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

[Stacey Spears]

Originally Posted by gregwwright

This is the disk

https://www.amazon.co.uk/Deer-Hunter-Blu-ray-Robert-Niro/dp/B07MZJGNCP/ref=tmm_frk_swatch_0?_encoding=UTF8&qid=1621181484 &sr=8-3

Thank you Greg. Is this disc region locked? Is here a US release or is the UK the only one? As long as it is not region locked, it should not be a problem.

[Fendergopher]

Originally Posted by Stacey Spears

Thank you Greg. Is this disc region locked? Is here a US release or is the UK the only one? As long as it is not region locked, it should not be a problem.

I don't think there are any region locked UHD discs yet, are there? The US UHD for Deer Hunter is by Shout Factory and has a different encode, so who knows if it has the exact same HDR setup.

[EddieLarkin]

Do LGs still crush blacks, to try and hide their near black issues?

I watched the SC Deer Hunter UHD on a calibrated E6. It was in a pitch black room so the brightness level wasn't too bad but there were some scenes (specifically the first roulette scene) that was indeed oppressively dark.

I then watched it on a Panny HZ2000 a year later and it wasn't anywhere near as dim and a ton of shadow detail was revealed that was hidden before.

Similar thing happened with the original Halloween on UHD. The scene where Myers approaches from behind Laurie at the top of the stairs after she sends the kids off, on the LG you can barely see him which I thought was very odd as I knew he was perfectly visible on previous editions. But same disc, same player, but into the HZ2000? All was revealed again.

Whether it's a cutting off of the black floor to hide near black issues, or just plain old shit tone mapping, I personally wouldn't touch an LG set again with a barge pole

[Stacey Spears]

Originally Posted by EddieLarkin

Do LGs still crush blacks, to try and hide their near black issues?

I watched the SC Deer Hunter UHD on a calibrated E6. It was in a pitch black room so the brightness level wasn't too bad but there were some scenes (specifically the first roulette scene) that was indeed oppressively dark.

I then watched it on a Panny HZ2000 a year later and it wasn't anywhere near as dim and a ton of shadow detail was revealed that was hidden before.

Similar thing happened with the original Halloween on UHD. The scene where Myers approaches from behind Laurie at the top of the stairs after she sends the kids off, on the LG you can barely see him which I thought was very odd as I knew he was perfectly visible on previous editions. But same disc, same player, but into the HZ2000? All was revealed again.

Whether it's a cutting off of the black floor to hide near black issues, or just plain old shit tone mapping, I personally wouldn't touch an LG set again with a barge pole

After Calman autocal, on the C9, LG modifies the 1D LUT near black to try and correct it. It does not do a great job given the variance among panels.

On the CX, and C1, you adjust near black using the shadow detail control that was introduced in Calman 2020 by looking at the Spears & Munsil dynamic range low pattern. We provided Portrait with SDR, HDR10 and Dolby Vision mp4 files to share with their customers since the current disc just has SDR and HDR10.

By doing the manual adjustment, you get the proper amount of detail coming out of black. You have to do this for SDR, HDR10 and Dolby Vision. You also need to use a loupe, in a dark room, to get the best results.

HDR10, out of the box, shows way too much shadow detail on the LGs.

[SeeMoreDigital]

Originally Posted by EddieLarkin

Do LGs still crush blacks, to try and hide their near black issues?

With reference to black crush, what do you guys think about this?

https://www.youtube.com/watch?v=mT-N...ature=youtu.be

[Geoff D]

Originally Posted by EddieLarkin

Do LGs still crush blacks, to try and hide their near black issues?

I watched the SC Deer Hunter UHD on a calibrated E6. It was in a pitch black room so the brightness level wasn't too bad but there were some scenes (specifically the first roulette scene) that was indeed oppressively dark.

I then watched it on a Panny HZ2000 a year later and it wasn't anywhere near as dim and a ton of shadow detail was revealed that was hidden before.

Similar thing happened with the original Halloween on UHD. The scene where Myers approaches from behind Laurie at the top of the stairs after she sends the kids off, on the LG you can barely see him which I thought was very odd as I knew he was perfectly visible on previous editions. But same disc, same player, but into the HZ2000? All was revealed again.

Whether it's a cutting off of the black floor to hide near black issues, or just plain old shit tone mapping, I personally wouldn't touch an LG set again with a barge pole

People keep mentioning they’ve had it “calibrated” when they protest at how their awesome TV (usually an older LG OLED, lol) couldn’t possibly be choking on tone mapping or nuking black levels or dimming too much with ASBL. Not a dig at you Eddie, you saw the light (or rather, the dark), I’m saying that “calibration” alone isn’t going to be able to overcome most of those issues and it’s a pain trying to convey that to people.

[gregwwright]

Originally Posted by SeeMoreDigital

With reference to black crush, what do you guys think about this?

Im trying not to hijack this thread with Oled/Calibration detours, but this is an SDR adjustment. My comments about the StudioCanal Deer Hunter disk was an aside about the DV look of the 4K UHD disk.

Specifically, the context is that this disk has HDR10 menus (and other screens) before launching the DV main title. But to my eyes the look of the film is so off Im not sure how useful it actually is!

However, I take @Geoff_D's point, I should try turning off ABSL.

And again, the 1080p version of the disc in the package looks convincing, so Im confused why the DV disc looks so bad(dark).

Regards
Greg

[Stacey Spears]

Originally Posted by gregwwright

Im trying not to hijack this thread with Oled/Calibration detours, but this is an SDR adjustment. My comments about the StudioCanal Deer Hunter disk was an aside about the DV look of the 4K UHD disk.

Specifically, the context is that this disk has HDR10 menus (and other screens) before launching the DV main title. But to my eyes the look of the film is so off Im not sure how useful it actually is!

However, I take @Geoff_D's point, I should try turning off ABSL.

And again, the 1080p version of the disc in the package looks convincing, so Im confused why the DV disc looks so bad(dark).

Regards
Greg

If you disable DV and play as HDR10, is it better, worse or the same as HDR10 with what you see crushed?

[gregwwright]

Originally Posted by Stacey Spears

If you disable DV and play as HDR10, is it better, worse or the same as HDR10 with what you see crushed?

It looks better in HDR10. Although it looks even better if you turn Dynamic Tone Mapping on which I dont normally do.

On my C9, for DV, to get a decent level of brightness, I need to use Cinema Home DV preset which skews the colours, so again I prefer to avoid that mode.

Obviously Im a home user so dont have 2 screens side by side so it is tricky doing this.

[Stacey Spears]

Originally Posted by gregwwright

It looks better in HDR10. Although it looks even better if you turn Dynamic Tone Mapping on which I dont normally do.

On my C9, for DV, to get a decent level of brightness, I need to use Cinema Home DV preset which skews the colours, so again I prefer to avoid that mode.

Obviously Im a home user so dont have 2 screens side by side so it is tricky doing this.

Thank you for taking the time to run the tests. Normally you see more in DV than HDR10, so this is really interesting.

DTM actually raises levels, so that could be one reason that DTM on is helping here, it is lifting shadows.

In our up coming skin tone content, there is a colorchecker chart in frame with each model. The lower left chip on the chart should measure ~132-134 nits. It does with DTM off and in Dolby Vision. With DTM on, it measures ~202 nits. This is what I meant by raising levels.

Originally DV was darker than HDR10. This could also be what you are encountering. Dolby will bend over backwards to not let any highlight appear clipped / crushed. Sometimes, in a less than ideal way. In our skin tone clips, there is an ~800 nit ping on the bolt that holds the colorchecker chart. A similar ping can be found on jewelry, such as earrings. The rest of the content is under 200 nits. These pings caused the Dolby image to be darker overall. The tool the created the MaxCLL value for HDR10 had ignored these outlier pings while Dolby did not. What we did was to use the HDR10 MaxCLL values in place of the default values Dolby calculated. This change resulted in Dolby matching the HDR10 levels overall. The bolt and earrings look fine too.

If you play the 10K nits HDR10 and DV montages on our disc and look at the fence in snow shot. The HDR10 version has really dark wood on the fence post. In DV, the wood is lighter like it should be. This is a more extreme example because the higher the nit metadata, often the darker HDR10 becomes to ensure it preserves all highlight detail up to 10K. I assume Deer Hunter is in the 1000-4000 mastering display range. If you pull up the info screen, what does it report for the HDR metadata?

I believe Cinema home uses a light sensor to compensate for the ambient light in the room. This most likely raises shadows and mid-tones. The C9 was pre-Dolby IQ, but it has something similar that is performing this action.

I wish I had two screens side-by-side, especially now that I have the Envy and want to compare Envy vs. TV/Dolby tone mapping and two screens would make it much easier!

[gregwwright]

Originally Posted by Stacey Spears

I assume Deer Hunter is in the 1000-4000 mastering display range. If you pull up the info screen, what does it report for the HDR metadata?

I re-ran the test for my sanity.

In HDR10 mode, info reports Max Luminance 1000 nits, Min Luminance 0.0001 nits
If DTM is off, it looks too dark, but turning DTM on, rescues the picture.

In DV mode, picture is way too dark unless you use Cinema Home, its comparable to HDR10 with DTM.

I find the mastering of this disk strange. The BluRay 1080p disk looks good, yet playing the UHD disk looks IMHO terrible. Yet no reviews Ive read have really mentioned this.

I wonder who signed the grading off on this, it cant be the director?

Anyway, this was all a detour from the fact the menus are definitely HDR10 and the feature is DV which is an edge case you were interested in!

Hope that helps
Greg

[KC-Technerd]

Originally Posted by Stacey Spears

I believe Cinema home uses a light sensor to compensate for the ambient light in the room. This most likely raises shadows and mid-tones. The C9 was pre-Dolby IQ, but it has something similar that is performing this action.

I believe the image only changes in response to the light sensor when the "AI Brightness" setting is turned on (on the 9 series OLEDs anyway). It's the last setting under "Picture Options" and I think it is only available in the menu when the Cinema Home picture modes (HDR/DV) are selected.

[Stacey Spears]

Originally Posted by gregwwright

I re-ran the test for my sanity.

In HDR10 mode, info reports Max Luminance 1000 nits, Min Luminance 0.0001 nits
If DTM is off, it looks too dark, but turning DTM on, rescues the picture.

In DV mode, picture is way too dark unless you use Cinema Home, its comparable to HDR10 with DTM.

I find the mastering of this disk strange. The BluRay 1080p disk looks good, yet playing the UHD disk looks IMHO terrible. Yet no reviews Ive read have really mentioned this.

I wonder who signed the grading off on this, it cant be the director?

Anyway, this was all a detour from the fact the menus are definitely HDR10 and the feature is DV which is an edge case you were interested in!

Hope that helps
Greg

SDR and HDR should look different as they are different grades.

Thank you for all of the details Greg, I appreciate it.

DTM does raise levels it should not, but it gives you a more pleasing picture in this situation, which is what matters. On our disc, it is raising ~134 nits up to ~202 nits.

[Stacey Spears]

Just an update. We may end up moving to 4-discs. It appears that the file size has ballooned because of the M2Ts container, minimum disc block size and number of files. What should take ~51 GB is consuming ~91 GB of disc space.

Right now we are trying to figure out if we should cut some montage versions and merge what does not fit on disc 1 onto the disc with the montage videos keeping a 3-disc set or just moving to 4-discs and not cutting anything.

[Geoff D]

Originally Posted by Stacey Spears

Just an update. We may end up moving to 4-discs. It appears that the file size has ballooned because of the M2Ts container, minimum disc block size and number of files. What should take ~51 GB is consuming ~91 GB of disc space.

Right now we are trying to figure out if we should cut some montage versions and merge what does not fit on disc 1 onto the disc with the montage videos keeping a 3-disc set or just moving to 4-discs and not cutting anything.

I can still recall when this was just gonna be one extra disc!

[Stacey Spears]

Originally Posted by Geoff D

I can still recall when this was just gonna be one extra disc!

The motion test patterns are consuming much of the disc space.

23.98p Motion - ~20 GB
59.94p Motion - ~30 GB
Skin Tones - ~20 GB

We can, of course, lower the bitrate a lot and shrink them down, but the goal is the highest quality possible. For the 24p stuff, we can force QP 0 in HEVC for the best quality possible. We can't do this on the 59.94, so the peak is set to 98 Mbps and average at 97 Mbps. This results in the QP being 4 or less on the stock ticker pattern, for example. Including at every nit level for HDR10 really eats up the space. Had we just done a single nit level, then we could probably get a way with one disc for everything.

For those that don't know all of the details on compression, the QP range is 0-51 with 0 being highest quality and 51 being lowest. When you set a bitrate, it simply adjusts the QP up and down to get as close to your bitrate as possible.

The Atmos tones are probably ~20 GB as well. The Atmos content is on the SDR disc.

I think I posted these stats earlier. Current disc has ~1400 HEVC files on it. Disc 1, before the split, has ~4500 HEVC files.

All of the content is now encoded. Once we start QCing the discs, we may have to re-encode / generate some patterns. That part will be quick. Given how much work it is to swap out files, we will wait and deliver any re-encodes all at once. It is specifically the Dolby Vision stuff.

Speaking of Dolby Vision, so far we have found three issues with it on BD. This has led me to buy the madVR Envy Extreme and bypass Dolby all together.

1. Low Latency is broken on BD. This we have reported before. We are including a hidden test pattern called Dolby Vision Quality Check, targeted at the press, to report on this issue. You can see a sample of the pattern in a recent video by Vincent on his HDTVTest YouTube channel. It shows a big red X when DV is not working correctly.
2. DV on BD is using 709 chroma upsampling. Dolby's tools create proper 2020 downsample, so you get a .5 pixel delay between luma and chroma in the vertical direction.
3. DV seems to be applying some type of horizontal chroma filter on UHD content. Does not seem to appear at HD resolution. This results in a total loss of high frequency chroma detail. Here is the chroma zone plate, the left and right sides of the image should not be luma only. Not sure if this is a player or display side problem yet. Hopefully I will know before the disc ships.

4. The Z9D has some unusual artifacts in Dolby Vision that I have not seen on other displays, which shows up on the Hue Shift test pattern. Looks almost like compression artifacts, it is bizarre.

We have added a rattle test for audio. Two sweeps, one 500 Hz to 200 Hz and 200 Hz to 15 Hz. The sweeps are center channel only in Atmos. The assumption is that the center channel is the most likely one set to small, so it will use the subwoofer.

[Geoff D]

QP = Quantization Parameter. As Stacey says, it’s a setting that dictates how transparent to the source your encode will be. The higher number the further away from the source you get and it throttles back the usual temporal transients like grain to compensate. That’s how I understand it anyway.

[Stacey Spears]

Originally Posted by Geoff D

People keep mentioning they’ve had it “calibrated” when they protest at how their awesome TV (usually an older LG OLED, lol) couldn’t possibly be choking on tone mapping or nuking black levels or dimming too much with ASBL. Not a dig at you Eddie, you saw the light (or rather, the dark), I’m saying that “calibration” alone isn’t going to be able to overcome most of those issues and it’s a pain trying to convey that to people.

Each generation OLED gets better. With the LG X and 1 series, you can actually adjust near black in Calman. It does this by modifying the 1D LUT. You visually look at our dynamic range low pattern and tweak until you can see level 65 flicker (using a loupe to see the white subpixel flicker on and off)

Out of the box, HDR10 is too bright near black and DV is crushed. I don't recall where SDR lands, but it needs tweaking too. You can make them all match pretty well when done.

This can't be adjusted on the 9, or earlier models. On the 9, LG applies a correction to the 1D LUT after its loaded, but its not ideal. This is why on the X and 1 they added the ability for the user to adjust it manually by looking at a pattern.

For HDR, you really need a minimum of a 33^3 3D LUT. 17^3 is simply not enough for HDR. You really want 17^3 just for the first 100 cd/m². Then for both of these, you also want Tetrahedral and not trilinear. For trilinear, you need even larger LUTs. madVR Envy uses a 256^3. Of course, this is a generalized statement. The less you do with a 3D LUT, the better. Ideally a true linear display using a 3x3 matrix is the best way to go. DLP may be the only tech that works that way.

[TbeRw01]

Originally Posted by Stacey Spears

Speaking of Dolby Vision, so far we have found three issues with it on BD. This has led me to buy the madVR Envy Extreme and bypass Dolby all together.

1. Low Latency is broken on BD. This we have reported before. We are including a hidden test pattern called Dolby Vision Quality Check, targeted at the press, to report on this issue. You can see a sample of the pattern in a recent video by Vincent on his HDTVTest YouTube channel. It shows a big red X when DV is not working correctly.
2. DV on BD is using 709 chroma upsampling. Dolby's tools create proper 2020 downsample, so you get a .5 pixel delay between luma and chroma in the vertical direction.
3. DV seems to be applying some type of horizontal chroma filter on UHD content. Does not seem to appear at HD resolution. This results in a total loss of high frequency chroma detail. Here is the chroma zone plate, the left and right sides of the image should not be luma only. Not sure if this is a player or display side problem yet. Hopefully I will know before the disc ships.

So is DV better than HDR10 even with those bugs?
I Always thought DV looked worse than HDR10 on my calibrated LG C8 in a pitch-black room, especially for the black levels; DV blacks look raised and crushed on my end.

[Geoff D]

Originally Posted by TbeRw01

So is DV better than HDR10 even with those bugs?
I Always thought DV looked worse than HDR10 on my calibrated LG C8 in a pitch-black room, especially for the black levels; DV blacks look raised and crushed on my end.

Even with all my eagle-eyedness there are very few artefacts I can see in motion with LLDV when compared to HDR10 and overall image quality is virtually identical aside from slightly dimmer brightness and slightly reduced colour saturation in DV. The only thing I’ve ever seen that impacted on real-world viewing and not torture-test material like Stacey's montage is the opening credits to the Shinning, as the blue letters go up the screen they’re filled with weird digital crud where the improper chroma is making the background bleed into the lettering, while the HDR10 is perfick, no chroma bleed.

Even with those issues I’d never, ever give up DV on the FEL titles that make a dreadfully encoded HDR10 layer look somewhat watchable, there’s not an outboard processor on the planet that can fix what StudioCanal did to the base layer of The Fog. Heck, even on Stacey’s montage both the MEL and FEL DV versions clean up some chroma noise that’s visible in some shots in HDR10.