[deronmoped]

I have noticed that contrast ratios for different PJ's can be rated differently. Now I can understand how LC can effect the contrast ratio, making it better. What are the other ways the contrast ratio can be different for a given PJ?


Are there any ways of improving them?


Thanks, Deron.

 

[darinp2]

Are you talking about CRT vs CRT here? Or are you including other PJs?


--Darin

 

[WanMan]

I think he might mean Full On/Off vs. ANSI, but I could be wrong.

 

[darinp2]

At the risk of having someone tell me I'm an idiot because I think (or understand) that ANSI CR does not do a good job by itself of describing what you should expect for black performance of a projector across the range of images you will see, I will repeat something I posted elsewhere.


In a perfect room there are really 2 contributors to black level. One of them is from the brighter parts of the image and is best described by the ANSI CR. The other is light leakage that creates a floor of "black" that the projector cannot go below and is best described by on/off. At any point one of the 2 can dominate the CR of the particular image, depending on how bright or dark it is and the particular makeup. Also, in a real room the reflections from the brighter parts of the image effect the ANSI CR. You really need to have a good idea of the performance of both types of CR to get an idea of how a particular projector will perform across different images (and then the 2 measurements still don't perfectly describe this), but CRTs have enough on/off CR that it is mostly ignored and ANSI CR is discussed, while at least DLPs among the digitals have enough ANSI CR that it is mostly ignored and on/off is discussed.


I wonder at what point reviewers and other people in the industry will understand this. There are some that do, but it sure doesn't look like enough to me.


One thing that is confusing is that people shorten both ANSI CR and on/off CR to "CR" when they really are not the same thing.


Also, what we care about around here are these numbers after calibration, but at least there seems to be some correlation between specs and real world numbers.


--Darin

 

[chrisautosound]

perhaps this helps, as an audio tech/writer I can explain by comparing audio and video terms and giving examples.


Most systems are non linear and require two measurements.


Signal to noise ratio = ANSI checkerboard contrast

simply put, with a signal, usually the maximum undistorted signal, how far down is the noise?


dynamic range = on/off contrast

what is the ratio of the largest undistorted signal to the noise with no signal?


A good example is VHS HI-FI. The noise was always about 35 dB below the sound, s/n=35

But the system could reproduce really loud and really quiet sounds for a dymanic range of >80 dB


both numbers are needed to descibe how audible hum and hiss were present even though the system cound reproduce very loud and very quiet sounds.


another example is a CD player. The zero bit track is no signal and is just hiss/noise maybe -110 below maximum output. But the 16 bit data on the disc limits the smallest signal to -95 dB.


Even the human eye works this way. We can see by starlight or broad daylight, a dynamic range of over 100 billion. but you can't read the label of a 100 watt light bulb.

 

[darinp2]

Quote:

Originally posted by chrisautosound But marketing will hype only the numbers that look good.

Thanks for the example.


While normally I would agree with the marketing comment it seems that in this case it isn't working out that way. I've done some measurements on my Sharp 11k which indicate an ANSI CR of higher than 700:1 and I've heard that TI has a minimum threshold for ANSI CR for anybody to build projectors based on their chips right now that is higher than the 320:1 that was mentioned as a very high ANSI CR for a CRT here recently. Yet, I don't hear much mention of ANSI CR for DLPs. While I think the ANSI CRs are high enough on DLPs that they don't need much improvement I do find it interesting that TI and some of the DLP manufacturers do not publicize these numbers more. Maybe they want to talk about the number that is higher (on/off), but that doesn't seem to apply to the way that CRT's specs are usually discussed.


On CRTs the on/off CRs are great, but I think people were used to talking about ANSI CRs because that is where it made sense to try to make improvements and where CRTs differentiated themselves from each other. I could be wrong here, but it seems that maybe more CRTs started getting speced with on/off CRs after digitals came along. Before that I'm not sure it made much sense to talk about on/off CRs for CRTs because they were all so good at it. If whatever CRT manufacturers are left wanted to quote the numbers that would sound the best for them, it would seem that they would quote on/off CR. Maybe that would make it difficult for them to differentiate between their high end and lower end models, though. It seems that ANSI CRs are more prominent for CRTs than on/off CRs, which wouldn't seem to follow the marketing hype theory. I could be wrong, though.


--Darin

 

[deronmoped]

Wow ask a question and you can get a real education here. Thanks guys.


What I was wondering about was CRT PJ's. If you check out the PJ specifications on CR for say a NEC XG 1350 LC they say the CR is 20,000 to 1. Now you compare that to a older 10PG which is also LC they say it has a CR of 17,000 to one. I guessing this is because the newer XG has a greater on/off range, better electronics. Does anyone know for sure?


I have done a couple of things to improve CR that are on the input end of things. Is there any modifications or just better components that will improve the CR on a 10PG or for that matter any CRT?


Thanks, Deron.

 

[WanMan]

Acually, its more likely to be because of the 'LC', which is a liquid-coupling between the CRT face and the C-element.

 

[Chuchuf]

You are both correct Wanman and Deron.

What effects CR are the max about of Lumens the PJ will put out divided by the lack of light Lumens the PJ will put out. This is all really theoretical as you really cannot measure the black side so as it goes to 0 ( it's really some very small number like .0000001fL), the CR approaches infinity.

A real world test is the CR measurement with a checkerboard pattern (non all on/all off). In this case, a CRT with an LC lens will measure a higher contrast ratio simply because there is less light scatter from the bright checkers and they will wash out or illuminate the black squares much less than a nonLC machine. Therefore the higher the RD under this test the better the blacks will be.

On a picture you will notice that the black scenes that have small amounts of light in them will stay blacker on an LC projector than a nonLC. This is one of the reasons the NEC XGLC'c command a much higher price than the nonLC's. It's really much more than the halo around the white letters that are on a black background with the credits at the end of a movie that folks can see the difference.


Terry

 

[deronmoped]

Terry


Both the 10PG and the 1350 LC (or is it a 1352) "are LC", but the 10PG is rated at 17,000 and the XG is 20,000. So LC is common and will not make the differance in the CR rating of these PJ's.


What I'm think the differance is, the XG being a newer PJ with better electronics can shut the output down closer to zero. The 10PG putting out a all black image is not as dark. Could this be true or are they both the same in the way of CR and it's just a sales ploy to market the newer PJ as being better then the older model?


One other thing will masking the face of the CRT help much with CR?


Thanks, Deron.

 

[darinp2]

Quote:

Originally posted by deronmoped If you check out the PJ specifications on CR for say a NEC XG 1350 LC they say the CR is 20,000 to 1. Now you compare that to a older 10PG which is also LC they say it has a CR of 17,000 to one.

Just wanted to say that these look like very minor differences to me. If both projectors had whites of 10 ft-lamberts the respective black levels for the on/off test would be 0.00050 ft-lamberts and 0.00059 ft-lamberts. I would say that once you reach levels of 17k:1 you should really look at the ANSI CR for the differentiation between the projectors. With either of these projectors you could probably set them up to be totally satisified with the black levels in completely black frames.


Even measuring the difference between 17k:1 and 20k:1 reliably would be somewhat difficult and I doubt people here could tell you which black level was which without a side-by-side. And even then it might be difficult.


--Darin

 

[HoustonHoyaFan]

The major reason that the industry created the ANSI standard CR benchmark was the concensus that On/Off CR had become a MIP !( meaningless indicator of performance ) There was certainly wide agreement that it told you nothing about the video performance of a pj.


The other concern about On/Off is the potential for cheating, because the benchmark is so simple.The fear was that manufacturers would start to put in 0Ire timing/modulation circutry which would effectively yield higher numbers on the tests but would have no concrete real world advantage. Such cheating was rampant in computer CPU benchmark in the late '80s. I believe with the high focus on On/Off CR in the digital world wou will start to see similar manipulations!

 

[darinp2]

Quote:

Originally posted by HoustonHoyaFan The major reason that the industry created the ANSI standard CR benchmark was the concensus that On/Off CR had become a MIP !( meaningless indicator of performance ) There was certainly wide agreement that it told you nothing about the video performance of a pj.

I agree with the cheating part and that is why I have mentioned true on/off (after calibration) as what we care about. Fortunately, there are enough people measuring these things and publishing them that the manufacturers have to consider whether it makes sense to claim something that will get panned. I have found pretty good correlation between specs and actual on/off CR with recent models. That is, actual on/off after calibration is always lower than the specs, but usually falls within an expectable range.


When almost every display had acceptable on/off (before HT digitals became popular) there probably wasn't much point in discussing on/off, since they were all good enough. However, anybody who doesn't think that on/off CR has anything to do with video performance doesn't understanding the physics. It is just a fact that a low on/off CR will keep your "blacks" looking gray in dark scenes, even if you have a good ANSI.


As I stated, there are two main factors to the black level and ANSI CR does not give you a good indication of performance in dark scenes.


Otherwise, please explain why a projector (Sharp 12k) that has higher ANSI CR than probably any CRT would still be considered to be lacking CR by most here if they could see both in a dark room.


Asked another way. Would a display with 320:1 ANSI CR and 320:1 on/off look any worse than one with 320:1 ANSI CR and 20k:1 on/off?


--Darin

 

[Chuchuf]

Ah yes.....seeing is believing....


Terry

 

[HoustonHoyaFan]

Darin

My comments on ANSI CR vs On/Off CR relates to CRTs, and was in response to the original poster's question about CRT On/Off CR. I agree that current d-pjs have issues with black levels, and the Off measurement is important in evaluating them. I am discounting filter modifications. etc.


The manufacturers will not cheat by lying about the specs, which they already do! They will cheat by including special hardware/software, which looks for a "long†duration 0ire signal, which tells them the unit is been tested, and lowering light output. This will give great On/Off CR numbers but will not have any real world benefits!


Both ANSI CR and ANSI Lumens were designed to stress test inherent problems in CRT technologies. Those benchmarks do not stress the d-pjs in the same way. IMO they are not directly comparable. 700:1 for the 12K is very impressive given that WSR just measured a Vision 20 Matterhorn at only 150:1.


I don’t believe based on just CR numbers alone I could tell you which pj looked better. I had the opportunity to A/B compare a Sharp 10k to my Sony D50 this past weekend. The black levels were not very different, by eyeball, because I had the black level (brightness) cranked up on the D50. The D50 had a better picture, but was only apparent on critical A to B comparisons. My only issue with the sharp was some kind of flickering noise, particularly around bright moving objects (HD football telecast) my eyes kept trying to focus on the flicker instead of the image!

 

[darinp2]

Quote:

Originally posted by HoustonHoyaFan Darin My comments on ANSI CR vs On/Off CR relates to CRTs, and was in response to the original poster's question about CRT On/Off CR.

I agree with your comments with respect to CRTs. I would have answered the question differently if it was just CRT vs CRT, but I wasn't sure and that is why I originally asked whether it included other projectors. When WanMan mentioned on/off vs ANSI I figured maybe it had to do with digitals mostly talking about on/off and CRTs mostly talking about ANSI.
Quote:

The manufacturers will not cheat by lying about the specs, which they already do! They will cheat by including special hardware/software, which looks for a "long†duration 0ire signal, which tells them the unit is been tested, and lowering light output. This will give great On/Off CR numbers but will not have any real world benefits!

Can you name any projector that does anything like this? I have tested quite a few digitals and I have yet to find one that does anything like this. The only cheating I know about was what Sharp reportedly did a while back, which was to change the projector settings between the measurement for "on" and the measurement for "off". They were panned on the forums for it, from what I understand. They are now claiming 5500:1 on the 11k/12k and I've measured 4200:1 at very close to a temperature of 6500. I consider that very good compliance with the specs. I still want higher on/off CR, though.
Quote:

I don’t believe based on just CR numbers alone I could tell you which pj looked better.

On that I completely agree. There is no one measurement that can tell you how good the images will look and even if you knew all the measurements the proof would still be in the viewing.


--Darin

 

[WanMan]

When Sam Runco & Company decided they were going to come out with a new standard to measure brightness across their product range they chose to use the ft-L method. I admit I blasted them for doing so, because their purported 14 ft-L claim is something completely lost to the average consumer, and still many videophiles. What I was also surprised at was that while this new [marketing] method was for brightness, they didn't decide to come up with a means for the other side of the coin.


Why then not also use ft-L (ft-mL?) for a measure of the minimum amount of light one will produce using a fixed-setup rule, which is exactly what they are doing for their brightness performance. Curious, as this would ultimately allow one produce (vs. another product) to show what performance it has against its ability to not produce (or block/hide) light from passing through the optical block.


The new term could be called ft-L contrast.

 

[darinp2]

Quote:

Originally posted by WanMan Why then not also use ft-L (ft-mL?) for a measure of the minimum amount of light one will produce using a fixed-setup rule, which is exactly what they are doing for their brightness performance.

I talked to a couple of Runco employees for quite a while at their CES party about this. Very nice guys BTW. We had a nice discussion about the relevance of each measurement. I think their main concern with doing this at this time is the opportunity for others to cheat. I've heard that it has been quite a while since the SMPTE specs have been updated and even though the on/off test is one of the most straightforward it might be nice to get some official rules to keep people from cheating. That is, even though I haven't seen any evidence of cheating in the last couple of years.


I would actually like to see 3 measurements. A 100 IRE ANSI CR (checkerboard) as now, an on/off CR, and a 10 IRE ANSI CR (checkerboard) at some gamma (2.2 or 2.5). The last one would be exactly like the current ANSI CR test except that the brightest squares would be 10 IRE. Or it could be 20 IRE (about 2% of the brightness of 100 IRE). I think that one could do a pretty good job of describing some things about the darkest scenes that still have non-zero pixels. For this "10 IRE ANSI CR" it might make sense for them to quote the measurement numbers compared to the normal ANSI CR to make sure that they haven't just blown out the 10 IRE level (low gamma for high rise near below 10 IRE) to get better specs. I would suggest a 1% ANSI CR kind of test, but then making test images would be the difficult part.


--Darin

 

[WanMan]

My suggestion was born from a need to be independent of the projection device in question. A flat, neutral-gain screen does not care about what technology is employed, but just cares about the photons striking its surface. Similarly, I'd think a tool that is independent (as opposed to dependent) would do more justice to consumer confidence in what they read (in terms of marketed product specifications).


I don't think I would be the first consumer to agree with the axiom that its better to have honest & independently-measured specification than actually reporting technology-specific limitations, which can be discussed separately on a product specifications.


Current market-deployed product specifications are far from the most beneficial position for a product or consumer to be in. Liberties are taking in the reporting by a manufacturer in their product's specification and this is played against consumer ignorance as there isn't an independent rule for all to follow. As a result, confusion sets in, misinformation and misinterpretation arises and only to the benefit of the manufacturer--never the consumer.


I'd be much more confident in a manufacturer reporting a 1 ft-L minimum light output for a given product knowing its honesty then what is currently reported that can be orders of magnitude off from the marketed value(s). I guess you can tell I am a consumer and not a business man.

 

[darinp2]

Quote:

Originally posted by WanMan I'd be much more confident in a manufacturer reporting a 1 ft-L minimum light output for a given product knowing its honesty then what is currently reported that can be orders of magnitude off from the marketed value(s). I guess you can tell I am a consumer and not a business man.

As long as that 1 ft-l is also given with the 100 IRE ft-lamberts at that setting and people understand that they can turn the bulb down, close the iris, add a neutral density filter, get a darker screen, etc. That is one thing I like about brightness combined with CR. It gives me an idea of where I can move the range by making other choices to combine with the projector.


I know I am repeating myself, but with all the concern here about cheating I haven't seen any evidence of any in the last couple of years. The 11k (12k in the US) is speced at 5500:1 and I've measured 4200:1 on mine with very balanced colors. I think that is very good compliance. I could choose a higher color temp and get closer to their specs.


--Darin