thank you i thought you said there no 1080p release

very cool! thanks

Did they just forget to fix the white point in the 1080p bdmv for certain scenes when converting down to bt.709? Or is white / red just suppose to not be that color anymore?   2 more comparisons. https://slow.pics/c/O1fSr9rZ For what it's worth, your web release of 3.0 + 1.0 shows the corruption as a dark red, not pink. So I'm making the assumption that the color changes in the BD aren't intended. Worst case scenario, someone takes the 4k bt.2020 and tries a conversion to 1080p bt.709. 

This is called the "other transfer" used for the video. Still, the first edition 3.33 was released in 2012, and 3.333 in 2021. I have not made any color changes. The 1080 version of bd edition 3.333 does indeed have these colors. The insane richness of the red, which turns into ruby ​​in the 4k version with bt 2020. This is the technique they used back in NGE.

I see. Just very odd and inconsistent, unless the 4th movie will also be having its color gamut adjusted in the BD release for it. Makes it seem like an on-disk conversion oversight that leaves the 3rd movie with different colors than the other three.

I agree, but the disc edition of the 4 movie will contain changes, as well as the first 3 movies (3.0-3.33-3.333). Therefore version 3.3 + 1.1 (?) May have a similar hue. In any case, we will find out only after its release.

I decided to dive a little deeper into the problem since someone said the 4k encode looked fine to them. It turns out, if you are using an earlier version of mpv, you might have a build before they added "better" gamut clipping. Or you are using a media player that does not try to compensate for out of range colors. https://github.com/mpv-player/mpv/commit/dc24a437fb9ba2485bc8c5866542bd36254410d4 https://github.com/mpv-player/mpv/issues/9071 By default the 4k encode (bt.2020) has its gamut de-saturated to fit inside bt.709 color space. This causes the pinkish hues and off-white (not d65).  If you add ``gamut-clipping=no`` to the config, you get the more "normal" looking reds (similar to old 3.33) with the bt.2020 4k version.  (3.33 1080p beatrice left, 3.333 1080p beatrice top-right, 3.333 4k v2 beatrice bottom-right) Some colors are a little different but it's much more in line with the older movie version. Unfortunately the white point in the 4k version is still completely off though with clipping turned off. The 1080p encode is unaffected either way and still has the pink hues / off-white, which makes me believe they didn't actually do a bt.2020 -> bt.709 conversion for the "other transfer", just clamped and flagged it as such. But as noted, there's no way to really know if the color change is intended until the 3.3 + 1.1 BD is out I guess.

@Mabby While not definitive, I can say that the 4KUHD converted down to bt709 using NVencC64 matches the colors perfectly for the Amazon Web release of 3.0+1.01. Specifically I took the scene of Misato telling Shinji he doesn't need to do anything, since it's a scene with deep reds that's shared between both 3.333 and 3.0+1.01. I took the actual 4K disc data, ran it through NVencC64 with "--vpp-colorspace matrix=bt2020nc:bt709,colorprim=bt2020:bt709,transfer=bt709:bt709" and the resulting 1080p video had reds that were spot-on for the same scene as depicted in 3.0+1.01. The 1080p bluray, meanwhile, has totally bizzare colors that just look wrong in every single possible way, and I feel sorry for people who bought that version. As a side note, I found out from the dev of NVencC64 that the "transfer=bt709:bt709" part is required to get the colors right for this because for some reason, this "SDR" version reports having bt2020 10-bit transfer characteristics but in fact those matricies contain bt709 data. Very odd, but it explained why the colors were totally off if I tried to let the encoder automatically convert to bt709. Here's my screenshots: https://cdn.discordapp.com/attachments/163412644013932545/884462519090040922/testreds.png Upper-left is NVencC64, upper-right is my encode using the 1080p bluray disc, bottom-left is my encode using the HDRTools plugin for AVIsynth (it got close but not quite right, though I'll admit I like the color of Asuka's plugsuit better in this version, though it's clearly incorrect), and bottom-right as it says is 3.0+1.01.

@nomakewan Thanks for sharing your corroborated findings! The reds in the 4KUHD do look like they match significantly better when ``gamut-clipping`` is off so it doesn't surprise me it probably converts down fine. However I still notice the white point is not a perfect D65 hue for known previous pure white scenes (i.e. 01:35:02 / 01:35:22). The matrices containing bt.709 data might explain why it's off and may just have to be manually corrected. Are you noticing the same issue with whites being off in your 4k -> 1080p encode? I did say it may be best just to take the 4k BD and encode down instead of using the funky transfer. :)

>this “SDR” version reports having bt2020 10-bit transfer characteristics but in fact those matricies contain bt709 data. The number of colors that the bt.709 matrix contains is part of the bt.2020 matrix. Because bt.2020 = bt.709 + "more shades". So most likely 1080 bd was created by trimming the number of colors to match the bt.709 level, rather than converting bt.2020 to bt.709.

@Mabby: Yes, the preview for the next movie is a "greenish" color. Here's a screenshot: https://i.imgur.com/QcnYRQu.png @Jensen: Well yes, but my point is that if you attempt to convert from bt2020 to bt709 by telling the encoder that the source material's metadata is correct, the result will be absolutely horrendous. Here is an example of what it looks like if you do that: https://i.imgur.com/lkvDVZA.png Everything is way blown out to the point where blacks appear greenish. However, if you tell the encoder that everything is bt2020 -except- the transfer characteristics, and instead tell it that the source and target transfer characteristics are both bt709, then everything works out peachy-keen. Also, in this release, the subs are one frame late across the entire film. I think I mentioned that in the last comments section, actually, that if you wanted to use my script you'd need to adjust them to be one frame earlier.

@nomakewan Invalid conversion occurs because the program is applying tonemapping for transfer 2086 (hdr 2020), and the source is sdr. Also, what does "one frame" mean? 1/24 second? I have looked at several places while creating the release and I have not been able to notice any nasty lags or anything like that.

O.o

@Jensen: If you actually go frame-by-frame, you'll see that on scene transitions, the subs will "hang" for one frame rather than disappearing on the last frame of the previous scene as normal. If you use Aegisub, you can load your video, then change the mode to frames instead of timestamps, and then use the Timing -> Shift Times function to shift all rows backward by 1 frame. Or yes, you can use a tool like mkvtoolnix to adjust the delay of the sub track negative by 0.04 seconds and that will probably do the same thing. Also the program isn't trying to do HDR as far as I can tell; the valid transfer characteristic that it detects from the source metadata is BT.2020 (10-bit). There are separate settings in this application for doing HDR conversion (hdr2sdr), which are disabled unless explicitly called in the command line.

@nomakewan I'm talking about the parameter "transfer". It might be 2084 (hdr) by default, not 2020 (sdr).

Right, and I'm saying that I don't think that's correct. Here are the full options; if you search for "vpp-colorspace" it'll skip to the section we're interested in. In a grey box it lists all the valid transfer characteristics for input and output data. Under those is the hdr2sdr functionality. Note also that smpte2084 is its own transfer characteristic. https://github.com/rigaya/NVEnc/blob/master/NVEncC_Options.en.md

Does anybody know what SBMV (Super Bit Mapping) Extended is? That is applicable to the standard BD of 3.333 and is also on the original 3.33 BD as well (I don't think 1.11 and 2.22 have it since their booklets do not state that information on the last page of the booklets like 3.33 and 3.333). I also have the 9th Anniversary Love Live! Forever Edition and Nijigasaki and those BDs show SBMV logo as well. I wonder if its some type of proprietary technology on certain Japan Sony players or something similar to MGVC (Master Grade Video Coding) for Panasonic? However MGVC discs have the KDM folder with additional color information (up to 12bit) on root of the disc in [.bin] format. I have the Fafner Ultimate BD-BOX and all three thus far Fafner: The Beyond BDs and all of those are MGVC discs so I went though hell and back and managed to import a Japan Panasonic DP UB9000-K. So far I've tested a Beyond disc which outputs at 4:4:4 12bit and Right of Left which outputs at 4:2:2 12bit. I can't help but wonder if SMBV Extended is something similar, however I plan to never go through the hassle of obtaining another specialized Japan BD player.

https://www.sony.com/electronics/support/articles/00109726 The Sony technology fo dithering video.

@Jensen Thanks for the information link. Once upon a time I tried to look it up but couldn't find it for some reason. I didn't bother try to look up again since I figured I'd get the same results as before. My first UHD player is in fact a Sony UBP-X800. I'll have to check the manual but from what I remember there's nothing in the settings for it. I've never seen any product here in the US with the SBMV logo on it so I was assuming its another exclusive Japan technology. Plus I got that player years ago and don't know if it's listed on the packaging or anything like that. I feel kinda silly for not being able to find that information myself but thanks!

@DamianV8501 I don't know why Sony describes this technology as "you need a special bd player", but in fact it is a filter through which the video is run before writing to disk and dither is performed there to prevent banding. MGVC on the contrary, it requires special devices that support this technology, since this is a way from the Japanese to record video with a color depth of 10 and 12 bits on a standart bd disc. This technology is based on bd3d, just the information about 3d is replaced with information about the color.

@Jensen Very interesting. I don't suppose there is actually a possible way to directly use the MGVC .bin files with the extra color information. I did a full disc backup with MakeMKV and my [BDROM] 蒼穹のファフナー 01 is 90.4GB with the KDM folder equaling 51.2GB and BDMV folder equaling 39.2GB. At least to my knowledge, everyone trying to figure it out on the MakeMKV forums gave up. I felt like I was missing out on what my Fafner discs had to offer and since there is no possible way without a special Japan Panasonic player (that I know of) which is why I ended up importing a Japan Panasonic DP UB9000-K. I will say it does indeed look very nice and didn't notice banding on the bird at the beginning of The Beyond OP1 but I just need to find time to watch, experiment, and research more with it. I feel like its quite amazing to add color information using bd3d method however it sucks that they wanted to keep it proprietary. It could've been a good band-aid fix for the standardized 8bit color for Blu-ray.

>At least to my knowledge, everyone trying to figure it out on the MakeMKV forums gave up. Naturally. All the "magic" happens in the analog filter of the BD player with the support of this technology.

True enough. I mean, what I was getting at was I guess it could potentially be possible to write software to handle it from it being reverse engineered and having capable hardware. Probably won't happen though.

The MGVC hack is somewhat similar to how Dolby Vision is handled for 4K UHD. The 4K disc has a second video track that is in 1080p with the Dolby Vision data. This can allow for up to 12-bit colour similar to MGVC but with better dynamic HDR.

Looks great! Still, I hope 1.11 and 2.22 will be getting 4KBD Conversion sometime! Same with Shin Evangelion!!! THANK YOU!!!!!

SBMV is mostly used by Sony Pictures. I saw that all of the Breaking Bad and Better Call Saul discs have them. And also the Sam Raimi Spider-Man Trilogy. I noticed that it's not a thing on 4K UHD, since there are no mentions of SBMV on the Sam Raimi Spider-Man Trilogy 4K-only set. It is a thing in 3D Blu-rays, however, as The Official 2010 FIFA World Cup Film in 3D does have SBMV. It's a 720p disc in HFR 60fps, but it's 3D-only. P.S: Checking the back covers of the Breaking Bad and Better Call Saul Blu-rays and I saw that the Canadian rating for them is only PG. That's strange considering both shows are rated 18 in the UK.

Seeing from those that have the BD of 4 the color issue is back but with a twist. If the translation is correct someone said that the reds in the 4k version are more orange than in the 1080p version.