Missing Shadow Detail, nothing below 2.5% black.....now 4.1%

[Boris Blank]

Arrrgh!

Hooked up my brand new Apple TV4k (Christmas prezzie from wife) set it up, installed MrMC and watched a few bits of various 4k videos / iTunes 4k movies I have.

Noticed a lack of shadow detail immediately, went to some 4k HDR test patterns I have and saw that I was not seeing anything 2% and below. Took a piccie linked below and this is what it the test pattern looks like (I boosted the brightness level to make it more obvious)

Looks as if either the Apple or MrMC is crushing the black level. Tried changing various settings, just can't get it to show anything below 2%.

https://www.avforums.com/threads/apple- ... t-25774874

Tried the exact same test files on my Panny UB700, works perfectly, as does non-HDR test patterns on the Apple via MrMC. TV is a ZD9 which I had previously calibrated for HDR using the same Test Patterns using an Nvidia Shield and the UB700. Swapped HDMI cables, HDMI sockets on the TV etc, nope, nothing showing under 2% out of MrMC/Apple.

HDR Test Patterns are R Macciola's HDR-10 UHD test patterns available from AVSForum.

Anybody know whats going on or has similar issues? Very frustrating to say the least.

Thanks for any help/advice.

[davilla]

no see picture

[Aesthetiker]

Same here with deinterlace inactive the issue is solved! Can you confirm this?

[davilla]

Same here with deinterlace inactive the issue is solved! Can you confirm this?

fixed in testflight for 3.4.1

[Boris Blank]

no see picture

No, me neither, curse those Flying Monkees, what did they do with it!?

EDIT: Put a link to a piccie I'd posted in another forum - sorry about that.

[Boris Blank]

Got it back!

[davilla]

Which exact test pattern are you using in the example? On a slow net for a few weeks so it it would take a few days to download all his test patterns.

For HEVC/HDR we are using AVFoundation in a pass-through which skips our internal renderer. Not much we can do as we are passing through decode/rendering direct to tvOS. Demuxing is handled by our demuxer. So ours should match the native player exactly.

One thing might be related to display luminance issue that wesk05 pointed out. See viewtopic.php?f=10&t=1468&start=140#p11799 and viewtopic.php?f=10&t=1468&start=140#p11799 .

Infuse took a different approach, moved from GLES to Metal for rendering and they are doing everything via their own LUTs, which can be very difficult to get right. See Kodi's 3DLUT handling for desktop. Lots of diddling required.

[Boris Blank]

The black level patterern I used in the piccie was from R Macciola's HDR-10 UHD test patterns available from the Calibration forum on AVSForum (but you need to pay for them).

I did also make use of free UHD/HDR Test Patterns again from AVSForums from this post https://tinyurl.com/y74m4pnz (with the exact same results, nothing below 2%) and luckily the creator has uploaded the individual files to google drive and they are free, Black level pattern is 123mb zipped - https://drive.google.com/drive/folders/ ... 8_5nLcRwHA

Results from Mehanik's test pattern using MrMC, no black bars below 84, roughly 2% (brightness level boosted for clarity, there are some room refections in the 48 to 80 bars);



I also tried the Black Level patterns with Infuse which resulted in both instances in messed up results, very odd. I'll check out the Wesk05 links.

Personally, I think its the Apple, but thats pure speculation. Thats about as far as my understanding goes though, I'll leave the techy stuff to brains wayyyyyy better than mine


Many thanks Davilla.

[Boris Blank]

Just an update, looks like the problem occurs when HDR kicks in.

The image below is taken a split second before HDR kicks in. It shows the full range of black bars for a second and then they vanish resulting in 2% black crush as in the first image above. (Ignore colors, this was just a play around)

https://btcloud.bt.com/web/app/share/invite/bDU239wlYp

[Boris Blank]

Managed to attach the Apple to another 4k HDR TV set, a Samsung KS7000. Exactly the same issue, the test patterns show no bars at 2% and below.

I also tried the latest TestFlight 3.4.1 dated 18th Jan, same issue.

I've not noticed anyone else reporting this problem but its there nonetheless.