Spears & Munsil 4K UHD HDR Benchmark (Test Discs) May 15, 2023

[Kris Deering]

Originally Posted by sapiendut

Kris, please post a link if you have it. I wanted to do a review of the disc when I received it but if I end up saying mostly what you already say, there is no point for me in writing it.

My article isn't a review per se, just an article on the disc and its contents. I didn't rate each pattern but went over what it has, some highlights and some authoring notes for the readers. I may do some blog posts on specific tests available on the disc on my website blog to highlight some of the cooler patterns on the disc.

[Kris Deering]

Originally Posted by Geoff D

There was on the 2nd Ed, dunno about the UHD.

There is one on the new disc as well in both Dolby THD and DTS MA. This is the only audio test on the disc this time around. S&M are looking to make a companion disc with a more complete audio test section in the newer immersive audio formats.

[sapiendut]

Originally Posted by Kris Deering

My article isn't a review per se, just an article on the disc and its contents. I didn't rate each pattern but went over what it has, some highlights and some authoring notes for the readers. I may do some blog posts on specific tests available on the disc on my website blog to highlight some of the cooler patterns on the disc.

Ah Ok. I’m planning to approach it more like why you need this disc.

[INdetectableMAN]

Hello my friends.

I wanted, as a curiosity, to bring you a doubt that has come to me after having been doing some tests using the reference material that brings the disc.

The test I made it on a Panasonic FZ800 and Panasonic UB9000 player with HDR OPTIMIZER ON / OFF.

The Panasonic FZ800 OLED always in Professional 1 mode and for all the photographs the same settings have been used, as is obvious.

At 1000 nits, there is no difference. However, setting the configuration to 4000 nits there is a more than noticeable difference, so I tend to think that the Panasonic with the HDR OPTIMIZER activated is executing a tone mapping different from the Panasonic FZ800 and also different from HDR10 +

Disk configuration:



HDR10 OPTIMIZER ON


HDR10 OPTIMIZER OFF


HDR10+


I understand that the optimizer ON is doing a different tone assignment, badly, or is it recovering volume of color instead?

What is correct, the Panasonic ZF800 or the Panasonic UB9000?

I have to say, that this during the usual viewing of movies has never happened to me, I have only been aware of this with this demanding reference material, which I intuit, leads to the maximum of its possibilities, the OLED panel and its limitations in terms of volume color.

[sapiendut]

Which target nits you use on the player (OLED, low nit LCD, etc)

[INdetectableMAN]

Originally Posted by sapiendut

which target nits you use on the player (oled, low nit lcd, etc)

oled

[Geoff D]

What you're seeing is the Optimiser turning the PQ curve into its own thing. I did some measurements with my colourimeter when I got the 820 and posted them in that thread, the Optimiser basically flattens out the steep PQ curve into a much gentler and much more linear SDR-style brightness curve. This graph is from the Medium 1000-nit Optimiser setting when fed 0-100% brightness patterns with 4000-nit mastering metadata, yellow line is the Optimised reading and red is the untouched PQ curve on the ZD9:

[Stacey Spears]

Originally Posted by sapiendut

Is there an audio sync wheel like the one available on the Wow disc?

It has the A/V sync from the 2nd edition plus one designed for use with the Sync-One2.

[Geoff D]

My copy turned up earlier, I unwrapped it like a kid on Christmas Day. It's all rather familiar and yet...different. The ability to change metadata/nit levels on the fly is awesome, but I have to echo what others have said re: the lack of an info screen. There are plenty of patterns that tread similar paths but there are several new ones and having some info for them would've been handy, not just as a separate piece of documentation - which will be appreciated, don't get me wrongo - but something that can be referred to on the fly.

I've been through the montage several times, toggling back and forth between the various versions, I'll probably kill the drive in the OPPO before the day is out! No banding on the opening fade up using DV LL on the ZD9, but in HDR10 I spotted a trace of it on those backlit shots of the windmill-type thingies, naturally the DV smoothed it out.

Regarding the mapping, even the DV with default 90 contrast on the ZD9 can't fully resolve some of the information on display, like chapter 2 with the horsies grazing in the snow. When it cuts from the previous shot of the trees to those horsies then I can see the highlight information in the snow for a second and then the DV mapping takes over and clips it out. It's still resolving way more highlight information than the >1000 nit HDR10 versions in my default Cinema Pro mode - which clips at ~1500 nits - but it's still losing some of the information that's there in DV.

Given that the Panny DV implementation seems to clip at 2000+ nits then I wonder if the OPPO's clips at 4000+ nits, this is why I'd love to get some DV-enabled versions of these test patterns. These very bright patterns also reiterate how the DV is dimming overall luminance compared to HDR10, that's a function of tone mapping to be sure but it really seems to dim the image on a global level which is very obvious when one can switch between the versions on the fly.

You know what the mad thing is? The HLG version of the montage is by far the most balanced presentation of it on my TV (HLG can be manually enabled on the ZD9). No clipped highlights whatsoever, even better than the DV, and lovely solid luminance and colour, so as much as I love HDR it's things like this that make you realise they missed a trick when it came what format they should use. Penton has mentioned that PQ has certain advantages over HLG, fair enough, but a relative luminance system really would've made HDR life so much easier.

[LordoftheRings]

▪ https://www.soundandvision.com/conte...ation-disc-hdr

[HDTV1080P]

QUOTES from Sound and Vision magazine:


"Stacey Spears and Don Munsil, creators of the acclaimed Spears & Munsil High Definition Benchmark series of test and calibration discs, have announced the release of the UHD HDR Benchmark, a 4K Blu-ray disc dedicated to the calibration and evaluation of 4K Ultra HD (UHD) displays that support high dynamic range (HDR)."



“There really isn’t any other way to make patterns that we can stand behind,” said co-creator Don Munsil. “We build every pattern using our own tools, written from scratch in C++. If a pattern needs to be generated directly in a very specific color space and data format, we generate it in that color space and format; we’re not limited to what you can do with off-the-shelf graphics software. In a few cases we’ve had to create our own format, because no existing file format could represent the pattern we needed to create.”

Read more at https://www.soundandvision.com/conte...f2rZmgD4ADK.99

[Scenic Labs]

The goal is to get the user guides up in fairly short order. First, for the novice to accomplish the goal of getting the best settings for their setup sans instrumentation. Just a little bit later, a more comprehensive guide will follow. And then off to translate the guides into every language that our dealers need.

Originally Posted by Geoff D

My copy turned up earlier, I unwrapped it like a kid on Christmas Day. It's all rather familiar and yet...different. The ability to change metadata/nit levels on the fly is awesome, but I have to echo what others have said re: the lack of an info screen. There are plenty of patterns that tread similar paths but there are several new ones and having some info for them would've been handy, not just as a separate piece of documentation - which will be appreciated, don't get me wrongo - but something that can be referred to on the fly.

I've been through the montage several times, toggling back and forth between the various versions, I'll probably kill the drive in the OPPO before the day is out! No banding on the opening fade up using DV LL on the ZD9, but in HDR10 I spotted a trace of it on those backlit shots of the windmill-type thingies, naturally the DV smoothed it out.

Regarding the mapping, even the DV with default 90 contrast on the ZD9 can't fully resolve some of the information on display, like chapter 2 with the horsies grazing in the snow. When it cuts from the previous shot of the trees to those horsies then I can see the highlight information in the snow for a second and then the DV mapping takes over and clips it out. It's still resolving way more highlight information than the >1000 nit HDR10 versions in my default Cinema Pro mode - which clips at ~1500 nits - but it's still losing some of the information that's there in DV.

Given that the Panny DV implementation seems to clip at 2000+ nits then I wonder if the OPPO's clips at 4000+ nits, this is why I'd love to get some DV-enabled versions of these test patterns. These very bright patterns also reiterate how the DV is dimming overall luminance compared to HDR10, that's a function of tone mapping to be sure but it really seems to dim the image on a global level which is very obvious when one can switch between the versions on the fly.

You know what the mad thing is? The HLG version of the montage is by far the most balanced presentation of it on my TV (HLG can be manually enabled on the ZD9). No clipped highlights whatsoever, even better than the DV, and lovely solid luminance and colour, so as much as I love HDR it's things like this that make you realise they missed a trick when it came what format they should use. Penton has mentioned that PQ has certain advantages over HLG, fair enough, but a relative luminance system really would've made HDR life so much easier.

[Geoff D]

Stacey said that y'all are working on it but after the 2nd Ed I'm gonna miss having it be part of the actual disc.

[INdetectableMAN]

Originally Posted by Geoff D

What you're seeing is the Optimiser turning the PQ curve into its own thing. I did some measurements with my colourimeter when I got the 820 and posted them in that thread, the Optimiser basically flattens out the steep PQ curve into a much gentler and much more linear SDR-style brightness curve. This graph is from the Medium 1000-nit Optimiser setting when fed 0-100% brightness patterns with 4000-nit mastering metadata, yellow line is the Optimised reading and red is the untouched PQ curve on the ZD9:

I already know that, but the question is: What is the right thing? HDR optimize on or off? Because the difference in color beyond preserving the highlights is obvious.

On the other hand, it surprises me, and a lot, that a Panasonic UB9000 player makes a different tone assignment to its same generation OLED.

[Geoff D]

Had a look at the banding (quantisation) tests and I'm still convinced that the Panny's Optimiser is adding some banding. It's slight so it's very hard to capture in a photograph but for my money it's definitely there.

Found something interesting re: DV improving artefacts. If you go to the peacock feather in the montage, look carefully at the inner 'eye' of the feather on the right side, in HDR10 on the ZD9 (regardless of nit level) I can see some little bits of chroma noise dancing about. But when I switch to Dobly, either MEL or FEL, the noise is cleaned up.

Stacey, you mentioned here https://forum.blu-ray.com/showthread...2#post16564692 when talking about future DV patterns that only one of the three DV files (MEL, FEL, and ???) would actually do anything on a low latency display, what did you mean by that? Is the player not reconstructing the full 12-bit stream from the base + FEL configuration because it's having to do too much work already e.g. the dynamic mapping? So even with a FEL config it's ignoring the extension data and showing the base layer + dynamic metadata only with an uprez to 12-bit at the end? I've gotta say, if that's the case then the Dolby processing engine is even better than I thought it was!

[Stacey Spears]

Originally Posted by Geoff D

Regarding the mapping, even the DV with default 90 contrast on the ZD9 can't fully resolve some of the information on display, like chapter 2 with the horsies grazing in the snow. When it cuts from the previous shot of the trees to those horsies then I can see the highlight information in the snow for a second and then the DV mapping takes over and clips it out. It's still resolving way more highlight information than the >1000 nit HDR10 versions in my default Cinema Pro mode - which clips at ~1500 nits - but it's still losing some of the information that's there in DV.

Given that the Panny DV implementation seems to clip at 2000+ nits then I wonder if the OPPO's clips at 4000+ nits, this is why I'd love to get some DV-enabled versions of these test patterns. These very bright patterns also reiterate how the DV is dimming overall luminance compared to HDR10, that's a function of tone mapping to be sure but it really seems to dim the image on a global level which is very obvious when one can switch between the versions on the fly.

You know what the mad thing is? The HLG version of the montage is by far the most balanced presentation of it on my TV (HLG can be manually enabled on the ZD9). No clipped highlights whatsoever, even better than the DV, and lovely solid luminance and colour, so as much as I love HDR it's things like this that make you realise they missed a trick when it came what format they should use. Penton has mentioned that PQ has certain advantages over HLG, fair enough, but a relative luminance system really would've made HDR life so much easier.

The OPPO does not clip the DV signal and I am pretty sure the Panasonic does not either. I don't use the HDR optimizer, could that be the cause?

The clipping on the Z9D will go away if you pull contrast down. You should not have to pull it down in DV, but you do on the Sony. LG is fine in that regard. Don't have any other DV displays to test. Something I have not done yet, but need to is:
1. Look at tone mapping ramp (white) and see where it is clipped.
2. Play chapter 2 and pull contrast down until clipping goes away.
3. Look at tone mapping ramp again and see how high it goes now.

I forget if contrast is different for DV and HDR10 on the Sony. Easy enough to copy settings if a different memory is used.

I fully understand on the pop-up help. For this disc we felt being able to change nit levels was more important this time around.

BTW, I finished the tool to create the DV binary (calibration metadata). Hoping to test it soon. Probably next week, after the holidays. Basically this bypasses Dolby's tool that generates the binary file that gets muxed into the EL. This will only work on the LG at this point. Hopefully others implement it soon.

[Stacey Spears]

Originally Posted by Geoff D

Stacey, you mentioned here https://forum.blu-ray.com/showthread...2#post16564692 when talking about future DV patterns that only one of the three DV files (MEL, FEL, and ???) would actually do anything on a low latency display, what did you mean by that?

Nothing to do with FEL or MEL. The calibration stuff must be done in the display, since you can't calibrate a BD player. So the special modes only work with DV processed in the display. Basically I can disable tone mapping and profile the display itself.

[Stacey Spears]

Originally Posted by INdetectableMAN

I already know that, but the question is: What is the right thing? HDR optimize on or off? Because the difference in color beyond preserving the highlights is obvious.

On the other hand, it surprises me, and a lot, that a Panasonic UB9000 player makes a different tone assignment to its same generation OLED.

Personally I would turn it off and let the display do the tone mapping. Right now you are most likely getting double tone mapped.

[Geoff D]

Originally Posted by Stacey Spears

The OPPO does not clip the DV signal and I am pretty sure the Panasonic does not either. I don't use the HDR optimizer, could that be the cause?

The clipping on the Z9D will go away if you pull contrast down. You should not have to pull it down in DV, but you do on the Sony. LG is fine in that regard. Don't have any other DV displays to test. Something I have not done yet, but need to is:
1. Look at tone mapping ramp (white) and see where it is clipped.
2. Play chapter 2 and pull contrast down until clipping goes away.
3. Look at tone mapping ramp again and see how high it goes now.

I forget if contrast is different for DV and HDR10 on the Sony. Easy enough to copy settings if a different memory is used.

I fully understand on the pop-up help. For this disc we felt being able to change nit levels was more important this time around.

BTW, I finished the tool to create the DV binary (calibration metadata). Hoping to test it soon. Probably next week, after the holidays. Basically this bypasses Dolby's tool that generates the binary file that gets muxed into the EL. This will only work on the LG at this point. Hopefully others implement it soon.

I dare say you need to double check the Panny, and it's not just me who noticed it as Vincent Teoh picked up on it as well, it's clipping the upper end of highlight information in DV and it should be spectacularly obvious with all that heavy nittage going on in your montage, chapter 2 should show it in all its glory. When I first mentioned this DV clipping problem in the 820 thread https://forum.blu-ray.com/showthread...g#post15961702 Kris Deering also thought it might be the Optimiser but it is literally disabled when playing DV, as is all post-processing picture enhancement on the Panny, it's all greyed out.

I noticed that I could pull the contrast down on the ZD9 to bring back the clipping on chapter 2 in DV (unlike the Panny which appears to be clipping at source, adjusting the TV's contrast made no difference) but the problem is that DV luminance is already measurably below what my properly calibrated HDR10 setting is putting out, i.e. I don't want to reduce the luminance in DV any further! The contrast settings are not the same between the various picture modes. This is why I'm jonesing for some DV 0-100% greyscale window patterns so I can measure them for their light output on the ZD9.

Originally Posted by Stacey Spears

Nothing to do with FEL or MEL. The calibration stuff must be done in the display, since you can't calibrate a BD player. So the special modes only work with DV processed in the display. Basically I can disable tone mapping and profile the display itself.

I see, thanks.

[LordSummerIsle]

Fantastic news!!!