New calibration disc - Page 24

Greg, I believe I answered incorrectly re "that's exactly what I do". It is not. Let's see if this helps. Here IS what I do.

Take for example a 0% amplitude pattern. It has a window that is encoded as RGB 16/16/16. The "frame" of that window is intentionally encoded below black, the maximum below. The original pattern's frames were created using RGB 0/0/0. Ursa suggested that technically this shoudln't be done and it would be more proper to use 1/1/1. Whether or not this is correct may be technically moot since no one's properly calibrated display is going to show levels of 1 or 0, valid or not.

To create the pattern, I take the RGB (photoshop) image and convert it to YCbCr studio levels using a C program. This YCbCr is fed to an mpeg encoder that accepts YCbCr natively, and created the mpeg structure from it.

RGB 0/0/0 converts to YCbCr 0/128/128 with my C program. This is meant to be "below black" information, and the lowest value of below black possible. Also, the mpeg encoder is set to accept values 0-255 so it does not clip the <16 information back to 16.

For a "0% amplitude" video black window, the level is RGB 16/16/16. That is converted to YCbCr as 16/128/128 before being encoded to mpeg.

The disc is intentionally created to be able to show above white (236-255) and below black (0-15) information. It is meant to be used to adjust so 16-235 is the tuned black and white level (+/-).

The issue came about when saying a RGB 0/0/0 (which is converted to YCbCr 0/128/128) was technically a valid level (or not).

Like Dave said, it's academic at this point since if adjusted correctly the display wouldn't get anywhere near actually displaying digital level 1, much less 0. It will clip anything below 16 (or slightly lower if one tuned to allow some BTB to show).

It was whether or not it was correct to use YCbCr 0/128/128 as the lowest possible value of black or was it more correct to use RGB 1/1/1 (YCbCr 1/128/128).

That help clear up what the question/issue was?

And I believe that answers the original question, and agrees with Ursa. The "maximum" level of black to be on a DVD in YCbCr should be YCbCr 1/128/128 (RGB1/1/1). For this discussion, call it "Maximum black" unless there is a better/proper name for it.

As to why such a value was used, I knwo I could have used say level 10 for exapmle to be sufficently "below black". I used Maximum Black in places where I wanted to ensure the display was displaying as black as possible. Places where teh display should, in as much as possible, be "off", not giving any color information.

If someone had their brightness too high, a 0% amplitude window would be visible on the darker BTB (and in my disc's case Maximum Black) background. If I was to have a BTB area/element for the sole purpose of being as dark as possible, I decided it might as well be the "Maximum Black" value. In addition to having that level as the background for the gray window patterns, I have one full screen "Maximum Black" window. Those are the only palces I remember having "Maximum Black" off the top of my head.

The Maximum Black pattern is intended to be used to measure on-off contrast. I know this does not account for the brightness setting (if higher), but my intent was to be able to measure a maximum on-off contrast ratio as-calibrated. So if one had their display set to display some below black information, this "Maximum Black" pattern will show that exact level, instead of the brightness of a level 16 pattern which could be slightly higher in luminance, resulting in a worse contrast ratio calculation.

So Claus, I'll straighten out the 0 and 1 thing in my rev.12 and pass back to you.

Greg, thanks very much as always for taking the time to give input.

Everyone else, the "Maximum Black" levels on the current DVD's are already RGB 1/1/1, YCbCr 1/128/128 and are correct.

Cheers,
Scott

Actually it's more simple than that. I just don't use RGB 0/0/0 when creating. No need to do it programatically. No need to clip, definately not at 2/253. If 1/254 are proper limits for YCbCr, that's exactly what they will be. Actually, black is already set that way. I will change white 255 to 254 in about 3 places.

S

OK guys. I've released Beta 6.0 Try not to get it all at once. By popular request it has a basic set of full field patterns. I do not recommend their use however unless you have avery specific need and fully understand why you need them over a window pattern.

The current revision of the documentation is also available on the download page to contributors.

Beta 6.0:

Corrected 2 pixel error on "Letterbox" line of alignment/Overscan pattern. Renamed "Letterbox" indicator to more correct 2.35:1
Changed opening screen and disc info/credits
Changed menu wording to Maximum White, Maximum Black.
Added 1 sequence of 10% step full field gray patterns, labelled.
Updated opening screen to production version
Updated credits screen to rolling credits with contributors, info, etc.
Reduced volume on auto gray pattern beeps

Scott

Rick: See docs, Q&A section: "Can I use the Get Gray DVD to calibrate my CRT device?". (Answer is, it's not designed for CRT's). But FWIW, all of the patterns with the exceptions of Y/C delay, Frequencey Bursts, andthe rolling credits are created from an identical process. If some "jiggle" (on unsupported CRT) and some do not, I would be curious as to which do and don't.

Best,Scott

Thanks for the help, but...No, it's not. Website is www.calibrate.tv

"tv" is the TLD (top level domain) not "com" . I chose the tv TLD for obvious reasons, even though it's not what tv really stands for...

Oops, missed your reply. Thanks very much for the technical detail as to the whys. Version Beta 6.0 of the DVD now has no occurances of RGB 0/0/0 nor 255/255/255. 0/0/0 was already out, I had to step 255 to 254 in about 3 places which is done.

Thanks again, Scott

OK guys, I think that will wrap it up, or close. Claus and I are on back and forth iteration #20 of the documentation and the current version is available on the website (again that's www.calibrate.tv). I've added the requested full-field patterns. I haven't left out much that was ever requested. I'll be going over the patterns one last time carefully and presuming nothing else is identified, version Beta 6.0 is a release candidate.

Time to look into PAL.

Don't laugh. I'm salivating over the encoder now. Only $499. I'll need a lot more contributors before doing it though

Well guys, it's been a long road getting to this announcement. Much longer than I originally expected. I guess that's the way projects go sometimes, especially one that's this popular. This is not a message to say the project is over, but one to announce it has reached a major milestone.

It's been educational and fun for the most part. The disc would not be where it is today without all the help and input from those here. It's taken about 4 1/2 months of picking away at it in all our spare time, but the result is what I think is a fine DVD indeed. All patterns have been scoured and checked, tested and triple checked. It has been tested and used by professionals and novices alike. It is already in use all over the world which is quite flattering for me. I've even made some friends in other countries from this project.

That said, this message is to announce that I am declaring the version Beta 6.1 as the first NTSC production release, GetGray Digital Video Calibration Version 1.0.

I would like to say a special thanks to Claus Wiebe who has worked with me extensively behind the scenes on the documentation. The final documentation draft version was #24. That's 24 revisions after I started "helping" with the documentation . I am proud of the finished documentation as well as the DVD. For those who followed along from the beginning, you know my original intent was to provide no docs other that some technical details in a text file. This is a far cry from that plan. Thanks to Claus's help and patience, together we have cobbled it into a professional piece of documentation. Claus continues to help me by helping others here on the thread. I don't have the resources to answer a lot of the questions, especially from new folks, but Claus has taken the time and has the patience to help. So again, a special thanks to Claus's participation. Maybe he'll be comfortable recommending his work now

I would also like to sincerely thank the experts here who have chimed in, especially:

Ron (dr_1394) for his outstanding input, programming, and behind the scenes assistance on the RGB to YCbCr details of which the disc would have been impossible without. We can thank Ron for having a DVD that is perfect in levels and color encoding.

Greg Rogers, extremely well known and highly respected in the AV community, who's input was timely, very welcome, and extremely valuable. To have someone of his caliber take time to contribute has been incredible.

Chris Wiggles whose original documentation works on Blacker than Black were my first learning experience on the subject a good long time ago.

Bob Sorel one of our resident AVS HTPC Forum Moderators, for giving me the impetus to go on and do the thing which I contemplated every time I switched calibration discs to get the pattern I wanted.

And to all the other folks I referred to early on as "big dogs" (a compliment ), as well as everyone who provided input and constructive criticism, thank you very much for the help as well.

The project is not complete yet. I like to think of it as a "living work" however this release makes it much less live . The next step is to deal with a PAL version. So any of you technical folks don't run away .

As for the future of the NTSC DVD, I remain open minded about it and will watch for opportunities to improve it with out complicating it, thus defeating it's original purpose. We'll keep an eye on the Ruby/DI issue and I might put something together for that purpose after dealing with PAL.

Now, for downloading the production version, please, don't everyone hit my server at once. If you don't need it right away, wait until you do, then download it to get your copy of the NTSC version 1.0 DVD. Note that the zip file contains BOTH the VIDEO_TS folder and the documentation pdf. You should burn them both to the DVD just as they exist in the zip file.

Everyone is required to read and agree to the License agreement and warnings before use. Page 1 of the document pdf.

See the first link of this thread for links to the patterns and documentation

next, on to PAL land..... the cycle begins again

Cheers,
Scott

No, 6.0 was the last Beta released to the public. Now we skip (back) to production version NTSC 1.0. It is uploading now. Be ready for everyone tomorrow. Moving credits take a lot of room. Disc is 70+Mb now.