This issue has come up on several threads on projector reviews, and I believe warrants a separate discussion.

The problem is that increasingly projector reviews are using the Rec. 709 color gamut as the standard to judge the color gamut of projectors. This information, for the time being is USELESS.

Kris Deering and Greg Rogers have posted here that currently all video material, including HD, is produced using the Rec. 601 color gamut; none is produced using Rec. 709.

The only MEANINGFUL color gamut comparison is Rec. 601 with that of the projector.

A case in point is three recent reviews of the Marantz VP-15S1:

1) Greg Rogers compared the correct Rec. 601 color gamut with the VP-15S1. All of the colors were very close, but slightly oversatured.

2) The UltimateAV review used the incorrect Rec. 709 color gamut (without even identifying that that was the one used), and came up with grossly oversatured red and green.

3) The Perfect Vision used the incorrect Rec. 709 color gamut ( and identified it as such) and also came up with grossly oversatured red and green.

Numbers 2) and 3) are utterly meaningless because there is currently no source material produced with Rec 709.

Only number 1) is useful because it uses the REC. 601 color gamut which is used for all video material today.

What conclusions can be drawn?

1) Rec. 601 is the ONLY color gamut that should be used in evaluating the color gamut of a projector.

2) Color gamut comparisons using Rec. 709 are useless because of the absence of any program material.

3) Reviewers should only use Rec. 601, and not Rec.
709 in evaluating color gamut

4) Reviewers should identify which color gamut they are using. To omit that info leaves the reader in a vacuum.

Anyone who knows of other useless Rec. 709 color gamut reviews is invited to post them.

I should mention that at some point everyone will switch over to Rec. 709, at which point comparing the Rec 709 color gamut with the proj. will be useful. But for now, it's Rec. 601 and only Rec. 601.

Robert I have never heard of REC 607...that said:

1. Is there any chance that by REC 607 you mean REC 601?

2. Is this such a significant issue because I (and assuming that you meant REC 601) I would suggest that the differences between REC 601 and REC 709 are smaller than other errs (i.e. color decoding within th video chain) given how relatively close the REC 601 and REC 709 color gamuts are.

Look forward to your thoughts.

Yes, I meant 601. I had 7 on my mind. Thanks.

If you look at how very divergent the color gamuts are on the Marantz VP-15S1 using Rec. 601 (Greg Rogers), and Rec. 709 (UltimateAV and TPV), the two gamuts can not be that close.

OK, I'll ask the stupid question....

I thought basically all HD (HD DVD, BD, ATSC, etc) was using REC 709 primaries? I know my IN76 switches to REC709 decoding when it gets an HD component input...

...Kris Deering and Greg Rogers have posted here that currently all video material, including HD, is produced using the Rec. 601 color gamut; none is produced using Rec. 709...I don't believe this correct. I know that gregr has pointed out most professional broadcast CRT monitors still use SMPTE-C (Rec 601) phosphors so a 601 mode may be closer to what the colorist/studio monitor operator saw on his/her monitor. The colorist may already have compensated for the monitor used. In addition, there are new LCD broadcast monitors which have a very accurate Rec 709 mode.

The bottom line is that hi def content is mastered to Rec 709!

Sure would be nice to have GregR himself chime in on this...

I thought that rec 709 had a wider gamut than did 607 anyway.

I am now totally confused.

Rec601 does not specify any primaries. So, technically, there is no such thing as a Rec601 color gamut.

The proper terminology is SMPTE "C" primaries, which are specified by SMPTE RP 145, SMPTE 170M, and SMPTE 240M.

I thought basically all HD (HD DVD, BD, ATSC, etc) was using REC 709 primaries? I know my IN76 switches to REC709 decoding when it gets an HD component input...

One VERY important aspect of this discussion: REC 601 vs REC 709 is about color decoding, which basically means the color conversion process from component to RGB. Color decoding does NOT have anything to do with color gamut, they are two completely different things. As Erik said, REC 601/709 isn't about color gamut, so whatever gamut the studio monitor had, you'd certainly still want to use REC 709 color decoding for HD material, and REC 601 color decoding on SD material. However, the color gamut used may be different in either case. I would believe the difference in color decoding between SD and HD has a lot more impact on the actual impression of the colors than the color gamut.

Basically, the ability of a display to either correctly decode component inputs (analog or digital) to RGB, or accurately displaying RGB inputs without messing with the color decoding, is extremely overlooked in reviews, especially considering the vast majority of displays available create errors in this regard, in one way or another. It's just not as easy to create a spec that everyone understands, as it is with color gamut. I believe a lot of reviewers are focusing so much on color gamut, that they fail to realize that color decoding is at least as important, if not even more important. A lot of reviewers highly praise certain displays which has very accurate color primaries for having "spot-on colors", even though color decoding is way off. I belive this is a major issue, which we should be a lot more concerned of than the differences in gamut.

Basically, your IN76 may very well shift between REC 601 and REC 709 _decoding_ based on input resolution, without changing color _gamut_ at all. I believe displays with the capability to actually change color gamut along with color decoding are very few. I believe the Samsung's do this, and it would seem that the new Sim2's is capable of this as well.

...may very well shift between REC 601 and REC 709 _decoding_ based on input resolution, without changing color _gamut_ at all. I believe displays with the capability to actually change color gamut along with color decoding are very few.

I believe this is the case also. The gamut of displays is usually fixed. YCbCr 480i/p sources should be REC601 and 720P or higher should be REC709. At least that is the assumed formats they should be in. In reality there are all kinds of situations where that is not the case, often due to equipment ignoring the input resolution and always applying a single matrix decode instead of two different ones when appropriate. Throw in upscaling players, and the situation gets worse. When a player scales SD sources to 720p or higher, it should also change the color matrix from 601 to 709 because 720p or higher is assumed to be 709. I know the Toshiba HD players did not do this on earlier firmwares (not sure of the current status). I suspect that part of the reason that HD cable/satellite feeds are so screwed up is because they ignore this distinction both in the cable box and at the cable head end. The fact that many displays, still after all these years, have broken color decoders turns this into a huge mess.

Kris Deering and Greg Rogers have posted here that currently all video material, including HD, is produced using the Rec. 601 color gamut; none is produced using Rec. 709.That's not quite what Greg wrote in his Sony VLP-VW60 review:

"However, professional high-definition CRT monitors still have SMPTE-C phosphors. As a consequence, the Normal mode will often produce a more accurate representation of what the telecine colorist saw while doing standard-definition and high-definition film-to-video transfers, and what is seen on HD studio monitors during broadcast video production."

The only MEANINGFUL color gamut comparison is Rec. 601 with that of the projector.

No where does either of the reviewers you cite say this.

A case in point is three recent reviews of the Marantz VP-15S1:

1) Greg Rogers compared the correct Rec. 601 color gamut with the VP-15S1. All of the colors were very close, but slightly oversatured.This is misleading. He compared the Marantz's color performance relative to BOTH the Rec. 709 and the Rec. 601 standards.

Tom-

You're right. My mistake.

I was talking to Dan Miller and he said that there are indeed very accurate REC709 LCD monitors available that are extremely expensive, which is why old SMPTE-C monitors are used. He says that when the prices of these monitors come down (a matter of years), everyone will be producing HD with REC709.

Oversaturation with respect to REC709 still tells you something - it will be way oversaturated with respect to SMPTE-C. You can pad the review yourself knowing that SMPTE-C is contained within REC709 gamut. Only if the display was undersaturated would having the hard data for SMPTE-C be needed. Nothing like telling us a display matches the DigitalCinema gamut as if that was a good thing - when it tells you it is oversaturated for REC709 and even worse for SMPTE-C. Better to tell us the gamut(s) the display can display - and note its comparison to the closest standard(s). It tells you what the design target was - no need to remeasure it outside the design target - when those standard offsets are already known. Sometimes a review can have too much information and measures - should we also start measuring Digital Cinema gamut compliance on the theory some displays can do it - even though no source material was mastered in it? Should we be leaving out the compliance to EBU - after all this is important to anyone in the EU! And your PhotoShop guru wants to know AdobeRGB compliance - they are not being served with reviews either!

Tom-

You're right. My mistake.

I was talking to Dan Miller and he said that there are indeed very accurate REC709 LCD monitors available that are extremely expensive, which is why old SMPTE-C monitors are used. He says that when the prices of these monitors come down (a matter of years), everyone will be producing HD with REC709.


Most color correction suites use a 3d correctional lut that will correct for something as straight forward as the differences between smpte and rec709 ( or pretty much any other video standard: EBU is actually closer to rec.709).As long as the display is up to describing the colorspace. You don't need hugely expenive LCDs either: the Apple cinema displays can be calibrated to rec.709 no problem : I've even managed to get a dell 2405 accurate at least in color terms. My preference is for a Sony FW900 which I use for PAL/EBU , rec.709 and film color correction.


I'd be very surprised to find anyone making color correction decisions for the 1080p mastering of a multi-million dollar film without sorting out something as trivial as their colorspace pipeline.

I can speak for the 20 or so prime time shows we do* for both the big 4 and other networks in HD and they are 709. In fact any downconvertor I have seen expects 709 input color space to produce compliant 601 SD. And since 601 still becomes NTSC, we must make sure any HD converts to legally compliant NTSC - those rules are still on the books until at least 2/17/2009 and the networks still enforce them.

As far as color correction, you can easily push beyond 709 without any problems but when downconverted you will have clipping and illegal colors. Therefore the TV industry generally stays within 709.

Now in feature film dailies, 3D LUTs are often used extensivily for print emulation so the dailies truly look like film. However it is rare a TV show will go to that expense.

Also until a month ago there was no reference LCD color monitor available. Yes there are a few claiming it but the industry has not accepted them. Last month Sony released it's new reference quality LCD display. We shall see. The reason we hang onto to CRT has nothing to do with cost. It's the fact that no LCD monitor is of a reference standard yet. A good quality image does not mean a standard. We have to know how the image will look at other facilities who colaberate in the post production process. To ensure that the industry uses calibrated monitors to SMPTE standards. No LCD monitor to date has been able to do that.

*This is before the %^*% writers strike!

Be careful not to mix up color encoding standards with the color gamut of monitors used during video production, telecine, compression, and other mastering processes. These are very different issues.

For standard-definition video (includes 480i/p) the color encoding standard is ITU Rec. 601, for high-definition video the color encoding standard is ITU Rec. 709. These standards govern the way that RGB video is converted to YCbCr video (or visa-versa), which is ultimately compressed (MPEG, VC-1, etc.) and distributed to us via DVDs, HD-DVDs, Blu-ray, ATSC over-the-air broadcast, NTSC over-the-air broadcast, satellite, etc. etc. The Rec. 601 and Rec. 709 encoding matrix standards also govern the way signals must be converted between SD and HD during production or in our homes in upconverters, scalers, upconverting DVD players, etc.

There is NO technical reason to ever get the color encoding wrong in production or conversion processes. When errors occur, and they have occasionally in production, and unfortunately way too often in consumer products, it is because someone goofed (fill in your own stronger words here!). There is no excuse for those errors other than humans make mistakes.

A completely different issue is the color gamut (chromaticity of the primary colors) of the monitors/displays used for production, telecine, compression, and for our home theaters. The "standard" for standard-definition program material is SMPTE-C, which has been in use since the 1970's. (Prior to that there was chaos ... that's a long story arising from the poor choice - another long story - of the original NTSC primary colors.) Rec. 601 does not address the issue of display color gamut, but the SMPTE-C primaries were "adopted" in the 1970's and became the defacto "standard" associated with Rec. 601 video (at least in the US). So to be precise the color gamut for standard-definition video is defined by SMPTE-C and not Rec. 601. Hence, technically we should refer to SMPTE-C primaries and color gamut, and not Rec. 601 primaries or color gamut. The color gamut for high-definition video is standardized in ITU Rec. 709, and thus can be referred to as the Rec. 709 primaries or Rec. 709 color gamut.

So ---- there is no issue (other than just human mistakes) associated with the Rec. 601 and Rec. 709 color encoding matrix equations. The issue is that professional monitors, which were about 99.999% CRT monitors until recently, do NOT have Rec. 709 primaries, they have SMPTE-C primaries. Therefore, when the telecine colorist is transferring a film to video, and adjusting colors so they look correct to him, he is looking at a display with SMPTE-C primaries, which produces decidedly different colors than a projector with Rec. 709 primaries. So if you want to see what the telecine colorist saw when transferring the film to video using a SMPTE-C monitor, you will need a display with SMPTE-C primaries.

(I've read some "experts" write that there is very little difference between the image colors with SMPTE-C primaries vs Rec. 709 primaries. That is hogwash! The difference is very significant. It is much more significant than the color differences between Rec. 709 primaries and the more saturated primaries used by many projector manufacturers.)

There are some electronic partial solutions to this issue, but for various reasons (that I won't go into) most of the industry appears to be waiting for new monitor technologies that will have Rec. 709 primaries.

It is much more significant than the color differences between Rec. 709 primaries and the more saturated primaries used by many projector manufacturers.)This sounds a little hyperbolic. The average RGB dE76 between SMPTE-C and Rec. 709 is under 8. That's a lower dE than I have measured from ANY non-CMS display I've ever seen. In my experience, most displays average RGB dE76 is 15 and higher. Below 10 is rare. What projector manufacturers are you thinking of?

This sounds a little hyperbolic.
:) Perhaps I should have been more clear about what I meant by "significant", or simply said "more important" (to me obviously, since we are talking about issues of perception). I think this is an example where it is important to not let numbers substitute for perceiving the "big picture". In many cases (it depends of course on image content and decisions by the telecine colorist as well) when material mastered using SMPTE-C primaries is viewed with Rec 709 primaries red objects and even flesh tones become overly saturated to the point that they are quite obviously unnatural. That to me is a very significant (important) error that destroys my desire to watch the source material without taking corrective action. From the Rec. 709 red primary to an even more saturated red (as used by many projectors) is not as significant (important) to me, the damage is already done. So even though the dC error (more appropriate to discuss than dE in this case) may be somewhat larger (magnitude depends on the projector of course) from Rec 709 to projector X red than from SMPTE-C to Rec 709, it isn't nearly as significant (important). Put another way, once the food is too spicy, it's not nearly as significant that it could be even more spicy.

Given the current state of video source material, it is more significant (important) to have (or be able to adjust) a projector with reasonably accurate SMPTE-C primaries than it is to have a projector with reasonably accurate Rec. 709 primaries. If you read my Sony VPL-VW60 review, you will see that point made several times about its improved Normal color space mode.

There is a new SMPTE Broadcast Monitor Report available

"Report Objective:

The objective of this report is to supply monitor manufacturers, equipment distributors video and film production and post-production houses and independent video content creators with the information needed in the switch from CRT to LCD monitors. In addition to a discussion of the existing CRT and LCD technology, the report will cover other technologies that may play a role in the future. Representative products from all major manufacturers intended for all niches within the professional monitor market are discussed."

http://www.insightmedia.info/reports/2007smpte-pvm.php

$999.00 [$799.00 for SMPTE members]

As can be seen from the cost of this report, getting the word "from the horse's mouth" can be a significant financial commitment for the hobbyist. However, this changeover is still going to take time. Even then, consumer displays should be designed to correctly reproduce rec. 601/SMPTE C and rec. 709. There will be DVD collections viewed for a long time yet, all of which were mastered in rec. 601/SMPTE C.

Best regards and beautiful pictures,
Alan Brown, President
CinemaQuest, Inc.

"Advancing the art and science of electronic imaging"

So ---- there is no issue (other than just human mistakes) associated with the Rec. 601 and Rec. 709 color encoding matrix equations. The issue is that professional monitors, which were about 99.999% CRT monitors until recently, do NOT have Rec. 709 primaries, they have SMPTE-C primaries. Therefore, when the telecine colorist is transferring a film to video, and adjusting colors so they look correct to him, he is looking at a display with SMPTE-C primaries, which produces decidedly different colors than a projector with Rec. 709 primaries. So if you want to see what the telecine colorist saw when transferring the film to video using a SMPTE-C monitor, you will need a display with SMPTE-C primaries.


Gregr:

Is my undersatnding therefore correct that we should -- at least for the immediate future -- be calibrating our displays to the SMPTE-C color gamut as this is what is used in the process of film to video transfer?

What about HD material that original video material (i.e. the above paragraph deals with the issue of film based material that is transferred to video, here I m asking about material that "starts its life" as video)...should be using an SMPTE-C color gamut or a REC 709 color gamut?

TIA.

I've been using 3dluts for color correction of film and video ( both rec.709 and EBU) for about the last 5 years. Assuming your monitor is reasonably healthy you can maintain accuracy for just about any standard ( video is easy the film standard(s) are the tough ones). Every so often someone suggests we get a broadcast CRT in to better see what the video stuff looks like and I have to re-educate them and point out how much easier it is to reprofile a workstation display and keep it up to date than do a coarse hardware cal on a CRT and get the scope out to check the voltages on the guns and all that palava.

This way of working is not a particularly big deal and is pretty much universal at most places that create or color correct .

I couldn't even tell you what phosphors my screen uses ( some sony flavour) all the profiling does is take a look at it and tell me if its possible to create a working LUT that gets it safely under delta4 for a given standard (usually its under 2 for the video standards without me even bothering to optimise the hardware.

Given the current state of video source material, it is more significant (important) to have (or be able to adjust) a projector with reasonably accurate SMPTE-C primaries than it is to have a projector with reasonably accurate Rec. 709 primaries. If you read my Sony VPL-VW60 review, you will see that point made several times about its improved Normal color space mode.


So given this Greg, one should calibrate all inputs using SMPTE-C fields for setting color correct? In relation to primaries/secondaries? Because the only projector I know of that lets you force what standard you watch is the Samsung projector that Joe did.

Marantz and Infocus give you the choice between NTSC and ATSC color space.

Just to help visualize precisely what we are talking about.

http://home.comcast.net/~tlhuffman/color_space.gif

Kinda looks lilke 'much ado about nothing' to me. :confused:

Does real world viewing translate into greater differences than this chart would show?

Wayne

So given this Greg, one should calibrate all inputs using SMPTE-C fields for setting color correct? In relation to primaries/secondaries? Because the only projector I know of that lets you force what standard you watch is the Samsung projector that Joe did.I've said for a long time that projectors should provide switchable SMPTE-C and Rec. 709 primary colors (if implemented correctly the complementary colors and all other colors simply follow from the primaries), in addition to whatever native primaries they may have. That is reasonably easy to do with a good CMS (color management system) in the projector. The Yamaha DPX-1200/1300 (1300 had a few improvements) had the best CMS for that purpose that I have seen. It provided the ability to enter the primary colors by their x,y values fairly accurately for users without measurement equipment (and of course could be adjusted more accurately with measurement equipment) and it would calculate the correct Lightness values for each color automatically (or that could be entered manually too). It also provided the ability to independently adjust the complementary colors in the same way, but that really wasn't necessary because the CMS was designed correctly. This was one major contributing factor to why I thought the Yamaha DPX-1300 was the best 720p projector (although it also had its flaws). The Sharp XV-Z series also provides a good CMS that will also allow you to get pretty close to accurate SMPTE-C and Rec. 709 primaries. It shouldn't be surprising that I gave very high marks to both the Yamaha and Sharp projectors for their color accuracy.

Anyway, to your question directly, I believe that for the moment, most of the time the SMPTE-C primaries provide more natural color than Rec. 709 primaries with the currently available HD sources. There are certainly exceptions, where non-CRT monitors were used for monitoring, or where electronic CMS systems were used for partial correction of the color gamuts, or where telecine colorists judgement calls are significant to the results, etc. Therefore, I think it is really beneficial to be able to switch between both color gamuts and pick the one that works best with each piece of source material. When set up to do that kind of A/B comparisons I think a lot of people will agree with me that the color differences are very significant after they have looked at a variety of source material (particularly those rich in reds and purples - and/or - that have more saturated flesh tones). Of course, not everyone will agree with me - and that's perfectly fine - the majority of people prefer the even more saturated native primaries of most projectors (else the manufacturers wouldn't be making them that way - you can be sure of that !!!). I don't feel like its my role to convince users what they should like (but to help educate those that want to know more about video technology) - it's my role to report how projectors perform and then let users decide what they want based on what they like. (I do feel it's my role to try to influence manufacturer's to make projectors that perform accurately to the applicable standards - but they can do that and still provide other switchable modes that they believe generate more sales than adhering to standards.)

Does real world viewing translate into greater differences than this chart would show?
Both the pro version in WMV and the HD DVD version of 'Digital Video Essentials' contain side by side color patches and split screen panned images showing 601 vs. 709 color decoding. You can decide for yourself if there is a noticeable difference and whether you find the difference worthy of concern. Of course, it helps to view it on good display. It's well established that human vision is more sensitive to subtle differences in hue than slight variations in brightness.

I have a bit of a problem with your term, "real world." If you mean the average consumer, I would assert that the masses don't really care very much about the subtleties of image fidelity. If you mean the program production community at large, there ought to be sufficient concern about observing standards. Without them, there can be no reliable communication of art. Should you be referring to the world of consumer display manufacturing, it's plain to see where that has already gone. These are all guesses as to your possible meaning. Perhaps you care to define your terms more clearly?

Best regards and beautiful pictures,
G. Alan Brown, President
CinemaQuest, Inc.

"Advancing the art and science of electronic imaging"

Kinda looks lilke 'much ado about nothing' to me. :confused:

Does real world viewing translate into greater differences than this chart would show?

WayneI don't know how you decided what the chart shows in terms of perceptible color differences. BTW, the x,y chart is not based on uniformly perceptible color differences, anyway. The CIE u'v' chart is more perceptibly uniform, but still isn't really very uniform. But even if the chart were perfectly uniform, how would you know what the differences represented unless you actually compared the two color gamuts on a variety of real material? So to answer your question, I believe the difference is much greater than your initial comment would imply.

Both the pro version in WMV and the HD DVD version of 'Digital Video Essentials' contain side by side color patches and split screen panned images showing 601 vs. 709 color decoding. You can decide for yourself if there is a noticeable difference and whether you find the difference worthy of concern. Of course, it helps to view it on good display. It's well established that human vision is more sensitive to differences in hue than variations in brightness.Color decoding and color gamut are two different issues. The chart that Tom posted is a comparison of color gamut, not a comparison of color decoding. Refer to my explanation of the two concepts posted earlier in this thread if that is not clear.

Thanks for the correction. Would you say the differences in appearance between SMPTE C and rec. 709 color gamuts would be more or less subtle than those between the two types of decoding?

Kinda looks lilke 'much ado about nothing' to me. :confused:

Does real world viewing translate into greater differences than this chart would show?

Wayne

To kind of add to this it will depend on the person. You can look at the charts till your blue in the face, but until you see the differences then you won't know if it is a concern for you or not.

I am not sure if this is a good analogy, but I use a Datacolor spectrophotometer for mixing colors. I believe it is accepted that a dE of 1 can be seen by most people (especially with light colors). Using four pigments for a color match, just one pigment being incorrect by a small percentage can cause a dE of 1 or greater. For some, it is not a big deal and for others its is the difference between black and white. My guess is Greg and some others around here are like me and it is like black and white. When you deal with colors on a daily basis, it becomes more noticeable. I hope this analogy works, if not then disregard.:)

Thanks for the correction. Would you say the differences in appearance between SMPTE C and rec. 709 color gamuts would be more or less subtle than those between the two types of decoding?When the color decoding is wrong (mismatched Rec. 601/Rec. 709 encoding and decoding) the brightness of green is probably the most perceptible error to most people, although some people will readily recognize other errors when comparing color bars encoded/decoded correctly vs incorrectly. So if you are looking at color bar test patterns I would say the decoding error is considerably more noticeable than the primary color differences. But with real sources perception is driven by content. If you are watching sports played on grass, or movies with a lot of fields and green landscapes, the color decoding error should get your attention. But as I said earlier, if you watching sources with familiar red or purple objects, or highly saturated flesh tones, then I believe you will readily notice the color gamut error. So the error that you perceive to produce the more unnatural colors depends to a considerable extent on the content and your familiarity with the colors of particular image features. Most people have a good sense of what flesh colors should look like so color gamut errors tend to really bother some people (like me) a lot. However, oddly (it seems to me) other people seem to be able to just tune out poor flesh colors. Perhaps this is a result of watching very poor color on uncalibrated consumer TVs for many years (which fortunately I haven't done in a very long time).

Here's the more perceptually uniform 1976 chromaticity diagram that Greg referred to.

http://home.comcast.net/~tlhuffman/luv.gif

GeorgeAB,

Thanks for your reply. I do have both the SD and HD DVD versions of DVE and AVIA also. I can check out the color patches and split screen panned images as you suggest in order to get an idea of the differences and whether this is truly 'much ado about nothing' to me.

What I meant by 'real world' was movies on SD and HD DVD, camcorder recordings, and broadcasts whether SD or HD etc...

Wayne

When the color decoding is wrong (mismatched Rec. 601/Rec. 709 encoding and decoding) the brightness of green is probably the most perceptible error to most people...
So does that mean for example that the source of complaints about the RS1 having too much green in the image (or neon greens, etc.) really is the result of bad color decoding?

I don't know how you decided what the chart shows in terms of perceptible color differences. BTW, the x,y chart is not based on uniformly perceptible color differences, anyway. The CIE u'v' chart is more perceptibly uniform, but still isn't really very uniform. But even if the chart were perfectly uniform, how would you know what the differences represented unless you actually compared the two color gamuts on a variety of real material? So to answer your question, I believe the difference is much greater than your initial comment would imply.

Thanks for the reply Greg. I haven't actually decided anything to be honest. I can only go by what I perceive to be true from a layman's perspective. I'm not a calibrator or a reviewer, just an ordinary guy who enjoys projectors as a hobby. I don't understand very much of what you or any of the other fellow members say with regards to this topic. :o

My question was an honest one. It appeared to me by that chart that the two standards are so close as to beg the question, is there REALLY a difference in skin tones and colors that will perhaps show the Canadian or American flag as red or pink tinged? Will people look either sun-burned or like martians? :confused:

I appreciate you relating that the differences will be greater than my amateurish attempt to interpret the chart in question. All I want to know is, is there anyone who can give me an example from a movie for instance whereby the difference is very readily apparent?

I'm not looking for a technical answer which I won't understand. I want a layman type response from anyone at all who can explain what the differences are in skin tones and colors etc... if they are noticeable to the naked eye.

In other words, is the difference between the two standards more of a concern of those who spend a fair amount of time looking at test screens and charts? I have had CRT RPTVs ISF calibrated but never a projector. I've always winged it with calibration discs and more recently with the SpyderTVPro to calibrate my projectors.

I do think that I know what accurate colors are and what good, natural flesh tones are. I'm wondering if it's not possible to get both of these results when calibrating to either of the two standards.

Wayne

To kind of add to this it will depend on the person. You can look at the charts till your blue in the face, but until you see the differences then you won't know if it is a concern for you or not.

Thanks Eric. :)

I'm sure some people are more aggravated by the differences than others.

Wayne

Thanks for the reply Greg. I haven't actually decided anything to be honest. I can only go by what I perceive to be true from a layman's perspective. I'm not a calibrator or a reviewer, just an ordinary guy who enjoys projectors as a hobby. I don't understand very much of what you or any of the other fellow members say with regards to this topic. :o

My question was an honest one. It appeared to me by that chart that the two standards are so close as to beg the question, is there REALLY a difference in skin tones and colors that will perhaps show the Canadian or American flag as red or pink tinged? Will people look either sun-burned or like martians? :confused:

I appreciate you relating that the differences will be greater than my amateurish attempt to interpret the chart in question. All I want to know is, is there anyone who can give me an example from a movie for instance whereby the difference is very readily apparent?

I'm not looking for a technical answer which I won't understand. I want a layman type response from anyone at all who can explain what the differences are in skin tones and colors etc... if they are noticeable to the naked eye.

In other words, is the difference between the two standards more of a concern of those who spend a fair amount of time looking at test screens and charts? I have had CRT RPTVs ISF calibrated but never a projector. I've always winged it with calibration discs and more recently with the SpyderTVPro to calibrate my projectors.

I do think that I know what accurate colors are and what good, natural flesh tones are. I'm wondering if it's not possible to get both of these results when calibrating to either of the two standards.

Wayne

Like you I too am a hobbyist in that I truly enjoy learning, reading and viewing all things hime theater -- the point being that I am NOT a professional calibrator...that said, over the last 3 to 4 months I have started to learn about calibration and can tell you that I agree with Greg's comments in that the impact of the color decoder error versus color gamut error depends on the source material.

To answer your question I have calibrated my RadianxeXD processor such that I have memory inputs that:

1. calibrated for color decoding and SMPTE-C color gamut
2. calibrated for color decoding and REC 709 color gamut
3. not calibrated for color decoding but calibrated for SMPTE-C color gamut
4. not calibrated for color decoding but calibrated for REC 709 color gamut
5. not calibrated for either color decoding or color gamut

(and yes, I know that I am missing the calibrated for color decoding but not color gamut but heck, there is only so much time).

By toggling between the different configurations the changes/erros are very noticeable...right now, as an example, I am watching the MIchigan verus Florida football game on HD cable and can tell you that the difference between the uncalibrated color gamut versus either the SMPTE-C or REC 709 color gamuts is huge while the difference between the SMPTE-C and REC 709 color gamuts is noticeable but not as large...the point being, you need to have this ability to notice the impact and , by having this ability, one learns what to llok for (i.e. in this case, color of the grass, color of the uniforms and flesh tones).

HTH

So does that mean for example that the source of complaints about the RS1 having too much green in the image (or neon greens, etc.) really is the result of bad color decoding?Nope. The green primary on the RS-1 is way outside (oversaturated) the standards. (There was an RGB color encoding/decoding bug on the early RS-1's but that was corrected, and not the source of the "complaints".)

Hey Greg

I think people here are still having trouble differentiating between color decoding and color primary position. They are two seperate animals but easy to get confused in general conversation.

Appreciate the input Joel! Sounds like a nice VP you've got there. :cool:

Wayne

Appreciate the input Joel! Sounds like a nice VP you've got there. :cool:

Wayne

I have a RadianceXD and am quite happy with it...setting it up as noted above took a lot of time and work (and I am sure that Gregr, Chris, Tom, and others) could probably get more out of it than me as I am still learning this stuff...nonetheless it has:

1. Taught me what to look for;

2. Taught me how important a properly calibrated video chain is; and

3. Taught me the numerous steps which can cause inaccuries/problems.

Now that I have seen the differences I will limit the number of calibration to just two; those being

a. Proper color decoding and SMPTE-C gamut for day time viewing;
b. Proper color decoding and SMPTE-C gamut for day time viewing
c. Proper color decodingand REC 709 gamut for day time viewing; and
d. Proper color decodingand REC 709 gamut for night time viewing,

but will not get around to setting up items b. and d. until the RadianceXD software is closer to final -- to be clear, The RadianceXD is terrific but Lumagen is still finalizing things and subsequent software releases sometimes require entire recalibrations...it is time consuming enough doing it for 2 setups that doing again for 3+ is not in the cards..

HTH

Hey Greg

I think people here are still having trouble differentiating between color decoding and color primary position. They are two seperate animals but easy to get confused in general conversation.Yes, I agree. It's been confused several times above. But I'm having trouble figuring out how to explain the difference and be more clear.

Color decoding is how signals are converted from YCbCr to RGB, color encoding is how signals are converted from RGB to YCbCr. All projectors with analog YPbPr or digital YCbCr inputs must have the ability to switch between Rec 601 and Rec 709 color decoding in order to be compatible with both standard-definition and high-definition input signals. Some projectors (not all) allow the user to manually select which standard is used for color decoding. Sometimes they call that a Color Space selection, which is probably what confuses a lot of people.

SMPTE-C primaries or Rec. 709 primaries refer to the desired color of the red, green, and blue primaries used by the projector. In today's projectors the actual (native) colors of the red and green primaries are almost always more saturated than the colors defined by either of those standards. In many cases, there is nothing the user can do about that in the projector (although an external CMS - what Joel is doing - may be used). If the primaries are wrong, the colors produced by the projector will be wrong (although some people prefer that the primaries are wrong - that is a different issue). But in some projectors an electronic Color Management System (CMS) is provided, which allows the video signals to be modified so that 100% red, green, and blue signals produce colors (virtual primaries) different than the native primaries of the projector. In that case, the CMS can sometimes be used to calibrate the virtual primaries so that they are at (or close to) the location of the SMPTE-C and Rec. 709 primaries. i.e. a virtual color gamut is created electronically that can be made to match the standards. In that case, if the CMS is designed correctly, all of the colors within the virtual color gamut are modified to match the standards, and the projector can produce accurate colors.

The Marantz was cited earlier as having this ability - it does not. The InFocus was also cited as having this ability - I haven't seen one of those in a while but earlier versions did not have this ability. The Sharp XV-Z projectors have such a CMS, as did the Yamaha projectors I mentioned earlier. I've been told the Samsung projector mentioned by Chris had the ability to be calibrated in this way, but I don't know what user adjustments it had. Not all projectors that advertise or claim a CMS system have the ability to be calibrated correctly to the standards because they may lack the ability to make all of the necessary independent adjustments. i.e. just because a projector allows you to change its primary (or other) colors doesn't mean it allows you to do it correctly to match (or get close to) the standards. Some of the Sony projectors have a Wide and Normal color space mode. In the VPL-VW60 the Normal mode is relatively close to the SMPTE-C primaries, but it wasn't in previous Sony projectors that had Wide and Normal modes.

I've been told the Samsung projector mentioned by Chris had the ability to be calibrated in this way, but I don't know what user adjustments it had.

I believe the Samsung had the most intelligent setup process of any projector: It's basically designed to do the work for you, no need to manually set up each format: You basically measure built-in test signals to determine the native Red, Green, Blue and White, punch in the x/y/Y values, and then the projector calculates which offsets needs to be made to get to the correct color space and color decoding. You can select between HD, SD and Auto. It has Blue-only, red-only and green-only to verify color decoding, if the source is perfect, the color decoding is spot-on. Color space is as accurate as anything you'd find from any projector with a "normal" CMS. The only thing we need Samsung to do, is to make the hardware actually work... But color-wise, I don't think it's possible to do it better than this, and personally I prefer this way of implementing correct color, instead of having to manually adjust color via a CMS, and I wish other manufacturers would copy this approach. If I understand it correctly, both DLP and plasma should be linear technologies, so in both cases it should be possible to calculate the offsets from the internal test patterns, without the need for manually adjusting a CMS. It does assume that the source works correctly though. Of course, if all manufacturers implemented this, it would be way too easy to be a calibrator... ;-)

I believe the Samsung had the most intelligent setup process of any projector: ...The way you described calibration is also what I understood. I just wasn't sure what user controls you had after calibration. Sounds like you can then choose Rec 709, SMPTE-C, or an Auto mode. Very good!

Yes, I agree. It's been confused several times above. But I'm having trouble figuring out how to explain the difference and be more clear.

Color decoding is how signals are converted from YCbCr to RGB, color encoding is how signals are converted from RGB to YCbCr. All projectors with analog YPbPr or digital YCbCr inputs must have the ability to switch between Rec 601 and Rec 709 color decoding in order to be compatible with both standard-definition and high-definition input signals. Some projectors (not all) allow the user to manually select which standard is used for color decoding. Sometimes they call that a Color Space selection, which is probably what confuses a lot of people.

SMPTE-C primaries or Rec. 709 primaries refer to the desired color of the red, green, and blue primaries used by the projector. In today's projectors the actual (native) colors of the red and green primaries are almost always more saturated than the colors defined by either of those standards. In many cases, there is nothing the user can do about that in the projector (although an external CMS - what Joel is doing - may be used). If the primaries are wrong, the colors produced by the projector will be wrong (although some people prefer that the primaries are wrong - that is a different issue). But in some projectors an electronic Color Management System (CMS) is provided, which allows the video signals to be modified so that 100% red, green, and blue signals produce colors (virtual primaries) different than the native primaries of the projector. In that case, the CMS can sometimes be used to calibrate the virtual primaries so that they are at (or close to) the location of the SMPTE-C and Rec. 709 primaries. i.e. a virtual color gamut is created electronically that can be made to match the standards. In that case, if the CMS is designed correctly, all of the colors within the virtual color gamut are modified to match the standards, and the projector can produce accurate colors.


Gregr:

The only twoi ideas that I can come up with to better explain the diferrence between the color decoding/encoding process and the color gamut are:

1. A block diagram which visually explains the process and thus allows one to see, for example, that correctly encoded/decoded YCbCr signals will look wrong with an incorrect color gamut; or

2. Actually post pictures from a setup showing the differences along the lines of what I did...I would do this but do not have the photography equipment to do this...perhaps someone here with such equipment (and the interest/time to do the various calculations) will be able to do so...

Cheers,

The InFocus was also cited as having this ability - I haven't seen one of those in a while but earlier versions did not have this ability.

I think I'm the one who mentioned IF, and there's either some confusion or I wasn't clear, IF provides selection of color decoding (Auto, REC601, REC709, RGB), but provides no way to change the primaries/gammut, at least not on the IN76 and not in the user menu.

Are you talking about the new Samsung SP-A800B 1080p projector? Why isn't anyone talking about this PJ? Is it really that bad? If the A800B does this, it will be pretty cool. The only thing left would be to get rid of the color wheel!

Yes, it will do this. The main reason noone is talking about it is that 1. It's not out yet, and 2. A lot of people have experienced or read about reliability issues with the 710, and are waiting to see if it will actually be reliable when it's out. Other than that, I suspect the SPH-800 have been in the pipeline so long that when it does become available, it won't be as competitive in other areas, not least contrast.

I owned the 710, but replaced it due to reliability. I would hope that some other manufacturer, preferrably Sim2, would duplicate the color design of the Samsung. I'm not sure it will ever happen, and even with other, more complicated designs, some manufacturers are getting increasingly accurate. As neat as the design of the Samsung is, what counts is what's actually up on the screen. I'm quite happy with the D80e, for instance.

I think a lot of people were disappointed that the A800B was announced as a DC2 projector at CEDIA when most projectors are already DC3 and some have been announced as DC4. I like the idea of a projector that can be calibrated to near perfection but I don't want to have to trade off too much to get it. It will be interesting to see how the 800 performs as they are doing a dynamic iris.

Kris-

I read somewhere that Joe Kane thinks the DC2 gives better shadow detail than the DC3.

Infocus has software for dealers to adjust primaries/ secondaries on some models. Marantz has a similar program but not as easily obtained. The adjustments for the Sasmsung 710 are in a calibrator/service menu. Either way these are features not available to the user.

Bob

One helpful way to think of color decoding errors is to think of them not in terms of how they are caused, but rather in terms of the symptoms they produce.

1) If you have a properly calibrated gray scale and have set Color/Tint correctly, but you still have hue errors on one or more of the secondaries then this is probably a color decoding error.

2) If you have lightness errors in one or more of the primaries, then this is probably a color decoding error. This type of error does not appear on chromaticity charts at all. It can only be diagnosed by using RGB filters with SMPTE color bars or by direct measurement.

The way you described calibration is also what I understood. I just wasn't sure what user controls you had after calibration. Sounds like you can then choose Rec 709, SMPTE-C, or an Auto mode. Very good!

Come to think of it: It has settings for BT-601, BT-709 and Auto. I'm 98% certain, but not 100% certain, that these affect both colorspace and color decoding. I never actually measured the difference, but as I remember it Joe explained it this way, as one of the arguments that his projectors are the ONLY projectors (at that time, at least) to fully do accurate colors, because other displays might have different decoding, but not different colorspace. I'll try to get confirmation on this.

In any case, the setup is extremely neat, and like I said, should have been copied by now. I would think that if you have accurate colorimetry in your signal path, this would be _easier_ (and thus, cheaper) to implement than any other hard-core CMS. I think we can all agree that if the picture is accurate in the first place, who needs a CMS? In my opinion, CMS is there to correct an error that shouldn't have been there in the first place. I believe the cry-out for CMS systems should have been a cryout for the manufacturers to simply integrate an accurate mode (which would obviously still need some level of individual tweaking, just as the Samsung does). At least in DLP's and plasmas, which, as I understand it, should be linear technologies, so calculating different colorspaces should prove functional in practice. Considering that manufacturers are actually starting to "give in" and give us SOME levels of CMS (although most are poorly implemented), it would seem that a lot of manufacturers ARE actually listening somewhat to the reviewers, installers and enthusiasts who want accurate pictures. They just don't understand how to give it to us. Would it actually be easier or harder to convince them to provide a "studio mode" that's quite simply accurate (only small tweaking needed), than to persuade them to provide a full CMS? If their signal processing for some odd reason has some non-linear behaviour (i.e. color decoding in Pioneer plasmas), even giving us acces to full CMS won't give us the full potential of accurate pictures. What the displays SHOULD have, is a "studio mode" that works the same way as Joe's Samsung design: Fully linear and accurate signal processing, with just the need to measure four simple test patterns. I guess this could even be used to easily implement an easy-to-use (oh, dare I even say it...) auto-calibration feature.

Or am I just dreaming?

2) If you have lightness errors in one or more of the primaries, then this is probably a color decoding error. This type of error does not appear on chromaticity charts at all. It can only be diagnosed by using RGB filters with SMPTE color bars or by direct measurement.

BTW, thinking of the original post, I believe this is a MUCH bigger issue with reviews (projectors or flatscreens) than the difference between 601/709 color space charts: A _lot_ of reviewers will calibrate greyscale, measure that the primaries and secondaries are in the right spot on the CIE, and then conclude that the colors are PERFECT, without even mentioning color decoding. It is certainly possible to have perfect greyscale and perfect primaries/secondaries, but a horribly wrong picture.

I think a lot of people were disappointed that the A800B was announced as a DC2 projector at CEDIA when most projectors are already DC3 and some have been announced as DC4. I like the idea of a projector that can be calibrated to near perfection but I don't want to have to trade off too much to get it. It will be interesting to see how the 800 performs as they are doing a dynamic iris.

It is delivered with auto-iris off, and from the hands of Joe intended to function without the iris. The iris is there because Samsung wanted to market as high a contrast ratio as possible, but Joe made sure that the control to turn it on is buried somewhere deep in the menus, to make sure most people would never find it... (those were his own words). I'm not an engineer, so I may be wrong here, but I believe that the gamma warping needed to make an iris work, might affect the linearity of the device. Don't know it that will actually affect color fidelity, but as Joe is equally committed to gamma as to color accuracy, it's certainly not something Joe would want you to use.

About the DC2, I think that any excuse Joe might give for using the older chip is just a cover-up for the fact that Samsung has been holding this PJ off the market, waiting for the right time to launch it... Pure speculation though. The indications of price seem fair though, even using a DC2. There will certainly be cheaper DC3's outthere, but the SPH-800 does use what seem to be a quite high-quality lens, compared to the 710 and any other 1080P below 10k$. All a matter of tradeoffs - like I said, personally I'd have preferred the same layout from a manufacturer that I trust more than Samsung, even if it was double the price.

Yes, I agree. It's been confused several times above. But I'm having trouble figuring out how to explain the difference and be more clear.

Color decoding is how signals are converted from YCbCr to RGB, color encoding is how signals are converted from RGB to YCbCr. All projectors with analog YPbPr or digital YCbCr inputs must have the ability to switch between Rec 601 and Rec 709 color decoding in order to be compatible with both standard-definition and high-definition input signals. Some projectors (not all) allow the user to manually select which standard is used for color decoding. Sometimes they call that a Color Space selection, which is probably what confuses a lot of people.

SMPTE-C primaries or Rec. 709 primaries refer to the desired color of the red, green, and blue primaries used by the projector. In today's projectors the actual (native) colors of the red and green primaries are almost always more saturated than the colors defined by either of those standards. In many cases, there is nothing the user can do about that in the projector (although an external CMS - what Joel is doing - may be used). If the primaries are wrong, the colors produced by the projector will be wrong (although some people prefer that the primaries are wrong - that is a different issue). But in some projectors an electronic Color Management System (CMS) is provided, which allows the video signals to be modified so that 100% red, green, and blue signals produce colors (virtual primaries) different than the native primaries of the projector. In that case, the CMS can sometimes be used to calibrate the virtual primaries so that they are at (or close to) the location of the SMPTE-C and Rec. 709 primaries. i.e. a virtual color gamut is created electronically that can be made to match the standards. In that case, if the CMS is designed correctly, all of the colors within the virtual color gamut are modified to match the standards, and the projector can produce accurate colors.

The Marantz was cited earlier as having this ability - it does not. The InFocus was also cited as having this ability - I haven't seen one of those in a while but earlier versions did not have this ability. The Sharp XV-Z projectors have such a CMS, as did the Yamaha projectors I mentioned earlier. I've been told the Samsung projector mentioned by Chris had the ability to be calibrated in this way, but I don't know what user adjustments it had. Not all projectors that advertise or claim a CMS system have the ability to be calibrated correctly to the standards because they may lack the ability to make all of the necessary independent adjustments. i.e. just because a projector allows you to change its primary (or other) colors doesn't mean it allows you to do it correctly to match (or get close to) the standards. Some of the Sony projectors have a Wide and Normal color space mode. In the VPL-VW60 the Normal mode is relatively close to the SMPTE-C primaries, but it wasn't in previous Sony projectors that had Wide and Normal modes.

I think that explanation is clear, if somewhat technical. Beyond just the fact that both have to do with the accuracy of colors, it's also confusing for some advanced people because the differences between the Rec 601 and Rec 709 matrices derives from differences in the gamut, the former based on the original NTSC gamut, and the latter based on the 709 gamut. So if someone is aware of that, and understands that the matrix differences are based on different theoretical gamuts to derive the best perceptual component video, then it's all the more confusing because now the explanatory separation of the two different questions of the component video matrix and the gamut get smashed back together again.