|
As this forum is rarely used anymore, we've locked it. Feel free to browse and read. Questions? Please reach out to us directly. Cheers! |
|
 |
|
|
| Author |
Message |
Phil Smith
Joined: 08 Mar 2006 Posts: 7717
|
| Posted: Fri Apr 11, 2008 4:14 pm Post subject: Over My Head: How do I read this CIE colorimetry chart? |
|
|
Am I the only one that doesn't understand this type of chart? An embarrassing admission, but I've never understood it. Can someone please explain it at the retard level?
|
|
| Back to top |
|
 |
Nashou66
Joined: 12 Jan 2007 Posts: 16171 Location: West Seneca NY
|
|
| Back to top |
|
 |
Person99
Joined: 09 Mar 2006 Posts: 4899 Location: Flower Mound, TX
|
| Posted: Fri Apr 11, 2008 4:31 pm Post subject: |
|
|
It is pretty simple, the 6 colored round dots are the the correct color points for the primaries and secondaries (in other words, what the display is supposed to produce).
The 6 colored squares with pluses are the measured ones from that. A CRT will usually be slightly "inside the triangle", most digitals look so bad because they are "outside the triangle"--in other words, oversaturated.
The coordinates at the top just the x y coodinates of those color points.
the swoop in the middle is color temperature. You want it to be D65 or 6504K across all luminance levels. All displays "fall apart" under 20 IRE.
D65:
http://en.wikipedia.org/wiki/D65
_________________ Dave
A train station is where the train stops. A bus station is where the bus stops. On my desk, I have a work station....
|
|
| Back to top |
|
 |
kal Forum Administrator
Joined: 06 Mar 2006 Posts: 18114 Location: Ottawa, Canada
TV/Projector: JVC DLA-NZ7
|
| Posted: Fri Apr 11, 2008 4:36 pm Post subject: |
|
|
This is my chart so I'll explain.
I'll keep it extremely simple, mostly because I really don't know enough to explain it in any more detail!
The coloured blob in in the background is CIE chart. It's basically all the colours on a 2D chart of hue vs saturation. Fully saturated hues lie along the outside edge with desaturated colors toward the center of the chart.
Inside this mass of colours you see a triangle with diamonds with plus's (green/blue/red) at the tips. The tips are called "primaries" as they're the R/G/B primary colours. The colours in that triangle are the colours you hope that your projector can display. The triangle with the circles at the tips is what my projector is actually capable of displaying. The circles half way along the triangle sides are the secondaries that my projector can display.
You can't adjust primaries or secondaries on most displays. You simply take measurements displaying all red/green/blue screens and then combo's of 2 at a time with a colorimeter and colorimitry software tells you what your primary and secondary values are. What your display can do is what you're stuck with. Some very high end scalers ($4000 Lumagen Radiance) and the Colour Management System (CMS) in some advanced digital projectors let you adjust primaries (needed by digitals as they're usually way off into the neon-like colours).
So what are the dots in the middle?
When you display from black (0 IRE) to white (100 IRE) on your screen you want to always try and not have tints or colours at all. You want it to be 100% grey all the way from black to white. Perfectly grey is the point labelled D65 or 6500 Kelvin colour temperature. The black line with the numbers are other colour temps. Lower numbers like 300K look red, higher numbers like 9000 look blue.
The little white circles are how my projector measured in colour temp going from 0 IRE (black) to 100 IRE (white). They're all piled on top of each other around 6500K so it's hard to see but there are a bunch of circles there. You take a measurement at each 10 IRE interval with a colorimeter and see where it falls. The idea is to get all your dots as close to the 6500K point as possible. For me they all fall with in the little blue circle which is the target you want to achieveThe idea is to get all your dots in there. It's pretty impossible to get EXACTLY 6500K for everything as no display is perfect. When you achieve this you have perfect greys from black to white - known as perfect greyscale tracking.
My 10 IRE dot is the one up and to the right. It's off by around 50 (DeltaE) points. Oh well. Such is life. It's the best I could do. 10 IRE is pretty dark. You don't even notice it really.
So how do you measure and adjust these?
It's actually quite simple. You need:
1. You need test patterns like the Digital Video Essentials: HD basics, available in either Blu-ray, HD DVD, or Standard DVD (ranges from $17-19, hover over for exact price).
2. A colorimeter. I used the one that comes in the Spyder2 Express Kit (around $60, hover over for exact price). Works great and cheap to boot. Yes, there are "better" colorimeters out there but the Spyder2 reads consistently down to 10IRE and most of us will only do a greyscale calibration once every few years as we change equipment.
3. Software to take the readings and graph things for you. I used the free HCFR tool. Works great. Can't beat the price. Make a donation if you like it!
That's it!
The software's not really obvious to use if you don't have a good understanding of the concepts I mentioned here. I'll be putting together a step by step "How To" in the near future to save people a lot of head scratching, including tips on CRT projector greyscale calibration. Unlike digital projectors, CRTs are a bit tricker for setting greyscale as things do not behave quite as linearly as in the digital world.
EDIT: The GREYSCALE CALIBRATION FOR DUMMIES guide is now up for those that haven't seen it already!
Kal
_________________
Support our site by using our affiliate links. We thank you!
My basement/HT/bar/brewery build 2.0
Last edited by kal on Wed Apr 30, 2008 7:41 pm; edited 2 times in total
|
|
| Back to top |
|
 |
garyfritz
Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO
|
| Posted: Fri Apr 11, 2008 5:14 pm Post subject: |
|
|
A few more points:
The primaries are R, G, and B. This is what your 3 CRTs are supposed to emit. The phosphors don't have precisely the right color so they don't fall exactly on the "correct" point. You can't change the color that the phosphor glows -- BUT you can filter the light with a colored lens, glycol, or gel, so the resulting light is closer to the correct point. See, the phosphor doesn't emit one single color like a laser. It emits a range of colors, most of which are not the "right" color. A proper filter can filter out much of the "wrong" colors, leaving you with mostly the "right" color.
The secondaries are cyan, yellow, and magenta. You can't control these separately with a CRT, because they're formed by combining two primaries.
Why do you care about getting your primaries on the "right" color? Because your projector is like a blind painter with 3 pots of paint, red green & blue. He's painting a picture by using the same proportions of R/G/B that a camera measured. If his red paint exactly matches what the camera calls "red," then his colors will match the original scene. But imagine if somebody stole his red paint and replaced it with purple! He can't see the results, he's just using the paint pots that the camera tells him to use. If the camera sees pure red, he uses only the paint from his "red" pot, but his resulting picture is purple. So you want your projector's red "paint pot" to match the camera's definition of red. The gray triangle shows the "right" colors that the camera used, which are defined by the SMTPE C standard. The closer your projector's primaries are to the "right" SMPTE C colors, the closer the resulting projected colors will look to the original scene.
CIE coordinates: you might have seen people refer to colors in X/Y coordinates. Those coords are X/Y locations on this chart. You can't see the Y axis on the chart above, but you can see the X axis. Supposedly all visible colors are represented on the CIE chart, and all can be referred to by the appropriate X/Y coords. D6500 is a particular coordinate, .313/.328 -- that's where Kal's dots are.
Note that SMTPE C and all current display technologies CANNOT display all visible colors -- because many colors are outside the gray triangle. Reality is more vibrant than our displays can reproduce. Kal mentioned the benefits of the expanded HD color gamut yesterday; that refers to the different primaries that HD uses. This new color standard is defined in Rec709. (However: Kal, maybe you can clarify for me. How is playing an HD source going to give you any more color depth or vibrance -- because you still have the same projector with the same primaries! Either your projector was right for SD and can't display the additional HD colors, or it can display the HD colors but in reality it was displaying them all along even with SD sources, or something in between. Plugging in a new PS3 doesn't change what your projector is capable of displaying.)
Color temp: this literally refers to the color that an idealized black-body object will glow at a particular temperature. Think of a piece of iron being heated in a torch: it starts out dull red, gets brighter red, then yellow, then white-hot, and eventually a bit blue-ish. Those colors are represented by the temperature of the object when it glows that temperature -- its "color temperature." Now a hot object will ONLY follow that curved black line. You can never get something "green-hot," or magenta, or even pure yellow. Just the colors that fall under that curved line.
And by the way, if you read the "color temperature" of your projector when it's displaying pure white, your colorimeter might give you a "temperature" -- but the "temperature" only makes sense if the color falls exactly on that curved line. If the color is off the line, then that "color temperature" reading is just an approximation of the closest point on the color-temp line. It's better to use the X/Y coords so you show exactly what the color is, even if it's not a genuine "color temperature" coordinate.
"Setting the grayscale" on your projector is a process of tuning the projector so that all levels of pure white, from very dark (IRE10) to full-bright (IRE100) fall as closely as possible to the D6500 "white" point. If your low IREs have a "cooler" color temp (which, paradoxically, is redder), then your shadow areas will be have a reddish tint. If your midrange IREs have a "hotter" color temp (bluer -- remember, blue-white is hotter than white-hot), your midranges will have a blue-ish tint.
And by the way, you can get your grayscale 100% dead-on perfect, from 10 to 100IRE -- and still have bad colors if your primaries are wrong. Even if your green is more "yellow" than it should be, and your red is more "orange" than it should be (which is what you get with unfiltered CRTs), you can find a point that produces pure D6500 white. But other points away from pure white -- like, say, skin tones -- will look wrong if the primaries are wrong. That's why you filter to get your primaries as close as possible to the SMPTE C ideal.
|
|
| Back to top |
|
 |
kal Forum Administrator
Joined: 06 Mar 2006 Posts: 18114 Location: Ottawa, Canada
TV/Projector: JVC DLA-NZ7
|
| Posted: Fri Apr 11, 2008 5:54 pm Post subject: |
|
|
Great write-up Gary!
FYI - Gary wrote up an article that explained the benefits of colour filtering last summer that I just re-read. Completely with CIE graphs of before/after with different types of filters. A must-read for anyone who wants near correct colours in their HT. See: https://www.curtpalme.com/forum_archived/viewtopic.php@t=5604.html
| garyfritz wrote: | | Note that SMTPE C and all current display technologies CANNOT display all visible colors -- because many colors are outside the gray triangle. Reality is more vibrant than our displays can reproduce. Kal mentioned the benefits of the expanded HD color gamut yesterday; that refers to the different primaries that HD uses. This new color standard is defined in Rec709. (However: Kal, maybe you can clarify for me. How is playing an HD source going to give you any more color depth or vibrance -- because you still have the same projector with the same primaries! Either your projector was right for SD and can't display the additional HD colors, or it can display the HD colors but in reality it was displaying them all along even with SD sources, or something in between. Plugging in a new PS3 doesn't change what your projector is capable of displaying.) |
Good question.
If you super impose a Rec601 (SD) color gamut on top of a Rec709 (HD) one you'll see that the HD color gamut triangle is larger - it goes out farther which therefore means the the HD color gamut is capable of display more vibrant colours.
Now my colour gamut shown in the initial post above is larger than both the SD and HD colour gamut. Which means (correct me if I'm wrong here) but if I put in an SD or HD disk that asks me to display a 100% red signal, won't it land exactly in the same spot since the signal will be, say, 255/0/0 (R/G/B) no matter which disc we use? Ie: In the spot that defines my primary? So how can SD or HD look any different?
That being said, I swear I see a wider colour gamut on HD that I've never seen in SD before. Many people say this. We're definitely not getting something.
Kal
_________________
Support our site by using our affiliate links. We thank you!
My basement/HT/bar/brewery build 2.0
|
|
| Back to top |
|
 |
garyfritz
Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO
|
| Posted: Fri Apr 11, 2008 6:46 pm Post subject: |
|
|
| kal wrote: | | Now my colour gamut shown in the initial post above is larger than both the SD and HD colour gamut. Which means (correct me if I'm wrong here) but if I put in an SD or HD disk that asks me to display a 100% red signal, won't it land exactly in the same spot since the signal will be, say, 255/0/0 (R/G/B) no matter which disc we use? Ie: In the spot that defines my primary? So how can SD or HD look any different? |
That's exactly my understanding. I believe a SMPTE camera will produce a 100% red signal if it sees a red that lands on the SMPTE red spot, and an HD camera will produce a 100% red signal if it sees a red that lands on the Rec709 red spot. And your projector is going to display the same color (your projector's pure red CRT color) when you give it a 100% red signal, whether that 100% red is 100% of SMPTE or 100% of Rec709. You would have to have different primaries in your projector, or have SD disks/players produce less than 100% red at a SMTPE-standard red spot, for SD and HD to look different.
| Quote: | | That being said, I swear I see a wider colour gamut on HD that I've never seen in SD before. Many people say this. |
I don't see how, unless it's a placebo effect, possibly driven by the (accurate) perception of more spatial detail.
You've got a colorimeter, Kal. Get SD and HD test disks that include primary colors. Measure what your projector displays with an SD "red" and an HD "red." I'd bet a lot that they're identical, even if the SMPTE "red" and Rec709 "red" aren't.
|
|
| Back to top |
|
 |
kal Forum Administrator
Joined: 06 Mar 2006 Posts: 18114 Location: Ottawa, Canada
TV/Projector: JVC DLA-NZ7
|
| Posted: Sat Apr 12, 2008 2:47 am Post subject: |
|
|
| garyfritz wrote: | | You've got a colorimeter, Kal. Get SD and HD test disks that include primary colors. Measure what your projector displays with an SD "red" and an HD "red." I'd bet a lot that they're identical, even if the SMPTE "red" and Rec709 "red" aren't. |
Good idea! I do have both test discs and the Rec709 is already done!
Kal
_________________
Support our site by using our affiliate links. We thank you!
My basement/HT/bar/brewery build 2.0
|
|
| Back to top |
|
 |
perisoft
Joined: 29 Aug 2007 Posts: 2920 Location: Ithaca, NY
|
| Posted: Sat Apr 12, 2008 12:18 pm Post subject: |
|
|
| garyfritz wrote: | | A few more points:Because your projector is like a blind painter with 3 pots of paint, red green & blue. |
That painter is totally screwed, then... somebody needs to give the poor bastard some yellow instead of the green!
_________________
|
|
| Back to top |
|
 |
perisoft
Joined: 29 Aug 2007 Posts: 2920 Location: Ithaca, NY
|
| Posted: Sat Apr 12, 2008 12:20 pm Post subject: |
|
|
| garyfritz wrote: |
I don't see how, unless it's a placebo effect, possibly driven by the (accurate) perception of more spatial detail.
|
I'm inclined to agree. Look at it this way - take a screenshot of an SD DVD playing on an HTPC, and then an HD DVD. Slice 'em both in half and display the opposite halves - obviously 255,0,0 on one side is going to be the same as 255,0,0 on the other. Red isn't different any more than white is.
On the other hand, composite NTSC is a whole other ballgame - Somebody in the business came up with a great acronym for what it really stands for: Never Twice the Same Color!
_________________
|
|
| Back to top |
|
 |
garyfritz
Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO
|
| Posted: Sat Apr 12, 2008 4:37 pm Post subject: |
|
|
| perisoft wrote: | That painter is totally screwed, then... somebody needs to give the poor bastard some yellow instead of the green!  |
Hm. You have a point there!! I've used that analogy for ages and never thought of the difference between light and paint.... but hopefully it gets the idea across!!
|
|
| Back to top |
|
 |
Phil Smith
Joined: 08 Mar 2006 Posts: 7717
|
| Posted: Mon Apr 14, 2008 7:13 pm Post subject: |
|
|
Ok, I think get it now. The 3 points of the triangle are the true primary values. The Xs are the primaries of the PJ being tested. The temp line charts color temp values. The area outside the triangle are colors an RGB display can't reproduce, but the human eye can see (correct?).
| garyfritz wrote: | | And by the way, you can get your grayscale 100% dead-on perfect, from 10 to 100IRE -- and still have bad colors if your primaries are wrong. Even if your green is more "yellow" than it should be, and your red is more "orange" than it should be (which is what you get with unfiltered CRTs), you can find a point that produces pure D6500 white. But other points away from pure white -- like, say, skin tones -- will look wrong if the primaries are wrong. That's why you filter to get your primaries as close as possible to the SMPTE C ideal. |
I hadn't thought that before. You alway here mention of "orange reds" and "yellow greens" when talking about non-filtered PJs. Even though it's pretty obvious, I never thought about how incorrect primaries would effect *ALL* colors.
I didn't know how to read the *chart*, not not understand color and color temp. Most of what was written I already knew, but I did learn a few new things and enjoyed reading it.
Thanks!
|
|
| Back to top |
|
 |
garyfritz
Joined: 08 Apr 2006 Posts: 12088 Location: Fort Collins, CO
|
| Posted: Mon Apr 14, 2008 8:05 pm Post subject: |
|
|
|
Right on all counts, Phil. And yes, bad primaries affect ALL colors, since ALL colors (except 100% pure red, blue, green, yellow, cyan, and magenta) are mixtures of all 3 primaries. That's why it's so worthwhile to color-filter your R & G -- it affects EVERYthing.
|
|
| Back to top |
|
 |
Bruce 09
Joined: 08 Mar 2006 Posts: 747 Location: Kamloops BC, Canada
|
|
| Back to top |
|
 |
kal Forum Administrator
Joined: 06 Mar 2006 Posts: 18114 Location: Ottawa, Canada
TV/Projector: JVC DLA-NZ7
|
|
| Back to top |
|
 |
|
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum
|
Forum powered by phpBB © phpBB Group
|
|