well maybe there is something that Lumagen can do to compensate for the secondary issue.

Noah,

Just get the Radiance dude! :D

It's the same for me I'm attempting to correct the gamut on a JVC HD1. All the primary points are very close to their CIE SMPTE-C point ( DelatE < 2 ). The secondaries intersect in D 65 but saturation yellow and cyan are very inside triangle CIE after many hours i can't tweak secondaries , with a picture slightly reddish or magenta

Jean Louis

I'm seriously considering it...

On the Lumagen HD deinterlacing performance - there are multiple factors that can affect deinterlacing performance on the Lumagens. The biggest one is lack of 1:1 mapping from the source input into the Lumagen (NOT the 1:1 mapping from the Lumagen out to display - TPAT tests assumed perfect for that). The TOPL and BTMR input size controls require careful adjustment if set to "SCALE" in order to get a perfect match. If set to "PASS", then it is a non-issue.

I've also had the same results of the secondaries gaining extra luminance. I hope Lumagen can alter the method so it doesn't affect the secondaries in this way. Unfortunately the CMS isn't very useful in it's present state, it does more harm to the secondaries than the correction to the primaries. We can only hope it will change.

Forwarded this thread to the Lumagen guys...we'll see what they have to say.

"I would NOT recommend using the HDP for the RS1, whose primaries are considerably more oversaturated than the Sharp DLP. You are likely to make the color worse, not better."

"Adding equal amounts of opposing colors will just desaturate the primary without affecting its hue. You will still have the problems with the secondaries."

Tom,

Are your comments on how good the color is based on CIE triangles, sensor measurements, and or subjective viewing?

I'm only interested in the latter for my RS1.

If the HDP is used to visibly improve on the oversaturated green, are the secondaries visibly worsened the same amount, or would they not be as noticeable?

IOW, might the HDP still be a win for the RS1 despite being far from perfect by the numbers?

Thanks

Are your comments on how good the color is based on CIE triangles, sensor measurements, and or subjective viewing?

I'm only interested in the latter for my RS1.

If the HDP is used to visibly improve on the oversaturated green, are the secondaries visibly worsened the same amount, or would they not be as noticeable?

IOW, might the HDP still be a win for the RS1 despite being far from perfect by the numbers?Noah: It is not a question of just not being good enough. It is that you correct some colors by screwing up others by an equal or even greater amount. It's a zero sum gain.

Assuming one's color vision is normal, I guess don't appreciate the distinction you make here between perceived and measured performance. It's not like testing for deinterlacing errors against artificial test patterns where failures against a test pattern might not have obvious implications for real program material. Measured color performance is a very accurate predictor of what you'll see on the screen, if you are sensitive to color accuracy issues at all (some people evidently aren't). For example, you are bothered by the overly intense greens of the RS1, and green is the primary that measures furthest out from spec. Subjectively, I guess that some program material will look better after the HDP correction and some will look worse, depending upon the composition of the individual scene.

I used ΔE numbers to assess the color performance of the HDP's CMS. These numbers are based on psychovisual studies in which people are asked to rank subjective color difference.

I can't think of any reason for why trading RGB inaccuracies for YCM inaccuracies would offer any perceived advantage, but maybe someone else can.

I don't understand why the cms desaturates the secondaries more than the primaries. I would have thought that bringing in the primaries would fix the secondaries...

Tom:

Thanks for your feedback on the VisionHDP.

As Gordon noted the color/hue/offset work differently in user and service mode. In service mode we do blank the other colors, but in user mode the colors are all passed through.

---

As to CMS in the Vision: We added what we could fit and I believe it can be used effectively to improve projectors like the RS1. I agree it's not optimal.

Note that the Radiance uses a totally different CMS architecture. It has a 3D palette and has independent control of the primary colors, secondary colors, plus 11 user-defined points along the grayscale curve. The Radiance will also soon be doing CMS in a linear color space, making it even more accurate.

I think the problem you are having stems mainly from the fact that we just do not have the gates needed in the Vision series to do the correction in linear color space. So as you change the matrix adding low levels of the other colors it affects the secondaries more than it should.

In any case, I believe you can use the Vision CMS for the RS1, but it will require more time and effort than the RadianceXD would, since you will need to iterate the CMS controls and the Color/Hue/Offsets to get close.

Some thoughts on using the Vision CMS:

- Try for improved primary and secondary points. Trying to get them perfect may not be possible.
- Use the CMS to dial in the primaries, but only to the point that the secondaries are reasonable. Split the difference if necessary. You may even choose to try to get the secondaries accurate and leave the primaries outside the triangle a bit.
- Use COLOR/HUE/Offsets to tweak. These are in YCbCr space (verses RGB) and so have somewhat different effect than the CMS controls.
- NOTE: COLOR/HUE affect all three color channels. COLRRED, HUERED affect the red channel only, and COLRGRN, HUEGRN affect the green channel only.
- Repeat until you have achieved the best possible results.

Unfortunately there is tail chasing going on here as there are significant interactions between the CMS and Color and Hue controls. So, you may not be able to get it perfect.

It is very important to note that you can stop making changes once the secondaries are correct. The graph shows that if this had been done for the RS1, the secondaries would have gone from over-saturated to correct, and the primaries would have been improved.

So, I'm not saying you can dial in the RS1 perfectly with the Vision series. I'm saying you can improve the picture significantly(such as removing the "neon grass effect" people talk about).

In any case, we appreciate the the feedback.

Jim: any possibility that you will do a calibration box as Tom suggested above "a new calibration-only solution that provides ONLY a full-featured CMS, color decoding, and grayscale/gamma adjustments. A product like that at an even lower price with no scaling, deinterlacing, switching, or transcoding would be a revolutionary product."

Jim: any possibility that you will do a calibration box as Tom suggested above "a new calibration-only solution that provides ONLY a full-featured CMS, color decoding, and grayscale/gamma adjustments. A product like that at an even lower price with no scaling, deinterlacing, switching, or transcoding would be a revolutionary product."

We are considering a calibration only box. However, we have not made any decisions on this yet. Certainly it makes sense as it would be less expensive.

We are considering a calibration only box. However, we have not made any decisions on this yet. Certainly it makes sense as it would be less expensive.Gosh, I just can't imagine that such a product could fail to be an enormous success. Every ISF guy on the planet would want one and they, in turn, would recommend it to their customers. The hobbyist market would go for it big as well.

No such product currently exists. You would literally be creating a new market segment with exactly zero competitors.

Jim,

"I think the problem you are having stems mainly from the fact that we just do not have the gates needed in the Vision series to do the correction in linear color space."

What about a calibration only firmware for the Vision units that forgoes the scaling/deinterlacing to free up more gates?

Shawn

Jim,

"I think the problem you are having stems mainly from the fact that we just do not have the gates needed in the Vision series to do the correction in linear color space."

What about a calibration only firmware for the Vision units that forgoes the scaling/deinterlacing to free up more gates?

Shawn

That would be truly great.

Jim,
...
What about a calibration only firmware for the Vision units that forgoes the scaling/deinterlacing to free up more gates?

Shawn

Certainly a good thought.

I just don't see this happening -- The cost of development is too high for any anticipated return we might see from doing this.

I think it's more likely we would start with the Radiance design and trim it down to this - mostly because the software of the Radiance implements HDMI and already has the features we want. Still not sure if we will do this, but it is on our list of potential products.

Certainly a good thought.

I just don't see this happening -- The cost of development is too high for any anticipated return we might see from doing this.

I think it's more likely we would start with the Radiance design and trim it down to this - mostly because the software of the Radiance implements HDMI and already has the features we want. Still not sure if we will do this, but it is on our list of potential products.

Well, if you can offer this for 1K, you've certainly got me as a customer. And I like your plan much better, since it would be much better to have HDMI connections rather than DVI.

Certainly a good thought.

I just don't see this happening -- The cost of development is too high for any anticipated return we might see from doing this.

I think it's more likely we would start with the Radiance design and trim it down to this - mostly because the software of the Radiance implements HDMI and already has the features we want. Still not sure if we will do this, but it is on our list of potential products.

Count me in! I don't mind the colors on the RS1 but if I can dial them in perfectly then I would go for it. I still see some scenes that don't look right with the over sat colors.

There is so much more than color correction that a Radiance can do to improve PQ and simplify the operation of your HT. It is truly one sweet machine and it just keeps getting better and better. Plus Lumagen has great accessable people to do business with. Just a little nerdy but hell, even I can't be perfect.

There is so much more than color correction that a Radiance can do to improve PQ and simplify the operation of your HT. It is truly one sweet machine and it just keeps getting better and better. Plus Lumagen has great accessable people to do business with. Just a little nerdy but hell, even I can't be perfect.

Nerdy! Well than I don't want to have anything to do with them

Tom,

Did you measure the primaries and secondaries for various IRE levels (such as 100 IRE, 75 IRE, 50 IRE, etc)? If so, did the secondaries have the same error across those IRE? For example, did yellow at 100 IRE have the same dE as yellow at 75 IRE?

I think the problem you are having stems mainly from the fact that we just do not have the gates needed in the Vision series to do the correction in linear color space. So as you change the matrix adding low levels of the other colors it affects the secondaries more than it should.
That would make sense.

Let's look at an example. If you add a tiny amount of blue to the red primary so its signal becomes (R=1, G=0, B=0.03), and also add a tiny amount of blue to the green primary so its signal becomes (R=0, G=1, B=0.02), then the yellow secondary would be the sum of those primaries. However, there are two different ways to calculate that sum:

1. by staying in the non-linear signal domain, so that (R=1+0, G=0+1, B=0.03+0.02) ==> (R=1, G=1, B=0.05).

or

2. by converting to the linear light domain and back, so that (R=((1^2)+(0^2))^0.5, G=((0^2)+(1^2))^0.5, B=((0.03^2)+(0.02^2))^0.5) ==> (R=1, G=1, B=0.036)

So the 1st yellow is brighter and less saturated than the 2nd yellow.

Tom,

Did you measure the primaries and secondaries for various IRE levels (such as 100 IRE, 75 IRE, 50 IRE, etc)? If so, did the secondaries have the same error across those IRE? For example, did yellow at 100 IRE have the same dE as yellow at 75 IRE?No, I didn't, but I can take a look tonight to see if it makes a difference. I would be very surprised if it did.

BTW, the secondaries have specified points for SMPTE-C and Rec. 709. But what if your primaries do not conform to either spec as is often the case? Surely the secondaries (which are derived from the primaries) would then need to fall on different xy points. I know that you can approximate this by drawing a line from the primaries through the white point, but there must be some math that allows you to precisely calculate the secondary xy points based on the observed primary xy points.

I've searched all over for this, but cannot find it. Does anyone know?

I know that you can approximate this by drawing a line from the primaries through the white point, but there must be some math that allows you to precisely calculate the secondary xy points based on the observed primary xy points.
I suppose that you could calculate the intersection point of the two lines. For example, to calculate the cyan xy, you would calculate the intersection point of these two lines: the line containing green xy and blue xy, and the line containing red xy and white xy.

Here is a link to some mathematical formulas:
http://local.wasp.uwa.edu.au/~pbourke/geometry/lineline2d/

These webpages might help to do some calculations:
http://www.analyzemath.com/Calculators/TwoPointsCalculator.html
http://www.analyzemath.com/Calculators_2/intersection_lines.html

My ISF tech and I are having trouble with 1080i60 -> Lumagen -> 1080p24/60. The test patterns are not even close to being resolved with 1080i60 input. The vertical patterns are fine and stable but the horizontal patterns are just solid blinking blocks. Also, 1080p24 -> Lumagen -> 1080p24 is fully resolved, pixel perfect. So I know it's not my cables or A/V receiver. Here is the thread I started concerning this...

http://www.avsforum.com/avs-vb/showthread.php?t=1056577

I see that you are using a LumagenHDP and a JVC RS2 projector. I would like to get more information so that we can duplicate this.

1. What brand and model test generator or player are you using?
2. What test disk are you using?
3. What pattern are you using?
4. How was the performance with real world video?

One thing to note. Real world video has to comply with Nyquist. You can't have adjacent horizontal pixels that transition from black to white or from white to black. Video generators and test disks can produce test patterns that violate Nyquist. Such patterns are useful for testing but they might not accurately represent the performance of a product with real world video.

Best regards,
Randy Freeman

I'm not sure why picture details one pixel wide would violate Nyquist, perhaps someone might be able to clarify this. Isn't the whole point of so many pixels of resolution being able to reproduce alternating dark and light picture details of that size? Granted, uniform alternating dark and light picture details close to the pixel width each will yield moire patterns or not reproduce some of the time but some of the time in live video the individual details should be reproduced, or "come into focus" if you want to say it that way.

I would not be comfortable with a calibration only firmware. You would have to download it when you want to calibrate, and then re-download the regular firmware when you are done. Going back to calibration to touch up something you forgot would be a hassle. And each download runs the risk of leaving the processor dead in the water if a PC crash or power failure hits at the wrong instant.

As far as fine horizontal line live video de-interlacing test patterns (single alternating black and white) go, I would think that any motion adaptive processor should be able to pass this.

Hi,
I have a question about the Mode function, as in MENU>OUT>MODE. Can anyone describe what the eight independent output configurations are?
Thanks.

It is true that a modern fixed pixel display can replicate a black pixel and then a white pixel horizontally. But on a CRT display there is no way to replicate a back and then a white pixel horizontally. The specifications for HDTV television were set a number of years ago when the CRT display was the the dominate display technology. The signal bandwidth of a HDTV signal was limited for "real world" video to match the limitations of a CRT display. This is helpful because it also limits the signal bandwidth that is needed for HDTV broadcasting. The bottom line is that video generators and video test disks can output signals that are higher bandwidth than "real world" video.

Best regards,
Randy Freeman

The eight output configuration memories can be used to store setup data for one or more displays. Everything that is in the output configuration menu can be stored in different output configuration memories. Press "Menu, Output, Configs, [output config]". This includes settings for things such as different output modes (1080p24, 1080p60, etc), different gamma calibrations (day vs night) and calibrations for two different displays.

Best regards,
Randy Freeman

support@lumgen.com

Thank you for the reply Randy. Are the eight user configured or are they preconfigured? If they are preconfigured, what do 0-8 represent?

Vince

Thank you for the reply Randy. Are the eight user configured or are they preconfigured? If they are preconfigured, what do 0-8 represent?

Vince

Vince you can get these answers on the Lumagen website as well as how to set them. You can customize them.

Ron

Some of the Radiance Technical Tips that are posted on our website provide good explanations of the Output Configuration Memories and examples of how to use them. See Radiance Tech Tips 1,2 and 4.

http://www.lumagen.com/testindex.php?module=manuals

Best regards,
Randy Freeman

support@lumagen.com