Projector calibrating and tuning with colour filters? Cine4home :)

You can use SMART III to do this (it comes D65 calaibrated by having it put against a D65 light source and the values of what Steve measures, you then put it into the spreadsheet). The meter that is shipped with the package read light levels all the way down the IRE ranges. Also Colorfacts with its light meter uses the USB port and by using its software, it automatically gives out the readings without having to input data manually. As with SMART III, its more time consuming as you have to actually place either the RED, GREEN or BLUE filter on the meter to get measurements for the different colour and takes a little longer to do and you have to manually input the readings from the light meter into the spreadsheet on the pc.


Yes, once you install a filter, you will then have top adjust as you are now putting a filter on the lens to compensate for that limiting colour. So when you take a new measurement, you will then have to adjust RGB levels to get the colour balance back to D65.


You can simply adjust the RGB levels to obtain the best contrast but with a filter you can obtain and even higher figure. You can without a filter try and get the max on/off contrast buit that will then make your colour balance off the intended mark your after for D65. A filter compensates for the limiting colour.


You can max contrast by using different modes like the one you stated but you will be again, like i said, missing the aim of trying to get closest to D65 reference. The whole point is trying to increase contrast while keeping it at D65 like Cine4Home is doing. You will have toi change the gamma response for that particular colour by adjusting thr gammar of that colour. Maybe you wont be able to do that using the projectors adjustments but programs like powerstrip can do that for you using a PC.


All Cine4Home is doing is using the proffessional Colorfacts light meter and software to test the projectors and by using filters they are obtaining higher contrasts with them while keeping the optimum colour balance throughout the IRE ranges.

What I am considering, after determining the outer lens-ring's size, is buying some quality color filters and some diachroic neutral density filters and then renting something like Colorfacts to calibrate appropriately.


I recognize the unit-to-unit differences in both the lamps and the set of LCD panels to which deviations in what may be needed from one unit to the next might present individual scenarios for calibration.


But, considering the price of something like the PT-AT700U is 'half' the price of what I paid for an uncalibrated 65" widescreen analog RPTV four years ago, I can see that $300 in rental and maybe $200 in filters can be justified ESPECIALLY considering what some are selling this projector for. And the same example applies to the VPL-HS51, too.


My biggest concerns in the post-purchase activity would be filter-lens attachment with available color-correcting & ND filters, and the quality of the detector to measure both color and brightness (at the lowest levels) for the best performance I can achieve via DIY.


I am particularly interested in how Eckardt(sp) chose to go the route of using Video mode instead of Cinema mode in order to start off with maximum lumens as he was adding color-correcting filters to the mix, where due to a lower-lumen condition of being in Cinema-mode one would tend to not offend the lumen gods by adding CC filters to a dim mode.


I suppose you could try both paths and save in user profiles, no?

Good feedback. What you're saying basically reinforces what I was thinking.

Using the Panasonic projectors as an example. I think the process works something like this:


1) Test the projector for contrast using no colour filters. Try all the different "modes" such as NORMAL, DYNAMIC, CINEMA 1, CINEMA 2, NATURAL, MUSIC etc. These modes are actually different gamma curves, correct?


2) Once you find the mode that gives you the highest contrast, try to work with that mode first.


3) Measure the colour balance and gamma curve.


4) Use colour filters to restore the colour balance back as close as possible to D65.


5) Since the colour balance will not likely be perfect from the application of a colour filter alone, use the individual RGB adjustments and fine tune the colour closer to D65.


6) Measure the gamma curve to see if it is correct after the filter and RGB adjustments. If not, then try another mode.


7) Try other modes which give higher than "stock" contrast and see what combination of contrast ratio, D65 colour balance and gamma curve results from your best effort to chose filters and adjust RGB.


8) Chose the best combination of mode, colour filter and tweaking which gives you the highest contrast ratio, closest to D65 colour balance and best gamma curve.


Question:


Irrespective of any tweaking, what is stopping a projector owner from turning up all the colours equally? Does this help or does it mess up D65 or the gamma curve?


This is what I think is being done. The reasons it works are many, but I think it has to do with considering which colour is the weakest, maxing out the colours while using filters to maintain as close as possible to D65 and also maintaining a proper gamma curve. Does this make sense? Are the modes for the most part simply different gamma curves?


This is the way I'm imaging that it works.


Ekkehart,


Can you give some feedback here?


thanx in advance

Regarding The Ferret's question, firstly in some modes you mighnt be able to achieve proper D65 if the 9500k temperature is way off the D65 mark, as that is stating already a cool temperature and at that temperature, it might be hard to adjust the gain and bias to get it set at D65. You can save the calibrated profile into the profile of the colour temperature selected.


What Horst has posted is a good idea and a very cost effctive way of obtaining a good colour balance throughout the greyscale BUT, like the guy stated in his article, your only actually tracking the greyscale using the colour temperature you have selected in the projector. So your not actually calibrating to D65, unless you had a reference, but infact calibrating the colour temperature your using as selected which means it mighnt be at D65.


Also cpc. All different settings in the projector will have slightly different settings for colour bias and so on. Usually, but not all the time, but the 6500k temperature is the best to start at. It all depends on the projector and how it was calibrated at the factory though, so maybe the cooler tempertaure of 7500k could represent it better at D65, but this is where you'll need a reference if you want it set as close to D65 as possible. If you dont have a reference, just use whatever temperature you think looks best to you and calibrate the greyscale until across the IRE is as flat as possible for the 3 colours. That will then help having a good greyscale with the temperature you have it at without a red push or whatever in the high IRE ranges for example.


You always take a measurement from where you have it set at ie. what you thinks best if you dont have a reference, and then by looking at graphs where you have entered the data, you can see what has to be changed and adjusted. For SMART III, it tells you what it thinks needs to be changed to get the best possible result but once you get used to it, you know what has to be changed by looking at the graphs. So once you have completed your first run, you then adjust the bias and gain of the projector to obtain the best colour balance and also what is needed to adjust gamma and so on, theres no real specific way in how you do it, but its best to make basic adjustments first and then take results and work from there.


You cannot think by adjusting all colours to equal amounts will make it at D65 like The Ferret said. Because the Blue is usually the limiting colour, but not all the time (but is always pushed the hardest), but its the most hardest of colours to see and usually has to be driven more while the green usually has a lower rating than the blue, so no, it doesnt mean you'll get D65, infact you wont at all if you make them all equal. You have to make all colours track well to create a grey for each IRE pattern across the greyscale.


Basically your trying to get the most light output by using the gain and the lowest output (bias) without losing detail down the bottom of the IRE range all at the same time keeping the colour balance tracking good and increasing the contrast ratio with filters helping you achieve that and also adjusting gamma to a curve thats at usually 2.2 for video material, otherwise having it too high can make it all wash out and vice versa.


You can do with a DIY kit but unless you have a reference, you wont achieve D65 properly. With something like SMART III, values are given to you which are referenced beside a D65 reference lamp so you can adjust to the proper reference and also obtain a good greyscale. I definately recommend it if your on a budget.

Tom, while I would always like a situation in which a product meets its marketing, I cannot remember an instance in which that has happened. The difference between measure out of the box and what is on the marketing sheet typically are in two different ballparks: one reality and the other a figment of some questionable imagination.


For instance, let's take the new Panasonic PT-AE700U. Marketing says 2000:1 contrast. I say, virtual contrast as you cannot hold a brightness and get that kind of CR from this product. Its actually modulating the light on the entire panel and thus extended the CR range.


BUT, not even this is something to be taking seriously. For instance, the out of the box was actually measured at 950:1. Heck, that is lower than the marketed CR for the model is replaces (PT-AE500U?) without the fancy lamp-modulation technology.


So, as a consumer I take pause. Calibration is calibration and its an activity one performs without criteria unless it enters into warranty violations (i.e. opening up the chassis). Now, these product manufacturers can go about disclaiming and caveatting their marketing, but I hardly see that happening.


I react to inaccuracies in marketing much like I do with Spam: its just not desired. But, I would think federal protections to some extent can protect us to insure that a reasonable delivery of marketed offerings is accomplished. A 27% deviation should not be easily waved aside as 'acceptable', and it took some endeavor to actually get to the marketed value, an activity that would surely become costly if done professionally.


Curious, did anyone actually measure the Full On/Off CR of a new HT1000 and compare it to the marketed value? And its because I so distrust marketing and marketers that I know if anything read by the manufacturer is desire it must be calibrated to great extent in order to achieve the stated performance.