AVS HD 709 - Blu-ray & MP4 Calibration - Page 16

It depends on whether or not his HD-DVD/BR player can output the HD (Rec. 709) signal from its component outputs (not all players can). The HDMI connection is preferred, of course, since it doesn't involve a conversion from digital to analog (with potential artifacts).

It took a couple hours and a little trial and error, but I just finished my first calibration with HCFR and the AVS HD-DVD test patterns. The software is fairly intuitive and easy to use. I was impressed. The test patterns were also great and I applaud the simple and easy to navigate menu structure. It is better than both Avia and DVE. Anyway, I went through three iterations and made some nice improvements in RGB levels and gray scale tracking, but I am a little puzzled by the final color brightness settings.

I used Tom Huffman's recommendation and set 75% Red Brightness at 21% of the 75% White reading, which turned out to be Y=14.0 for Red. In the end, this produced a "Color" setting on my Sony 55A3000 that was 8 clicks lower (from 53 to 45) than I had previously set with the DVE color bars and a blue filter. After the HCFR calibration, I went back to the DVE disc (HD-DVD version) and confirmed again that the color setting would be 53 using the blue filter. Where can this difference be coming from ? Is Tom's 21% value not applicable for all television types ? Or is my i1 Display LT colormeter perhaps defective ? Or is my technique defective ?

The odd thing is, I am getting very good color temperature and RGB levels from 30% gray on up... based on the output data from HCFR. The bottom end needs some more work but from 30% and up, Delta E is 1.1% or less. Color temperature is at about 6250k at 10% and 20% gray, but dead on 6500k thereafter. My individual color points from the CIE chart are as follows:
R: .626, .333
G: .291, .598
B: .148, .065
C: .220, .335
M: .321, .162
Y: .421, .495

I was also happy and surprised with the contrast setting. I set contrast to produce 35 fL (again per Tom Huffman's guide) and lo and behold it was the same contrast setting I had been using. I guess I have a pretty good eye.

The only real disappointment is Gamma, but I am not yet sure how to improve it. It is consistent but only 2.07-2.08.

My biggest concern at the moment is the color brightness. Any ideas or trouble-shooting suggestions are appreciated.

The colour filters supplied with DVE/AVIA/THX glasses are not all the same and there can be variability in them.

When I use my Avia filter it suggests a 4-5 click higher setting than what I read with my DisplayLT.

Since there have been several comparisons of the LT against high end probes and it has proven to be fairly accurate, I just trust the readings. The fact that the picture also looks good confirms the reading.

Semi OT here: I ask here because it looks like we have some pattern experts commenting.

I keep im my mind the dream to have a calibration pattern suitable for satellite calibration.

My provider (Sky Italy) has one, but the experts think of it as an unusable one; this:



Considering that Murdoch will never waist more than one channel for test pattern usage, we could make a proposale to use only one but more usable pattern.

Getting into the pattern discussion, I think what the experts complained was:

1) With only one pattern, you need to move the probe on the screen continuosly, and this can bring wrong measurements

2) The current pattern misses 10 points grayscale.


Having said so, would like to share with yous and look over to have an opportunity to have one usable single pattern.

What do you think? Could we think about proposing a better pattern? Usable? How to deal with plasma calibration with such a pattern (I ask because it looks like plasmas need small windowed patterns..... there may be a trick?)

The problem with using that pattern for a plasma is that the APL is still probably to high to do proper intensity readings.

I suspect that there might be a fair amount of leakage from the adjacent colours into the sensor as well.

Thanks... it became clear to me this morning that I was seeing the color decoder error and I came to the same conclusion in terms of checking all colors, not just Red. Are you essentially saying I should choose a Color setting that minimizes brightness (Y) error in all three primaries, as opposed to being perfect with one primary and potentially far off on the other two ? I also remember reading that color brightness error is less noticable in one or two of the primaries, but I don't recall which. I'm guessing it is most noticable with blue and that is why blue filters are used per DVE and Avia instructions.

I was using the REC 709 setting in HCFR, but I may need to double-check other settings. For instance, should the sensor be configured for an LCD display ? The only other options were CRT or plasma.

Where do I find the Delta Luma graph ? I don't see it in the toolbar or in the tabs at the bottom of the HCFR menu. Is this the "Saturation - Luminance" graph ? If so, it looks like I just have the reference lines for RGBCMY. I do have a curve on the "Luminance" graph and I am about 2-3% higher than the reference from 20% to 80% gray.

Lastly, is my Gamma of 2.07 about the best I can expect with the A3000 without making adjustments in the service menu ? I was hoping for 2.20 but I don't think there is much I can do to change this.

And thanks for the tip about adding RGB Y to get Gray Y. I guess that should have been apparent, but the light bulb had not yet clicked on.

That's the point: assuming you can use only one pattern, how should it be designed? Can you makes some suggestions? Who knows, we could ask them to replace current patter with a better one...

I have noticed the same thing Dave. I'm not quite sure what causes it. Those patterns are hard to encode and also hard to deinterlace correctly. I think we are dumping that pattern in the next release. They cause havoc with MPEG encoders.

I noticed the pinkish hue on my A3000 as well, though it was fairly uniform. I could still make out all the lines in each pattern, but I needed to get within a foot or two of the screen to distinguish each line.

If I had to do it all with one pattern, then what you posted is a decent start. They need to add 10% gray boxes for grayscale.

Make sure you keep a box for each primary and secondary, and the geometry stuff isn't a bad idea. Make sure you have a white/gray box that corresponds to the primary boxes so you can calculate color decoding correctly... in other words, if your primaries are 75% then the gray box should be 75% too (I would use 75%).

Also keep in mind that you have different sized TVs and your boxes must be large enough to fit the sensor in for smaller TVs. I don't know that I wouldn't use square boxes in order to optimize the space you have available. Tinker with the layout to get your measurement boxes as large as possible to make sure sensor has plenty of room and minimize stray light leakage into the sensor. Also, try to keep the APL to a reasonable level.

It is important to realize the limits of using one pattern. For instance, this pattern wouldn't work well for RPTVs (as alluringreality stated) or any TV that doesn't have really good uniformity. As long as you understand the limits, then you may be able to get decent results from such a pattern.

Thanks for the spreadsheet. I will put it to good use. I like how you have color-coded everything. Very nice !!!

I feel dumb because I am still not seeing the "Delta Luma" values. Is there something I need to activate in preferences, or is it called out by a different name ? Is "Delta Luma" the same as "percent separation from luminance target" as called out in your spreadsheet ? Perhaps I simply did not take the right kind of readings to get this information. I ran grayscale in 10% increments, and then primaries and secondaries at 75%.

I was thinking about Tom Huffman's guide and I realized that his guide is written with the assumption that you have access to a complete CMS. If not, then you can't follow his steps verbatim. It probably didn't matter what primary was adjusted initially with the color control because he was assuming you would be able to adjust brightness of the other two primaries as a final step. Given we can't do this without going into the service menu, I will use your spreadsheet and make the best compromise I can. I think I will err on getting blue brightness the closest but allow for some deviation if there is significant brightness error in red and green.

This stuff is kinda fun. And thanks again for taking the time to help get me back on track.

I found the Delta Luma values (after switching the view window from "grayscale" to "primaries & secondaries"), and boy is it off. My Delta Luma is at -45% to -50% on all colors !!! I thought I was using 75% windows to measure gray and red, but I think I must have used a 100% red window to get this error. And I must have repeated the error.

The 100% red window seems to be the only possibility because I wrote down my 75% gray reading of Y=66.5 and then calculated that I needed a Red Y of 14.1, which is exactly what I measured after lowering the color setting. Oh well, I guess this is good practice and I am certainly learning from my mistakes.

hwjohn, thanks for commenting.

What APL stands for?

APL = Average Picture Level. An all white screen would have the highest APL and and all black screen the lowest.

Exactly.

ColorHCFR asks for IRE and these discs show "% grey". I'm confused. There is apparently a difference between the two, yet these discs are supposed to work with ColorHCFR. Am I missing something in ColorHCFR setup?

PS3 has RGB Auto and RGB Limited. It also has YpbCr and "super white" settings.

My Samsung 71 has Color gamut choice of Wide or Auto. It also has xvYCC.

Which combination of PS3 and TV settings is correct with the HD709 disc?

The two terms are often used interchangeably. For a digital display, they are equivalent.

jvincent hit it on the head again. We use % because, strictly speaking, IRE is only applicable to analog signals (such as component). If it asks for 10 IRE, then 10% is the corresponding pattern on the disc.

It isn't so much about what is right for AVS HD, but which settings allow your calibration to be correct for both games/XMB menu and BD/DVD. Do this:

On your Samsung 71:
I can't give specific instructions on this as I don't own one, so consult your manual or owners thread for more info.
1) Make sure it is expecting video levels (16-235).
2) Color gamut should be set to the setting that most closely matches Rec. 709 primaries.
3) xvYCC isn't applicable to anything... yet. Turn if "off" if that is a setting.

On your PS3:
BD/DVD Output: YPbPr or Auto
RGB Full/Limited: Limited
Superwhite: On

They are often used interchangabley, and thus incorrectly. They are not the same except under a special (not normal) circumstance. Whether or not the HCFR authors know the difference I can't say. Surely they do.

The issue is generally about what the measuring hardware/software is expecting. If the software is expecting IRE, then one must know if the software means IRE with pedestal or not. No pedestal; IRE=%stim. But almost all software is expecting the pedestal so for example, 10 IRE is not = 10%stim As you get closer to 100%, then they get closer (100IRE=100%). But except at 100%, they are not equal.

This generally only affects gamma calculations. The software may say to display and measure a "10 IRE" pattern. It is expecting a certain luminance level for that reading. It will assume what you displayed is what it expected. If you are using a source pattern than is putting out %stim, but the software is truly expecting IRE, then the luminance level will be off, and thus the software will calculate the gamma incorrectly.

I don't use HCFR so I don't know if they built in the ability to switch what it is expecting (i.e. IRE, IRE with pedestal, % stim) or not.

In any case %stim is NOT equal to IRE for digitals. For example the 10 IRE pattern from Avia will not be the display the same luminance/brightness on screen as the 10% stim pattern from my DVD which has %stim level patterns. FOr digitals, the digital values for 10IRE are not equal to the digital values for 10% stim.

You can see a chart of the relationship between IRE and % stim, including digital values for each in the appendix of my manual avail at www.calibrate.tv.

HTH,
Scott

Thanks to all of you that responded. As alluringreality points out, when the calibration is done in ColorHCFR, the text box asks for "IRE", but when results are displayed, they are displayed as %gray.

Thank you! I need to read about the Sammy's settings more, as you suggested. Thanks to this reply, I will know what to look for.

These PS3 settings are all exactly opposite of what I have been using for calibration. Jeez. Thanks! Back to work!!

Oh, and as far as games vs BR DVD playing goes, I am only interested in calibrating for BR DVD color accuracy.

I also want to set up my DirecTV input but only have the test patterns on the BR disc. Can I just use the settings obtained from calibrating for BR DVD, or is there some other combination of BR/TV color settings that would make a different calibration?

Maybe Tom will see this and respond. I have been thinking about this as well. Yes, I agree, it IS fun.

My opinion of why he chose Red as the primary to use in setting Color is that it is in between the other two, thus making any blue or green error only "one off", whereas using either blue or green to set Color makes the opposite end of the spectrum (green or blue) errors "two off". Also, blue error is the most tolerable to our eyes. Red and green errors bother more.

So, if you don't have CMS, you are as close as you can get to having all 3 in the ballpark by using red (although I have wondered if there isn't actually a intermediate point someplace between red and green (closer to red, probably) that would make all the errors comparable. And even if you do have CMS, by having your errors only one off, you have to make less of an adjustment.