Can you point us to a publicly available test clip that shows this effect?
Only in his own opinion. He says in the comments
profile 5 is wrong only on iTunes because apple is a stupid company that doesnāt care about quality.
which means he doesnāt understand how Profile 5 is supposed to work. Yes, if their content is prepared as Profile 5 with optimised shaping and they stream it with no shaping it will be wrong (very wrong) but if they just take a HDR10 original and package it as P5 with no shaping all they are losing is the extra 1.5 bits of precision Profile 5 is supposed to add. If anyone can tell the difference between 10bits and 11.5bits, especially on a film like Barbie Iād be very surprised.
I suspect heās very good at guessing the causes of the differences between renderings that he āmeasuresā and mostly getting it wrong.
Heās not very good at explaining his methods (unless you want to spend hours wading through AVSForum posts). I doubt whether the washed-out look he claims heās getting (I canāt see it in the comparison pics) is because Apple arenāt using shaping.
Just my 2d.
You can use the clips that Kontrarian shared in this thread. Dropbox
āAliens (1986) Special Edition [UHD DV Profile 7.6] Test Clip.mkvā for example. Itās showing the same issue Iām getting when ripping a Dolby Vision Blu-ray with MakeMKV.
1 Like
I canāt reproduce this on my Philips OLED. Borders are black as it gets. Can you post a log?
Logs: https://paste.osmc.tv/fusenogije
I tried taking a screenshot but that just produced something garbled. Not supported with DV?
Not supported with 4k hevc. Thank you.
Yeah, itās very subjectiveā¦ and itās nauseating.
UHD, HDR, DV, etc. are just standards (and very fluid to say the least), and just like any other, how well it is utilized ranges.
Most UHD discs were gimmicky at first.
Upscaled, so the benefit in detail is mostly just a better codec and bitrate, HDR color grading thatās controversial, and sometimes blown out aggressive HDR, etc.
Then some people did it right; get a good scan from the original elements, do a tasteful color grade, etc.
Itās so hit and miss.
Is the Arrow version better than the Criterionā¦ etc?
And now that DV is getting ground itās DV vs HDR10, but most of the DV releases are arguably just crippling the HDR BL, etc.
It can all be rather disorientating and overwhelming for most.
And people want to inject WEB DV into UHDs that didnāt have DV etc and that introduces all kinds of new issues.
And then everyone has an opinion on how it should look.
Maybe Apple has it right and everyone else has it wrong?
For all these reasons I stopped doing encodes and just remuxed my discs.
If it doesnāt look right, wellā¦ thatās how it was in the source, no second guessing my handbrake settings or something else like that.
It used to drive me crazy!
Iād watch a movie and then something wouldnāt look right, maybe some black crush or something, and it would continue in the next dark scene as well and I totally lose emersion because I start thinking maybe I used poor encoding settings, etc.
Now when I see stuff, I notice, but I donāt dwell on it because I know thereās really nothing I can do about it and I just enjoy the movie.
You guys do amazing work dealing with all these issues.
Can you remind us which display you have? Canāt tell from the EDID.
He has an LG B7 OLED.
1 Like
You said above you have a B7. Why are you bothering with the Scope skin?
Tested it because Iām deciding which projector to get next and whether I want to bother with lens memory and things like that for constant height with different aspect ratios.
Currently use LG B7 but also have a Sony VPL-HW50 projector.
1 Like
Let me get this straight. You have grey bars on the LG, on the Sony, or both?
The Sony doesnāt do DV, so only on the LG.
Please upgrade, following the instructions at the top of this thread. Let us know if itās fixed.
The upgrade also intruduces the GUIPeakLuminance setting requested by @mceron and others having issues with bright subtitles on HDR.
1 Like
I noticed something today about DV performance from the Vero V that has me wondering whether itās managing DV correctly/differently from CE. My daily driver for DV playback of mkvās is an Ugoos/CE. I hooked up my DV test Vero and noticed that highlights had a lot more brightness compared to CE. At first I found that pleasing, and then I realised: the highlights are so bright because CE is doing a better job of rendering highlight detail. Iāve attempted to illustrate this via a shot from Harry Potter and the Deathly Hallows part 2. I have a disc-sourced HDR10 rip, processed via dovi_scripts to add a CM v4 P8 RPU. There is an underground labyrinth sequence where the gang are on a sort of rollercoaster car which has a very bright headlamp. When I watch this in HDR10 on my Panny HZ980, there is next to no detail in the headlamp. Watching the dovi_scripts conversion via CE (mainline or latest CPM build), I get to see that there is a design detail of concentric rings within the lamp. However, when I watch this on the Vero, the design detail in the headlamp is largely missing, giving me pause for thought as to whether DV from the Vero is correct?
I dug out an old DSLR to take the following shots that give an indication of the issue. The first is from the Ugoos, the concentric rings in the headlamp clearly visible:
Here is a shot giving an indication of what the headlamp looks like via the Vero, the concentric rings have gone. The overall brightness level in the second shot is brighter, but this is taken during a moment when the light in the top of the frame is coming on and I may not have matched exactly. Iāll investigate further whether the overall brightness profile of DV content is elevated vs CE, in addition to highlight detail.
EDIT: just compared to the ATV/Infuse, Infuse is giving detailed rendering of the headlamp in DV, essentially same as CE. Does look to me like the entire brightness profile of the shot is elevated on the Vero vs the other platforms.
further to previous post, there is an even more glaring example in Goblet of Fire. Just after 9mins in, there is a shot of Robert Hardy in the glare of floodlights. Watching this in HDR10, his face is a burnt-out over-exposure on my Panny HZ980 (even in filmmaker mode - my TV canāt deal with the highlight brightness in this film).
The DV conversion played via the latest test build on the Vero still shows him with a burnt-out face, even though the TV and the AVR register a DV signal. Playing the DV version via Infuse or CE, Hardyās face is rendered far more sympathetically - itās an example of how comparing HDR10 vs DV for a given display can result in the HDR10 version looking unwatchable.
Looks good with the Scope skin now. 
2 Likes
Based on those two pics, the difference is quite, er, subtle. Would most people notice it without a side-by-side comparison? Not trying to diss your report, thatās a genuine question to get a better idea of the scale of the difference.
As it happens, Iāve spent a lot of today looking into this, as also reported by HDRpotato. The only device I can compare with and actually get meaningful measurements from is a well-known cheap s905x4-J box that CE refuse to support but running CE-ng just fine.
If I measure the output of that box, itās actually brighter than the output from Vero so Iām a bit stumped. Maybe it would be different if I ran CE-ne with the dovi.ko that came with it. Iāll have a close look at your two pics to see if thereās a systemmatic difference we can use.
If you have a moment, can you clip that bit from Goblet of Fire for us? Say, 30 seconds each side of the bit where itās āglaringā?
And, just checking, your TV is set to DV mode 1, right?
it was a struggle to get shots that fairly represent what I can see. The in-play difference is obvious, via the Vero the headlamp is a blinding glob of light, via the other platforms you see the detail. But the Goblet of Fire sequence is far more obvious in highlighting an apparent issue with the Vero in how metadata is being managed. The TV is indeed running in its display-preferred DV mode 1.
Iāll snip a clip from GoF and send you a link.
1 Like
We donāt manage metadata, we just pass it through with P7 and P8 streams. For the GoF clip, is that also over-cooked if played as HDR10?