"High Definition Benchmark" BD Edition by Stacey Spears and Don Munsil - Page 4

The ramp is useful as a cross-check, and allows you to quickly see that clipping is going on. It's not intended to be a precise clipping measurement. Use the numbered bars to see whether the display is clipping, and if so how much.

The point we were trying to make was that when everything is adjusted correctly, the black-white-black ramp will not have a thick stripe of even brightness in the middle, and the white-black-white ramp will, and this is not an error. Again, it's intended to be a handy quick reference. You can put up this pattern and see instantly if there is significant clipping, once you know what to look for.

Myself, I go for the bars and barely look at the ramp for clipping purposes. Stacey likes to use the ramp for coarse adjustments, then fine tunes by looking at the last few bars. But that's only useful once you know what it's supposed to look like and can tell when it starts clipping. To do that, you'll need to use the bars at least for the first time or so.

Does that make sense?

The ramp goes from levels 1 to 254 to 1. Each level is 3 or 4 pixels wide. The center level, 254, is 1 pixel wide. If you are on a 1080p display, then the center line should be 1 pixel wide.

You also want to ensure the ramp is as smooth as possible. Certain spots on the contrast control may show banding in the ramp. As you adjust, the banding may get better or worse. Go through the full range of contrast and see if you can find the smoothest ramp that is below the clipping point. On my DLP you will see a coarse ramp, then finer, then finer and then perfect. It snaps into place. So 8 bands, 4 bands, 2 bands and then no bands.

Don, I remember you from back in the hey day of the Sony VPL-VW10HT and the excellent FAQ you authored that was so helpful for user calibrations. So I have taken notice of your new endeavor. I was out of the front projector world for a few years, but have recently purchased a Panasonic AE3000 and used the only calibration disc I have - the Avia Guide to Home Theater, 1999 - to make some basic adjustments. Although the PQ is very good and a big step up from the 10HT, there may be some room for further refinement.

In a nutshell, can you tell me a few advantages of your disc over the old Avia and specifically what I might want to look for to further optimize the AE3000? Many thanks.

-Rod

Factory reset of -83 performed. Denon out of signal path, -83 connected via HDMI direct to PRO940.



Yes there are significant shifts as output resolution is changed (summarized at bottom of this post). The 940 does not have different resolution storage, AFAIK.


There is no shift (>+/- a pixel or 2) when changing output resolutions on the PS-3.


There are no independent resolution settings that I can find. There is an overall positioning adjustment, however, from what I can determine that apples universally.


Then, Houston we have a problem. See summary below.


The -83 is set to 16x9 - not 16x9Auto.


CEC has always been set to OFF.


I'd be interested in that data.

Summary of test results

BDP-83 - Using the HD Crop pattern from the menu:

Rez	Top	Bottom	Left	Right
720p	17	18	20	30+
1080i	17	17	26	30+

BDP-83 - Using the Crop pattern from 127:

Rez	Top	Bottom	Left	Right
720p	9	7	11	12
1080i	9	7	11	12

PS-3 - Using the HD Crop pattern from the menu:

Rez	Top	Bottom	Left	Right
720p	19	20	23	29
1080i	16	19	24	30

PS-3 - Using the Crop pattern from 127:

Rez	Top	Bottom	Left	Right
720p	17	19	21	28
1080i	15	17	23	27

Added Weirdness...

When changing in/out of "direct", these numbers will vary wildly - Most noticeably on the L/R values, up to +/- 6 pixels!! The T/B values also move, but they are in the +/3 range. There appears to be no correlation for this, other than each time I enter/leave direct mode - the 1080i and 720p values are different from before.

I received this response from Oppo, after emailing them and referring them to the posts in this thread:

"Each resolution has different horizontal and vertical frequencies. When you adjust for one frequency subset, you will inadvertently mess up the defaults for another frequency subset.

You can see this with CRT monitors: at different resolutions and output frequencies will cause the size of the image to change."

Comments? Thoughts? Suggestions?

Don has a 720p LCD with HDMI in. I will bring the OPPO over this evening and we can see how they interact.

Can you take a look at the DVE patterns and tell me title and chapter of the patterns you are using? Its MUCH easier to find them that way!

What OPPO said about the timing, and CRTs, is why I was asking. There may be additional information sent over HDMI that is keeping the PS3 more or less in the same position, but that is just speculation at this point.

well for home PC you can use the disc to test how it handles 1080i ATSC broadcast I suppose or 1080i camcorder footage.

lots of LCD still don't do 24p

and there are 1080i blu-rays, especially concerts of any sort, many seem to be in 1080i

Well, for starters this disc is high-definition. It's impossible to check 1080p sharpness, for example, with a standard-def source. And standard-definition content usually goes through somewhat different signal paths, either in the player or in the display. Getting your display and player adjusted properly for playing back DVDs will not necessarily produce correct results for Blu-ray content.

Otherwise, I'd say about Avia what I said about DVE. Avia is a good disc. We know Guy Kuo and have used Avia extensively over the years. That said, we like to think we've made some useful improvements with our patterns. Our chroma alignment pattern, for example, is very easy to use and accurate to at least a quarter pixel. The Avia chroma alignment pattern is very hard to use and not nearly as precise. This isn't Guy's fault: he had to create the pattern in RGB and was at the mercy of the chroma downsampling applied by the video encoder. By writing our pattern directly in YCbCr and in a format the encoder could take in directly, we could position the chroma lines exactly where we wanted them.

Right now we're doing an article about how to choose which color space to output from your player into your display. Like so many things in video, the answer is "it depends." You need to view a variety of patterns in each mode to find out which component is doing the chroma upsampling best, has the least rolloff, best sharpening, etc. The test patterns you'd need are on our disc.

Similarly, some people have a choice of having their player or their display or a video processor do the deinterlacing of interlaced (and interlace-flagged film) content. Which is better? Our disc can tell you the answer.

When you change video presets on your projector, what exactly happens? Our disc can help you find out.

As far as optimizing goes, I would not want to tell anyone this disc is going to make a night-and-day difference in the final picture. For some displays it absolutely will, for others not so much. If the display is already dialed in pretty well, we can't dial it in any more. But if you want to know things like which deinterlacing mode to use on your player, or what color space to use, or what happens to the sharpness when you change video modes, or what difference a different player might make in your setup, this disc will help you with those answers.

And if you can't tell whether the center line is 1 pixel wide, you should do what I do: roll your eyes at Stacey and use the bars on the bottom of the image.

In all seriousness, if you move the contrast control around, you'll see what Stacey is talking about. When you turn up contrast past the clipping point, you'll see a wide white band start to show up in the center of the ramp, and it gets wider as you turn up contrast. Turn down contrast, and it gets narrower. When it stops getting narrower, your display is not clipping. Unless the clipping is happening earlier in the signal chain, in which case you can turn down contrast all you want and the middle line won't get any narrower.

The advantage of the bars is that you don't need to know what the ramp should look like to read them. The advantage of the ramp is that it's compact and very quick to use once you know what it should look like. That's why we have both.

Well, actually, I just wanted the bars, but Stacey made me put in the ramps.

Yeah they both are working for me now, I just wanted to clarify what the width should be on the white ramp, and it makes total sense that it should be narrow to prevent clipping. When I change the contrast/picture control on my set, its hilarious/distracting because when the contrast is too high all my whites turn pink, apparently due to my weirdo white balance settings. Good times.

Sorry... had to make a trip down to ATL yesterday and did not get back until this morning. I know how hard DVE is to navigate through

Stacey, would you sacrifice the visibility of above white bars (eg > 235) to improve the banding? In other words, how important to film are the above white sections? Also, can one not increase overall contrast ratio by clipping at the 235 white point?

Cheers,

Mark

Hi Mark,

It is possible to increase the contrast ratio by clipping the 254 down to 235. It is not something I personally do, but I know others who are trying to get the biggest CR possible.

When I had the Samsung SP-A7000, the banding would come and go. Once I found the clipping point, I had 8 distinct bands. One click down on contrast changed it to 4 bands. Another click 2 bands and then finally 0 bands. If you went one click further, the bands would return.

Then there is Marantz 11S1 and 11S2. The 11S1 was similar to the Samsung, you could find a sweet spot. The 11S2, on the other hand, has banding no matter what. One concern has been the DC3 vs. DC4. My Samsung 800 has a DC2. As soon as it is available in the US, I will have the 900 with the DC4, so it should be easy to eliminate the DCn as the culprit.

You can increase measured contrast ratio by clipping at 235, because typically contrast is measured between 16 and 235. I personally think contrast on modern projectors is so high that a small amount of lost contrast is not a big deal compared with highlight clipping.

In our perfect world, a display would actually compress the above-235 material with a shoulder curve, just like a CRT. So if the display actually has, let's say, 1024 available levels, 235 would be mapped to around 990 or so, and the remaining above-reference stuff would be gently rolled off, with appropriate dithering. That would match the behavior of a properly-calibrated BVM pretty well.

In practice, there can be useful visible info all the way up to 254, so we advocate making sure it's all visible. But there are diminishing returns, in a sense. Each additional level above 235 adds a little less. So if I could avoid visible banding by clipping the top two or three levels, I'd do it in a heartbeat (though I'd still be annoyed with the display manufacturer for making me do it). But if I had to clip off everything above 235 to get rid of banding I would be unlikely to do it unless the banding was really bad (and then I'd start shopping for a new display). Hard clipping at 235 is obvious and ugly in lots of real material, so it's not something I'd want.

Thanks guys, that's very helpful.

Mark

I just wanted to add one very major point about the original Avia (in addition to what Don covered above), is that all the patterns on the original Avia disc are limited to 16-235. There are no below-black elements (below 16), or peak white elements (above 235) on the original Avia disc. This makes Avia useless for testing clipping of peak whites, or clipping of below black. The absence of below-black elements makes setting black level more difficult.

Another issue is that the grey patterns on the original Avia are not perfectly grey, so they aren't the best choice for calibrating greyscale, thus why Guy recommended (back in the days of the original Avia) turning the color saturation down all the way while using the grey patterns for greyscale alignment.

Avia II has patterns that go the full range, and have below-black and peak white elements, as does DVE. And of course the Spears and Munsil disc does too.

In any case, I would recommend all of the above test discs, except the original Avia which if you do not have it already, I would get Avia II instead for sure.

This was a limiation with the MPEG2 encoder used at the time. This is why Guy came up with the 2% idea to work around the encoder limiation. It was a great idea and everyone has done it since.

Guy worked closely with them to improve their encoder for AVIA Pro, which translated into AVIA 2. The MPEG encoder is also to blame for error in the window patterns.

For a second there, I was wondering what was wrong with my set as I tried setting both Brightness and Contrast to their respective extremes, and could still not see half the test bars on many patterns. Turning 'Super-White (HDMI)' on in my PS3's menu fixed that - I figured that was just some kind of boost, not that it'd be necessary to see half the range!

Hello,

I'm using your disc to test my Oppo BDP-83 and calibrate my set.

My question: should I use the "Standard" mode on the t.v. for calibration even though I always leave the set in "Movie" mode for my viewing? I have a Samsung DLP rear projection t.v.

Thanks,

snudley

You should always calibrate in the mode you plan to use. Each mode changes the settings of the picture controls.

Kudos guys on finally getting this disc out. I've been waiting for you guys to produce a disc for years.

I have a question about the CUE pattern. I'm only seeing a VC-1 encoded test clip. Should there be an MPEG-2 and AVC clip as well? Or am I confused? I'm pretty sure that's it.

Would be nice to see some film-based 480/60i material for torture testing players like the Oppo at converting to 24p. De-interlacing and off-cadence 24fps mixed stuff.

I wish that test discs had an actual clapboard to help sync the receiver audio delay up. It seems to be very hard for me to sync up audio my watching people talk, whereas a clapboard seems more ideal. Maybe not...I dunno.

Go to the menu called setup and select SD VC-1 or SD MPEG-2. This will switch the the source adaptive, and edge adaptive, clips from 1080i60 to 480i60.

just to be sure, this disc does not have 1080i50 video(not film) test patterns ?

No, the video/interlaced test content is 1080i60 and 480i60.

Am I correct in saying you need to access the service menu of the TV to make use of this disc? I'm referring to those settings that can scare the heck out of you and can mess up your TV if you don't know what you are doing. I have a 720P plasma hooked via HDMI, and because of that connection I only get to make sharpness and balance adjustments using Avia. If there is a way to make color adjustments I haven't found it. Fortunately, looking through the filter they don't need adjusting.

Maybe.

There are patterns on the disc that will help you figure out which color space to use for BD. e.g. 4:4:4, 4:2:2 or RGB. We have not provided the details on that yet, it is in an upcoming article for our website.

The disc can help you figure out if you should let the player or display deinterlace your SD content, assuming you are going to use the same BD player for that.

Your display does not let you adjust saturation or hue for HD? What about brightness and contrast?


For HD, this is how it should be.

This works well when editing video. It may not be the best for setting lipsync. We created one pattern for it, but we want to further refine and support the various resolutions and frame rates so you can test lipsyc for all combinations.

In theory? no. In reality? Unknown. I do know that for PIP you need at least AVC and VC-1 because one platform on the market would show PIP correctly for AVC and not for VC-1. They have fixed the bug, but not sure if every player built on that platform will get the fix. Next time we need PIP in all CODECs to be safe.

Through HDMI, the only adjustments displayed for me are brightness, sharpness, contrast, and advance white. If I’m connected via component, then saturation and hue will be displayed and I can make the adjustments, but they aren’t displayed thru HDMI. Adjusting contrast with Avia gives zero results when looking at their pattern. It’s been years so I can’t recall with 100% accuracy, but I believe Avia’s contrast pattern is unusable with a plasma. Thanks.