There has been a lot of discussion on the forum about on/off lately and in particular how much is enough. Some forum members want CRT level performance, others feel that 3k:1 is enough.

I've also heard respected members in the industry make comments that, there is an "obsession" with people wanting more contrast from digitals. Even the SMPTE study goup that follows display devices that are replacing CRT's don't seem certain if a specific contrast spec should even be specified and in fact they recently said, "The WG should determine whether in fact specification of contrast ratio is required".

So I'm curious to hear what people think and why.

I don't know how many people know this, but no too long ago I had the opportunity to do a shootout between an RS1 and a C3X 1080. After equalizing the brightness between the two projectors, it was apparent that the C3X's performance in many low APL scenes suffered in comparison to the RS1. I attribute this difference in performance to the difference in on/off contrast between the two machines.

Still, the RS1's performance in low apl scenes can still be improved on a great deal. I hope that I can be happy with what the RS20 will provide.

The real answer is that until black is really black and until a fade to black means that you can't see anything no matter how hard you try, there will still be room for improvement.

I voted 30'000:1 but in practice I could go for less depending on other factors, in a few years time 30'000:1 will seen very low.

Mark, The question is impossible to answer in the form of a number.

If a projector has, lets say 300 lumens and a CR of 100k, its black floor will be very low. If you double the CR by halving the black level it will probably be unnoticeable. You could go on and on and the difference to the image would be nil. Do the same by doubling the lumens and its a very different story.

I've never heard anyone claim 3k to be enough. I certainly don't.

The real benefit of CR is when borne in relation to lumen output.

I voted 250K and up. The reason there is an "obsession" with contrast, is that more contrast is always more pleasing (to my eye). Of course the next issue is "what is contrast" - on/off, ANSI, intra-scene, etc...

High on/off is mostly useless to me as a measurement unless you like pausing fade to black so you can marvel at how black it is. Contrast at 1% APL is a more realistic measure of things to me.

There has been a lot of discussion on the forum about on/off lately and in particular how much is enough. Some forum members want CRT level performance, others feel that 3k:1 is enough.

I've also heard respected members of the industry make comments that, there is an "obsession" with people wanting more contrast from digitals. Even the SMPTE study goup that follows display devices that are replacing CRT's don't seem certain if a specific contrast spec should even be specified and in fact they recently said, "The WG should determine whether in fact specification of contrast ratio is required".

So I'm curious to hear what people think and why.

Mark, I had assumed this was a genuine post, and as such gave a genuine answer.

I now appears it may be just a vehicle to reopen the same stuff that has been done on the other thread.

Maybe that wasn't your intent, and LG simply hijacked it. If that's the case, sorry for assuming the worst.

I've seen enough extended dark scene material on my RS1 to recognize that it's handicapped in those scenes. I'll be very curious to see if the high contrast iris settings on the RS20 of say 40k:1 will be "enough". In scenes that aren't extended and don't allow the eye enough time to recover the RS1 seems fine, so it really takes extended scenes to really notice the limitations. A couple of good examples are the horror flicks, "Wind Chill" and "Aliens vs Predators: Requiem".

I don't know how many people know this, but no too long ago I had the opportunity to do a shootout between an RS1 and a C3X 1080. After equalizing the brightness between the two projectors, it was apparent that the C3X's performance in many low APL scenes suffered in comparison to the RS1.

I cant believe you will ruin a potentially interesting thread by restarting something that should be confined to the other thread.

Now its graduated to a "shootout". We both know, for the obvious reasons, it never happened

If by enough you mean enough to accurately represent the consumer software "mastering intent", then I would say 15,000:1 since IIMU that most mastering monitors (CRTs) are in the 10K:1 to 12K:1 range.

When we talk about CRT FPs with insane on/off numbers (>750K:1), IIMU we are talking about gamma adjusted devices. CRT FP when marketed new were speced at < 40K:1 and properly setup were ~10K:1.

I haven't seen either Wind Chill or AVP:Requiem. I just put them to the top of my Netflix Queue.

I don't know the answer to this question. That is a major reason why I am thinking of picking up an RS20. I very much prefer the DLP look, and am thinking of the new C3X1080 for a new theater I am planning. Playing with the RS20 will give me a better idea of what the increased CR gives me. If it is significant, I will probably hold out for LED.

Mark, I had assumed this was a genuine post, and as such gave a genuine answer.

I now appears to be just a vehicle to reopen the same stuff that has been done on the other thread.

Maybe that wasn't your intent, and LG simply hijacked it. If thats the case, sorry for assuming the worst.

It was a genuine post. One that I was going to start even before the other thread. The reason that I've been mulling this over is based on comments that I read on SMPTE and also respected people like Pete Putman. I get the impression (perhaps wrong) that most of the industry feels that 10k:1 or so is "good enough".

I think LG was just discussing the on/off differences between a C3X and an RS1 rather than which one is better. So it seems like those comments fit and aren't a hijacking of this thread. I know when I replaced my old 2k:1 projector with the RS1 that the difference was night and day and not contained to the mostly extended dark scenes that I mentioned earlier. With that significant of a on/off contrast difference the intra-image contrast differences extended well into brighter material. My guess is the higher the on/off the darker the material one can view before the image starts to fall apart.

I've been calculating this very question for my own purposes and have determined that around 50,000 or above would satisfy my craving for great black level and satisfying brightness. To determine the level of black which (presently) offers me very satisfying blacks I lowered on screen foot Lamberts, when I reached about 9 fL's blacks in all but total black screen look terrific and I'm sure that black level which, in this configuration calculates to about .00036 fL's, would be very satisfying in most all situations. Now to get minimum starting max brightness of 15 fL's at the upper end I'll need a pj with a contrast ratio of about 41,000 with my present throw. That means a pj with advertised CR of about 50,000 with initial real calibrated lumens of 600 available at that CR would satisfy my contrast needs just fine. No 50,000 so voted 80,000

If by enough you mean enough to accurately represent the consumer software "mastering intent", then I would say 15,000:1 since IIMU that most mastering monitors (CRTs) are in the 10K:1 to 12K:1 range.

When we talk about CRT FPs with insane on/off numbers (>750K:1), IIMU we are talking about gamma adjusted devices. CRT FP when marketed new were speced at < 40K:1 and properly setup were ~10K:1.

The interesting thing about the SMPTE study group that I referenced is there was a write up in this month's magazine discussing an update of their current status and findings with regards to various display parameters in preparing for the shift away from CRT. They cite the importance of many display parameters including gamut, gamma, grey scale tracking, etc. But then openly wonder if a specification for contrast is even needed.

It was a genuine post something that I was going to start even before the other thread. The reason that I've been mulling this over is based on comments that I read on SMPTE and also respected people like Pete Putman. I get the impression (perhaps wrong) that most of the industry feels that 10k:1 or so is "good enough".

Thats what seems to be going on. The fact is that most new leaps in CR are achieved by changes in black level that are of such small parts of a fL that we get to the point that astronomical spec differences make no difference to the viewing experience. When we get to the point of having huge CR on high end units, it will get interesting. The new Christie looks to hit 10k as does the Meridian and JVC 4k units. The new C3X1080 is spec'd at 30k.

Within 2 years CR will be a meaningless parameter anyway. A couple of LED iterations (not the TI being currently shown) already look set to achieve this.

On/off needs to be accompanied by high ANSI CR, and vice-versa, for contrast nirvana.

I think LG was just discussing the on/off differences between a C3X and an RS1

No he wasn't, and its blatantly obvious.

No he wasn't, and its blatantly obvious. It was simply vile opportunism.

I have my own opinions about that, but rather than destroy another thread, I'll just say that I'll leave it up to others to decide.

The interesting thing about the SMPTE study group that I referenced is there was a write up in this month's magazine discussing an update of their current status and findings with regards to various display parameters in preparing for the shift away from CRT. They cite the importance of many display parameters including gamut, gamma, grey scale tracking, etc. But then openly wonder if a specification for contrast is even needed.


The problem for the industry in adopting a contrast specification at-this-moment-in-time is that whatever standard would be adopted now would later be regretted as technology progresses. Few, if any, currently available projectors provide satisfying contrast. Why adopt a standard that is unsatisfactory to everyone?

The problem for the industry in adopting a contrast specification at-this-moment-in-time is that whatever standard would be adopted now would later be regretted as technology progresses. Few, if any, currently available projectors provide satisfying contrast. Why adopt a standard that is unsatisfactory to everyone?

But then again they could set a minimum standard. They seem to be wanting to specify gamma (although they didn't say what that number should be) so you would think they would specifiy a minimum level of on/off needed to get good results with that gamma.

Is this study that you refer to available anywhere that I could read it?

.

I've also heard respected members in the industry make comments that, there is an "obsession" with people wanting more contrast from digitals.

Thats very true. Whilst I would always want more of everything, CR is a single component to a good image. Uniformity, color, MTF, ANSI, sharpness, motion handling etc all play their part.

There are a number of machines available that have a CR no higher than 4k (some lower) that make most PJs discussed here, and the high end forum, look like toys. Image quality is a composite result of many parameters.

Is this study that you refer to available anywhere that I could read it?

It's in the latest SMPTE "motion imaging magazine". I think they have it online via their website but you'll hve to join to read it. There are a lot of good reasons to join but not really for that particular article.


I haven't seen either Wind Chill or AVP:Requiem. I just put them to the top of my Netflix Queue.


Neither are notable movies and I wouldn't recommend wasting time watching them with friends. But they both contain extended dark scenes that are on/off acid level tests.

Btw, both contain extended dark scenes with little to no bright pixels so both could also be used for showing the benefits of a DI because there will be little to no downside in the form of BC. I suspect that if someone with a DI watched some of the scenes with the DI engaged and then again with it off that it would make a huge improvement to have it engaged.


Thats very true. Whilst I would always want more of everything, CR is a single component to a good image. Uniformity, color, MTF, ANSI, sharpness, motion handling etc all play their part.


Yes this is true but this thread is only examining on/off and isn't meant to be a comparison between projectors other than just comparing on/off itself.

Yes this is true but this thread is only examining on/off and isn't meant to be a comparison between projectors other than just comparing on/off itself.

I know what you are saying.

The point I was making is that CR cannot be meaningful in isolation, nor can any other parameter. I didn't mean it to appear as a comparison, just an example of my point regarding image as a whole.

Well, I posted this in another thread, but it never really got a response. My "theory" based on what scientific data (NASA data on HVS capabilities) we have:

The Human Visual System is capable of perceiving a total dynamic range on the order of well over 10^12:1 (without damage), or upwards of 10^15:1 through it's entire range.
http://www.avsforum.com/avs-vb/showpost.php?p=14173560&postcount=31
http://msis.jsc.nasa.gov/sections/section04.htm#_4.2_VISION

If you were to say, for reproduction purposes, define/limit "Peak White" as the luminance of white paper in sunlight, a display that could reproduce that would need an On/Off or sequential CR of 10^10:1 (100,000,000,000:1, 100 billion:1) to limit it's darkest black to a level equal to the lower threshold of human vision.

And while a given cell of film or digital capture can not capture this entire dynamic range (film being 5-7 stops or 32-128:1, and digital perhaps being 10-12 stops or about 1000-4000:1), that doesn't mean that across several scenes, or even frames, that the total dynamic range of a composition can't change by more that, quite the contrary.
http://www.clarkvision.com/imagedetail/dynamicrange/
http://www.clarkvision.com/imagedetail/dynamicrange2/

Based on that information, we can see that given current image capture technologies, and "ideal" projector, one that could completely reproduce any given recorded image, would have a simultaneous CR capability of around 10 stops or so, or about 1000:1. But it would also need a sequential CR capability to reproduce a wide dynamic range which can be captured and perceived, which at a minimum IMO is around 1,000,000:1 (based on black being the minimum we can perceive and white being at least a "comfortable reading" level).

FWIW, according to the NASA data, the "Absolute Threshold" of sight is 10^-6 millilamberts, or effectively 10^-6 ftL, so that makes the CR of SMTPE reference:absolute threshold 12/10^-6 = 12,000,000:1 .

So my answer is 12million:1 is enough :D

There has been a lot of discussion on the forum about on/off lately and in particular how much is enough. Some forum members want CRT level performance, others feel that 3k:1 is enough.

I've also heard respected members in the industry make comments that, there is an "obsession" with people wanting more contrast from digitals. Even the SMPTE study goup that follows display devices that are replacing CRT's don't seem certain if a specific contrast spec should even be specified and in fact they recently said, "The WG should determine whether in fact specification of contrast ratio is required".

So I'm curious to hear what people think and why.

I chose 250K to one. Oh, and I want it without an iris. :)

Cliff

Quote:
Originally Posted by Mark Petersen View Post

Yes this is true but this thread is only examining on/off and isn't meant to be a comparison between projectors other than just comparing on/off itself.
I know what you are saying.

The point I was making is that CR cannot be meaningful in isolation, nor can any other parameter. I didn't mean it to appear as a comparison, just an example of my point regarding image as a whole.

The question seems to make sense to me....To bring up another car analogy..If someone asked what is the fastest car from 50-100 mph, or what is decent 0-60 time would it be appropriate to give a response? Even though all of the other parameters of a good car were not addressed handling , cornering, braking and lap times for example. The question is just a basic question. I for one would like CRT type on/off contrast. The higher the better. plus lots of lumens to light up a big screen:D What we want and we we are going to get are 2 different things!

Well, I chose 30,000:1 (are we talking native or static :) ) My goal is watching movies and my yardstick is, does the hardware allow a "suspension of disbelief"? With my theater and the RS2 the answer is yes 30,000:1 is enough. Now, if only the the software would catch up, I find myself avoiding perfectly good movies telecined with poor black levels.

On a linear display, the higher the CR, the better.

Infinite would be enough.

Anyway, I voted for 100k:1.

The black level of my RS1 is visible immediately on blackout, but I'd think 100K:1 would make the slight amount of ambient light in my room the limit.

I'm not sure 250k wouldn't be discernibly better, but I suspect I wouldn't be willing to pay the higher price it would command.

Just curious, about what is the measurement limit for a 700-1000 lumen projector? Is it possible to accurately measure a 1,000,000:1 contrast ratio?

Does anyone know the real world contrast ratio of a 'sony xbr800' which I own?

Until any digital 'looks' the same I will never be satisfied. I am spoiled by the crt blacks.

Non dynamic 8000:1 contrast is high enough to consider other factors like (Sharpness/Ansi/Color/Motion/Luminence) as more important when picking my next PJ.

My 'sharp 20000' to me in high contrast mode is good enough. Never accurately measured it so I have no honest number to claim. All I can say is that I compared it to an RS2 which I had for a little over 2 months and the real world viewing impression between the 2 of them was minor at best. If the RS2 rated at 30,000-1 was a small improvement at best over the sharp then I can say that cr values far above what the present projectors can produce is needed to make a substantial difference toward crt black levels which is the holy grail at this point.

Digital has a looooooong way to go unless the technology changes to something different from anything we have today.

Just curious, about what is the measurement limit for a 700-1000 lumen projector? Is it possible to accurately measure a 1,000,000:1 contrast ratio?
With lots of money and/or long exposure in a light-tight room. You can focus the entire output of the pj on a receptor with lenses also. Forget the screen.

Just curious, about what is the measurement limit for a 700-1000 lumen projector? Is it possible to accurately measure a 1,000,000:1 contrast ratio?

A trick for measuring black level on high contrast PJ's is to place a white card just a few inches in front of the lens and measure from that small spot. Or place the sensor a few inches from the PJ pointed directly into the light path.

Enough is when it no longer matters/is talked about/discussed and quarter mil or better should hopefully accomplish this. You could have 12.5ftLs with a black level of 0.00005 or double both if a bright pic is preferred. I think that would completely satisfy the diehards. Dave

I refuse to pick a number because this can not be answered in a vacuum. The answer is it depends. Joe Kane increases screen size or goes gray in order to get acceptable blacks from his Samsung, which I think would be for him .01 ft lamberts or less and he would accept a fairly low on/off number Me. When watching a bright projector, most times the blacks will look fine regardless of how high they might measure, but I would want to run a CRT like gamma of say 2.6 which would I think require 60,000 or so to one to avoid black crush. But FTW (which I think the House of Representatives did today), I want a fade to black screen to be black, not gray. So I want a really high on/off or low black reference level regardless of how bright the brights get. So I would insist on say a million to one if possible. The higher the better. For normal viewing and my preferred gamma, 50,000 would not be enough. 60K might be OK but 80K to 100K OK. The RS20 will not be enough in the on/off department even if its 50,000 at reasonable brightness, say 12 ft lamberts. Most everybody here really doesn't know. They are just spewing forth whatever including me too I guess. Until you've seen the higher stuff, you just can't say no no no, I only need X, anything else would be wasted on my unworthy eyes.

I refuse to pick a number because this can not be answered in a vacuum. The answer is it depends. Joe Kane increases screen size or goes gray in order to get acceptable blacks from his Samsung, which I think would be for him .01 ft lamberts or less and he would accept a fairly low on/off number Me. When watching a bright projector, most times the blacks will look fine regardless of how high they might measure, but I would want to run a CRT like gamma of say 2.6 which would I think require 60,000 or so to one to avoid black crush. But FTW (which I think the House of Representatives did today), I want a fade to black screen to be black, not gray. So I want a really high on/off or low black reference level regardless of how bright the brights get. So I would insist on say a million to one if possible. The higher the better. For normal viewing and my preferred gamma, 50,000 would not be enough. 60K might be OK but 80K to 100K OK. The RS20 will not be enough in the on/off department even if its 50,000 at reasonable brightness, say 12 ft lamberts. Most everybody here really doesn't know. They are just spewing forth whatever including me too I guess. Until you've seen the higher stuff, you just can't say no no no, I only need X, anything else would be wasted on my unworthy eyes.
There are some good points here but don't forget it's all linear. Bottom line: how many ft-famberts do you want at 100 IRE and how many do you want at 0 IRE? Projector brightness, screen size, gray screen and all the rest gets you either to the 100 IRE end-point or the 0 IRE end-point. Then CR gets you the IRE span.

all this talk about black level is making me want to put back up my 9500lc I had it on a 78 inch wide screen.....it was like looking out a window on 1080p Bluray .. but got involved in these size matters debates and got a digital. I just wish the 9500 wasn't so darn heavy or I would use it for myself and use the larger screen & digital for family and friends..


Joey

I voted 250k:1, like the majority. My reason is simple: It's all about marketing. ;)

And of course CM (and others saying more or less the same) is right:

"The point I was making is that CR cannot be meaningful in isolation, nor can any other parameter."

"There are some good points here but don't forget it's all linear."

? Pretty much everything related to subjective human perception requires exponential changes of the parameter in question.

Does anyone know the real world contrast ratio of a 'sony xbr800' which I own?...Most properly calibrated consumer direct view CRTs are <<10,000:1.

Can anyone actually see the difference between 80,000:1 and 100,000:1 without directly comparing the images? Or is it just a numbers thing?

I believe that if folks have seen the look of a well set up CRT and what high on/off does for the image depth (despite poor ANSI and modest light output) they would not want to accept less as a good place to stop.

Art

"There are some good points here but don't forget it's all linear."

? Pretty much everything related to subjective human perception requires exponential changes of the parameter in question.
I should have been more explicit. Human visual response IS logarithmic. But I was talking about the effect of constants like screen size, screen gain, projector brightness on contrast ratio. If you change a constant, both ends of the contrast ratio are changed in a linear fashion and you end up with the same ratio.

In a ideal world, enough to where any further increase would be either imperceptible or impossible. Because if a further increase would be possible and perceptible, you know we'll want it and knowing it's possible will make some people unhappy with whatever projector they currently own. If it gets to a point where a further increase would be imperceptible, then who cares? Who knows what that number is? Also, this assumes all other aspects of picture quality are also at that level.

Who knows what that number is?

I've got a pretty good guess...

I've got a pretty good guess...

Hmmm.....let me know when you've got a REALLY good guess.

Can anyone actually see the difference between 80,000:1 and 100,000:1 without directly comparing the images? Or is it just a numbers thing?

There seems to be very little study done in this area so we don't know what it takes to produce a noticeable difference. It might be that once you get to 80k:1, you need to go to 200k:1 to get a noticeable difference. Again, who knows?

Hmmm.....let me know when you've got a REALLY good guess.

12,000,000:1 based on SMPTE recommendation and NASA info:
http://www.avsforum.com/avs-vb/showpost.php?p=14764571&postcount=23

Can anyone actually see the difference between 80,000:1 and 100,000:1 without directly comparing the images? Or is it just a numbers thing?That would probably be tough. 80k:1 and 1 billion:1, no doubt as long as images where the differences really are get used. But 80k:1 to 100k:1 wouldn't be much difference. I don't consider 1600:1 to 2k:1 or 8k:1 to 10k:1 to be that much difference either (same percentage improvement) and with the higher on/off CR numbers washout from ANSI reduces the difference in many images more.

--Darin

12,000,000:1 based on SMPTE recommendation and NASA info:
http://www.avsforum.com/avs-vb/showpost.php?p=14764571&postcount=23


Now, we're talking! I would also guess that would be enough to produce a noticeable difference, but over 80K:1,
200k:1? 500K:1? Who knows?



.

... I get the impression (perhaps wrong) that most of the industry feels that 10k:1 or so is "good enough"...That is certainly the sense I get from my contacts in post production. It is pretty clear that if one's display has less on/off than a mastering monitor (< 12K:1) one will not have accurate dark scene contrast in lower APL scenes.

It is not clear, what are the visual benefits in viewing current sources/content on a display with significantly higher on/off contrast than a mastering monitor?

12,000,000:1 based on SMPTE recommendation and NASA info:
http://www.avsforum.com/avs-vb/showpost.php?p=14764571&postcount=23If this is a serious answer, since we are discussing display devices maybe we should subtract the fully dark adapted HVS limit from the number since that can take more than 30 minutes. I suspect we will not see any movie with 30 minutes or more of full blackout. :D

That is certainly the sense I get from my contacts in post production. It is pretty clear that if one's display has less on/off than a mastering monitor (< 12K:1) one will not have accurate dark scene contrast in lower APL scenes.

It is not clear, what are the visual benefits in viewing current sources/content on a display with significantly higher on/off contrast than a mastering monitor?

My guess is the production houses are wanting to ensure that sufficient contrast exists so that shadow details are preserved and nothing is crushed. For enthusiasts on the other hand it's more than being able to see the contrast differences it's also about preserving the realism of those contrast differences. As an example, enthusiasts will probably not be happy until projected video black is the same or close to perceived perfect black.

If this is a serious answer, since we are discussing display devices maybe we should subtract the fully dark adapted HVS limit from the number since that can take more than 30 minutes. I suspect we will not see any movie with 30 minutes or more of full blackout. :D

I agree, does anyone know what that limit is? On the other hand though, some horror flicks like "AVP: Requiem" and "Chill Factor" contain extended dark scenes (most of both movies) that while not being blackout would still cause the eye to become dark adapted. So even for those movies we're still probably looking at the darker region of the HVS.

Edit: I should probably add that both movies seem (to me anyway) "overdone" as far as dark content is concerned. I would think that a good cinematographer would try and avoid movies from having this many scenes devoid of bright content and would ensure that at least some bright content is placed in most scenes even if the scene is very dark overall.

I agree, does anyone know what that limit is? On the other hand though, some horror flicks like "AVP: Requiem" and "Chill Factor" contain extended dark scenes (most of both movies) that while not being blackout would still cause the eye to become dark adapted. So even for those movies we're still probably looking at the darker region of the HVS.

Edit: I should probably add that both movies seem (to me anyway) "overdone" as far as dark content is concerned. I would think that a good cinematographer would try and avoid movies from having this many scenes devoid of bright content and would ensure that at least some bright content is placed in most scenes even if the scene is very dark overall.

I totally agree. These brighter elements add contrast to darker scenes and also permit the darker areas to appear blacker due to perceived contrast.

Black in and of itself is pretty uninteresting visually. The interplay of light and shadow is what really sets a mood.

I totally agree. These brighter elements add contrast to darker scenes and also permit the darker areas to appear blacker due to perceived contrast.

Black in and of itself is pretty uninteresting visually. The interplay of light and shadow is what really sets a mood.

Very true, most of the better horror flicks (Alien, Aliens, etc.) have lots of scenes that seem very dark but there is usually something in the scene that is bright enough to create some contrast while also reducing the perception that video black has a little glow to it. It's been awhile since i've watched it (and I have no intentions of sitting through it again ;) ) but the second Starship Troopers movie seemed to have the same problem (extended dark scenes devoid of bright content). None of these movies will set a benchmark as far as production quality is concerned :)

That is certainly the sense I get from my contacts in post production. It is pretty clear that if one's display has less on/off than a mastering monitor (< 12K:1) one will not have accurate dark scene contrast in lower APL scenes.

It is not clear, what are the visual benefits in viewing current sources/content on a display with significantly higher on/off contrast than a mastering monitor?I understand what you are saying and you did say "current", but I think this points to wanting the mastering monitors to be improved in this area as consumer devices are improved in this area. Mastering houses seem to be moving away from just sticking to CRTs and we may see them updating their displays used for mastering more often than in the past. There is the issue with 8 bit sources and banding, but at least with DCI content I think there should be enough bits to go much deeper than 12k:1 without those problems, if the mastering is done right.

I see it kind of like mastering on audio equipment that can not come as close to real life as some audio equipment a person can buy for home use. In that case I would say that the goal would be to improve the audio equipment used to do the mastering, not keep the consumer stuff from being able to get closer to real life. And I see the on/off CR issue kind of like in audio having a low end hiss that is always there. If the mastering equipment always has a low end hiss and the artists are not allowed to actually deliver silence, then that issue should be addressed.
If this is a serious answer, since we are discussing display devices maybe we should subtract the fully dark adapted HVS limit from the number since that can take more than 30 minutes. I suspect we will not see any movie with 30 minutes or more of full blackout. :DI think we need to look at how much it takes to take a person out of that dark adaptation. In my experience a person can have a lot of dark adaptation 30 minutes into watching a movie. Probably not as much as if they went into a dark cave for 30 minutes, but still quite a bit. So, the lower threshold after watching a movie for 30 minutes is relevant, but I don't know how close it is to the threshold on that chart. I know that with a CRT that we got around 700k:1 on/off CR with from some measurements, and a white level around 6-7 ft-lamberts, it only took a few seconds to be able to see the edge of screen near the screen border in the extended blackout in the Abyss chapter of Finding Nemo. That was from watching the part just before the blackout and then letting it go into the blackout.

--Darin

"There is the issue with 8 bit sources and banding, but at least with DCI content I think there should be enough bits to go much deeper than 12k:1 without those problems"

Doesn't bit depth just define the granularity of the range of color/luma?

Won't the display map the source's luma range to its full CR?

If this is a serious answer, since we are discussing display devices maybe we should subtract the fully dark adapted HVS limit from the number since that can take more than 30 minutes. I suspect we will not see any movie with 30 minutes or more of full blackout. :D

I see the smiley, but horray! Serious commentary on my post, last couple times I posted that theory it was greeted with a hoard of crickets chirping.... :cool:

You've got a very valid point. Here's my take, What is "enough"? In the strictest sense, we want the machines "black output" to be imperceptible. Ideally, that shouldn't be contingent on any time limit. We all know people around here, if you can see the screen, even if only after 10, or 20 minutes, people will want more.

In that vein, 12M:1 is "enough".

However, you're absolutely right, we could tackle it as an engineering problem. So, we can go back to the data, "white paper in starlight" is spec'd as 1e-4ftL. But what is that? Well, the RS1 can do say 12,000:1 (for easy math) that makes it's black level at SMPTE ref white 1e-3fL. So "white paper in starlight" is 1/10 the brightness of the RS1's blacks, or equivalent to about 100,000:1.

It usually takes a good bit for our eyes to adjust to starlight, so we could say 100,000:1 is "good". I'd venture that that's about the threshold where many people, especially those without bat caves, would be "satisfied", or at least would be the point where the PJ would not be the limiting factor for them.

That said, I think the goal we should be hoping for is 10x that, or 1M:1, which would be 1/10th "white paper in starlight".

As for 12M:1, I'll agree that while "enough", it is in fact "more than enough", probably unattainable and also unnecessary.

I think we need to look at how much it takes to take a person out of that dark adaptation. In my experience a person can have a lot of dark adaptation 30 minutes into watching a movie. Probably not as much as if they went into a dark cave for 30 minutes, but still quite a bit.

That's a very good point, and one reason why I'd lean to the "conservative" side, ie more CR is required to be "enough". Lets face it, a DC2 DLP with no DI is capable of producing the "illusion" of perfect black on a quick fade. But that's not enough for us because that illusion fades in seconds (if not less).

So, the lower threshold after watching a movie for 30 minutes is relevant, but I don't know how close it is to the threshold on that chart. I know that with a CRT that we got around 700k:1 on/off CR with from some measurements, and a white level around 6-7 ft-lamberts, it only took a few seconds to be able to see the edge of screen near the screen border in the extended blackout in the Abyss chapter of Finding Nemo. That was from watching the part just before the blackout and then letting it go into the blackout.

--Darin

Like I said, if in a cave you can ever see the screen, we here won't be satisfied :D

Of course if we ever see these 1-10M:1 machines, let's hope they think ahead enough to put dynamically dimming LEDs on them so the PJ's LEDs aren't lighting up the screen :eek:

"There is the issue with 8 bit sources and banding, but at least with DCI content I think there should be enough bits to go much deeper than 12k:1 without those problems"

Doesn't bit depth just define the granularity of the range of color/luma?

Won't the display map the source's luma range to its full CR?It should, but now with limited steps that full CR can bring things out. That granularity is important to whether people can see single steps. As an example, there is some stuff in the blackout coffin scene in Kill Bill 2 that may not have been visible on the CRT used for mastering that I think would be visible on a digital display with a million:1 on/off CR, if the black is set to its deepest in that scene. If I raise the brightness setting on my JVC HD1 there I can see big blocks in spots. Dithering could have been used in this area, but doesn't look like it was. I haven't checked the scene with my Panasonic AE1000 to try to get a handle of the actual encodings for the different things in that scene though.

I think CRTs hide some things that show up on digitals for other reasons too, but limited on/off CR combined with low bit depth can be a problem when displaying with higher on/off CR. At least if the source had higher bit depth maybe the encoder would have made the transitions smoother, even if that difference wasn't visible on the CRT.

--Darin

It should, but now with limited steps that full CR can bring things out. That granularity is important to whether people can see single steps. As an example, there is some stuff in the blackout coffin scene in Kill Bill 2 that may not have been visible on the CRT used for mastering that I think would be visible on a digital display with a million:1 on/off CR, if the black is set to its deepest in that scene. If I raise the brightness setting on my JVC HD1 there I can see big blocks in spots. Dithering could have been used in this area, but doesn't look like it was. I haven't checked the scene with my Panasonic AE1000 to try to get a handle of the actual encodings for the different things in that scene though.

I think CRTs hide some things that show up on digitals for other reasons too, but limited on/off CR combined with low bit depth can be a problem when displaying with higher on/off CR. At least if the source had higher bit depth maybe the encoder would have made the transitions smoother, even if that difference wasn't visible on the CRT.

--Darin

Big dog, I agree 100%. Even the Kill Bill scene shows TONS of blocking when I look directly into the lens that does not show up on the screen.

Cliff

So it looks like roughly 50% of those who responded to the poll would like 250K:1 or more on/off.

So here is an associated question: If the RS20 used a DI and kept the same iris apertures as it currently has, the on/off would be somewhere around 300k:1 (50k:1 * 6). Darin had postulated in another thread that given this level of dynamic range that it might be possible to close the iris completely for full black outs without causing noticeable flaws. So this would raise the on/off to infinite. With that in mind do people think that an RS-20 with a DI would be "enough"?

A better question would be how many people would like to see the RS20 with a switchable on/off DI. Buit we know the answer to that, most but not all. No one really knows how much is enough.

Most here can tell when they have enough to drink. How much good sex is enough?

Would you rather see it all native? YES. But how much native is enough" And then bonus time for the DI to prove the native wasn't enough.

A better question would be how many people would like to see the RS20 with a switchable on/off DI. Buit we know the answer to that, most but not all. No one really knows how much is enough.

Most here can tell when they have enough to drink. How much good sex is enough?

Would you rather see it all native? YES. But how much native is enough" And then bonus time for the DI to prove the native wasn't enough.

250K+ to one. :)

15+ foot lamberts at 10ft wide SMX.

Not to compound all of the other features I'd like to see but I actually think that is a realistic expectation of projector manufacturers. They have the technology to make it happen. There is PLENTY of content out there that takes advantage of it. Hell, there is TONS of content.............

Cliff

250K+ to one. :)

15+ foot lamberts at 10ft wide SMX.

Not to compound all of the other features I'd like to see but I actually think that is a realistic expectation of projector manufacturers. They have the technology to make it happen. There is PLENTY of content out there that takes advantage of it. Hell, there is TONS of content.............

Cliff

If the RS10 really will throw 1000 lumens of callibrated light then you've got your 15 foot lamberts... at least for a few hundred hours. I have a 10' SMX also and I'm waiting to see if the RS10 can really do 1000 or I'm waiting for the new 2.0 gain SMX material.

I don't know the answer to this question. That is a major reason why I am thinking of picking up an RS20. I very much prefer the DLP look, and am thinking of the new C3X1080 for a new theater I am planning. Playing with the RS20 will give me a better idea of what the increased CR gives me. If it is significant, I will probably hold out for LED.

I think a total re-evaluation of projector technologies is in order after this fall's crop of pjs are released...projectorcentral.com has measured two new LCDs at ~400:1 and ~450:1 ANSI, which is very close to competing DLP pjs (unless they've upped their game substantially, which hasn't happened in quite some time), I can only hope LCOS is following suit. Also, the Epson is supposed to be better than the other LCDs, if it comes in over 500:1 ANSI it will compete with almost all of the DLPs. Its a pretty big "if"...but I'm just saying we need to abandon old ideals if they are, in fact, outdated and no longer true.

If the RS10 really will throw 1000 lumens of callibrated light then you've got your 15 foot lamberts... at least for a few hundred hours. I have a 10' SMX also and I'm waiting to see if the RS10 can really do 1000 or I'm waiting for the new 2.0 gain SMX material.

Ahhh yes, but does it measure 200K to one on/off? :)

Hopefully I'll have one in the near future to measure the light output on SMX. I'll be more than happy to post the results.

IMO, the RS1 measured 8 foot lamberts to my 10 with a brand new bulb.

Cliff

Hopefully I'll have one in the near future to measure the light output on SMX. I'll be more than happy to post the results.


I'd LOVE to hear.

I don't know what foot lamberts I get, but it's bright enough that photographic flash bulbs and strobe effects in movies make you wince. I guess that doesn't take a whole lot. I also had a guest complain, "Ahhh... too bright!" in a particularly bright scene. But again, it's all relative to what the level was before the bright scene.

Cliffy. I am going to send you an RS20 to play with, not an RS10. But I suspect it will be late December at the earliest. My green tube neck board had a dreaded spot kill failure and I have a small circular burn at green tube center. Not visible except in a flat green field. Parker is rebuilding the neck boards and has come up with still more improvements. Gotta think hard about ever retubing my err light cannon.

Cliffy. I am going to send you an RS20 to play with, not an RS10. But I suspect it will be late December at the earliest. My green tube neck board had a dreaded spot kill failure and I have a small circular burn at green tube center. Not visible except in a flat green field. Parker is rebuilding the neck boards and has come up with still more improvements. Gotta think hard about ever retubing my err light cannon.
Well, then... can you send that RS10 to ME??

Why is it so quiet in here?


When are you guys gonna get your arse's over here to take measurements!!!!!????? ;)

Cliffy

At Cedia, one thing that was consistently repeated at projector demo's from various manufacturers was the fact that if you want to fully realize the on off and AnSI contrast levels provided by current projectors at current on off/ANSI levels, you need a totally blacked out theater. Walls, ceiling, carpet.

That tells me, from theaters I have seen on this site and others, that most theaters are not set up to derive any benefit from further contrast improvements in projectors.

That tells me, from theaters I have seen on this site and others, that most theaters are not set up to derive any benefit from further contrast improvements in projectors.

If you can make shadow puppets on a black screen in your theather, you would benefit from a projector with higher contrast. So far I haven't seen any projector were I couldn't produce shadow puppets, not even the RS2, not even close...

When it comes to on/off specifically, you would benefit from improved contrast even in an all white room, as long as you don't have any ambient light present.

I too am sometimes amazed at the lengths a lot of people have gone to, even in creating a dedicated theater, only to make decor choices that I would think obviously compromise image quality. Including a great number of professionally designed theaters as well!

Drexler,

Main point being that even in a totally blackened, light controlled room, the light that comes off the screen reflects off walls, ceiling and floor.

I truly doubt what you are saying, regarding an all white room. I think the manufacturers were pointing out that, particularly in an all white room, contrast levels that current PJ's obtain would be wasted.

When it comes to on/off specifically, you would benefit from improved contrast even in an all white room, as long as you don't have any ambient light present.


Exactly !

Art

Drexler,

Main point being that even in a totally blackened, light controlled room, the light that comes off the screen reflects off walls, ceiling and floor.

I truly doubt what you are saying, regarding an all white room. I think the manufacturers were pointing out that, particularly in an all white room, contrast levels that current PJ's obtain would be wasted.

You should not doubt it for on/off.
ANSI, yes... a white room will compromise that.

Drexler,

Main point being that even in a totally blackened, light controlled room, the light that comes off the screen reflects off walls, ceiling and floor.

I truly doubt what you are saying, regarding an all white room. I think the manufacturers were pointing out that, particularly in an all white room, contrast levels that current PJ's obtain would be wasted.

In dark scenes and fade to blacks, there is little or no light to be reflected on the walls. Ergo, the color of the room will have very little or no effect on the image. High on/off contrast projectors gives the image deep blacks in dark scenes that you will benefit from even in an all white room.

In brighter scenes however, the room will have a huge impact. ANSI contrast is a measure of intrascene contrast in very bright material and is thus severely compromised in a white room.

Adding to what others have said, the reason that a white room doesn't hurt the on/off CR is just basic math. If the reflections were adding even 20% then the on/off CR would be the same because 1.2x/1.2y is the same as x/y. ANSI CR gets hurt and so that can bring down where the crossover images are where the washout best described by ANSI CR dominates the absolute black floor best described by on/off CR, but you can still benefit from higher on/off CR for blackouts and very dark scenes. Even with something like a star field, the number of pixels that actually are on to display stars can be a very small percentage (so not a whole lot of light to bounce around the room).

Also, there are screens that can help ANSI CR to viewers in viewing positions in light colored rooms. Both gray screens and screens with directionality (like the new Supernova that rejects light from above and below very well).

--Darin

So it looks like roughly 50% of those who responded to the poll would like 250K:1 or more on/off.

So here is an associated question: If the RS20 used a DI and kept the same iris apertures as it currently has, the on/off would be somewhere around 300k:1 (50k:1 * 6). Darin had postulated in another thread that given this level of dynamic range that it might be possible to close the iris completely for full black outs without causing noticeable flaws. So this would raise the on/off to infinite. With that in mind do people think that an RS-20 with a DI would be "enough"?

No.

I think it by far would be enough for the vast majority (95%+) of people, and even the vast majority of viewing tasks. However, it isn't a high-dynamic range display either. There are always improvements to be made, and I am sure they will be made.

Frankly, I think that 30,000:1 is more than enough for the VAST majority of people. But I still voted for the maximum because there are always improvements to be had for those chasing the best, and chasing what's possible. 250:000:1 doesn't come close to what we see in just our regular day to day lives. So as much as 10,000:1 or 30,000:1 is probably more than enough, and could even satisfy (but never fully if pressed) me, there is always room for improvement, and always tasks and viewers somewhere who will appreciate that capability.

So no, specifically, while such a projector could be enough for perhaps the pickiest CR fanatics right now, that may change if you wanted to do HDR for instance, in which case it wouldn't be nearly enough.

Is 1080p enough? Sure, maybe it is for DVD viewing. But is it enough for 1080p content, or 4k content? Does it match IMAX? Does it match real life? I think there's always going to be a place for more. And since the folks here are not average in their knowledge or wants, I don't think that would be enough for most people, even if that were done NATIVELY without the DI, which I think could be objectionable.

That all being said, if you asked is 30,000:1 enough in this economy right now, I'd say: yes. goodness yes it's more than enough! :(