madVR - high quality video renderer (GPU assisted)

[Klaus1189]

Do you have a s90c?

[flossy_cake]

an s90c the most broken TV on this planned.

a TV so intoxicated that it changes the primaries and secondaries if you change the gamma!

Damn I was looking at S95C just now and actually it seems pretty good on paper, perfect viewing angles and strong brightness like G4 but for much cheaper. But then I see Vincent talking about the low light skin tones not being as saturated and a little lifted due to an uncorrectable gamma boost at 10% stimulus and I was put off cause this is the issue I have been battling for years on plasma on LCD and I want to be finally free of that issue (the Panasonic OLED I had briefly had very good low light skin tones, good saturation and hue stability across the luminance range - I want that!).

PlasmaTVForGaming says S95C has a low latency 60fps->120fps interpolation mode available in game mode which would could be better than 60hz BFI. Yet to be tested. Some people are having issues with the one connect box HDMI signal dropouts.

So I want the G4 but it's way too expensive right now, maybe the upcoming black friday thing will help. C4 just doesn't have the viewing angle for my needs. I also want that heatsink on the G4 cause I want 10 years out of this TV.

[flossy_cake]

Here are a couple of test images that I use to evaluate skin tone stability on low light scenes. Check for green/bluish tints in certain luminance ranges of the skin

https://a.l3n.co/i/nFY1xP.png
https://c.l3n.co/i/nFYteZ.png

So many plasmas and LCDs look like crap on these types of scenes, but OLED much better stability of hue/saturation.

[huhn]

and i use a colorimeter. the screen has fake good results thanks to De 00 not caring about the error this screen is proven to and everyone using it. if older recommended calibration settings you ended up very high itp for one of the skin colors and primaries&secondaries where off.

you can archive good result maybe not with the scaler...
but the calibration controls are subpar and do stupid things only has 2 gamma option which have nothing todo with reality 2.2 and bt 1886 both are neither 2.2 nor 2.4 and change depending on the picture mode because why wouldn't they... if you want 444 in SDR this screen is a total disaster.

[Klaus1189]

Do you have a s90c?

So do you have one?

if you want 444 in SDR this screen is a total disaster.

Why?

[huhn]

bt 1886 is closer to 2.1 then 2.4 even at the max change to bt 1886 -3 (i don't remember if it is + or minus) it doesn't read 2.4 so it can't do gamma 2.4 in PC mode using a CMS. so that mode is now dead to me outside of a 3D LUT. 22 point to maybe fix that is blocked why wouldn't it...

game mode can do 2.4 but this will change the primaries and secondaries making calibration a nightmare if you have to redo all primaries settings when you change the gamma.
if you use report bt 2020 it just ignores it so good luck with that...
fmm works like a charms and subsamples like the sub mode it is.

yes i own a s90c. replaced my LG CX because the batteries where empty mistakes have been made.

[whitestar999]

an s90c the most broken TV on this planned.

a TV so intoxicated that it changes the primaries and secondaries if you change the gamma!

And yet it has a very good rating on rtings which is what many consider before buying TV.
https://www.rtings.com/tv/reviews/samsung/s90c-oled

[flossy_cake]

The thing is I've been burned by glowing reviews more than once, so those reviews are just more of a general guideline for me personally. There are usually a bunch of quirks that don't get mentioned in the reviews. And I think there are regional variances with not insignificant differences.

What huhn said about fake good results with colorimeter is true in my experience too (not with Samsung OLED specifically, just in general). On my plasma if I measure colour patches it measures well but looks bad on actual content. I have to use wrong picture settings to get a "right" image. It wasn't like that for the Panasonic OLED — that one had perfect colours out of the box, all content looked good in terms of colour accuracy. I would never get a scene where skin tones looked wrong. So I didn't even bother measuring it.

Panasonic plasma on the other hand always had terrible colour on low light skin tones, had many of them, low end and high end models, always rendered those scenes poorly, and always in the same way with blue green cyan tints in the low light skin tones.

[flossy_cake]

I should say this: it depends on what content you watch. If you watch mostly sport, or news, or reality TV shows, or sitcoms, or videogames, most of the time everything is brightly lit and you won't see what's really going on in the low end. But for heavy drama and movies with a lot of dimly lit scenes it becomes apparent when the stability of hue&saturation across the luminance range is not right, and it's mostly noticeable on skin tones because our vision is adapted to analysing faces.

Some of it has to do with lighting the scene with incandescent (tungsten) bulbs + celluloid, vs LED + digital cameras. So even if your TV has colour problems the content itself may never use those ranges of colours anyway if it was all nicely lit by the director with tungsten bulbs and the camera was set up a certain way.

The worst TV I had by far was a Samsung VA LED 6400 somethingorother from about 2013, the skin tones on that display were so bad, I could not believe they would expect people to watch any movies on that thing, all just desaturated green and blue tints on low light scenes. Probably most buyers just use it for news and sport, morning chat shows or whatever where its all brightly lit and the issues aren't apparent. But try to watch a movie like Mystic River on it and it won't be watchable. That movie is the best torture test cause they colour graded it to be as desaturated as possible , so if your TV isn't good in the low end it will look like trash but hey the TV measures good on grey scale and 75% colour bars so it's the movies fault right? :rolleyes:

I recall there was an OLED with metamerism or something where it measured good but looked bad.

Edit: and don't get me started on LED bulbs, those ones also measure 90% accurate colour rendering (tungsten being the 100% reference point) but if you look at a swatch of reds it's on only 30% accurate on those reds so skin tones, wood, red apples, meat etc. all look weird orange tint. That's why I still use 100% CRI halogens for overhead lighting in kitchens and bathrooms , cause when I look at food or my face in the mirror I don't want to see desaturated crap .

[huhn]

And yet it has a very good rating on rtings which is what many consider before buying TV.
https://www.rtings.com/tv/reviews/samsung/s90c-oled

yes and the stuff they show is not faked at all. the words they say are obviously wrong they literally show a de 00 of over 3 and say there is no issue. and if they switch to itp and do a real verification they may need to revaluate there statement all of this is not PC mode anyway.

every TV has major issue. this one is just stupid in PC mode on top of it.

[flossy_cake]

every TV has major issue.

Yes, definitely 100%. That's why I gave up on high end models cause I hated spending all that money on something I wasn't ultimately satisfied with. If I'm not going to be satisfied, might as well go low end and cheap. I'm still dissatisfied, but at least I didn't pay a bunch of money to be dissatisfied. The issues don't seem as bad when you didn't spend a lot of money. "Cheap and cheerful". So maybe I will just go with a B4 or C4 , or maybe even a used C1 as that is the last model with 120hz BFI which looks similar to 240hz sample and hold.

[Klaus1189]

replaced my LG CX because the batteries where empty mistakes have been made.

What? I don't understand :scared:

[huhn]

good.
but yes i was getting this one because my remote batteries where empty. this is not a joke. just enough hassle for me to get a new TV.

[tp4tissue]

an s90c the most broken TV on this planned.

a TV so intoxicated that it changes the primaries and secondaries if you change the gamma!

Hahaha. Yea S90c is buggy.

It's perfect for Tp4 though, because he only uses PC graphics mode.

The secondary twisted magenta issue in SDR huhn talks about is present in all modes except pc graphics, colorspace set to normal.

If you let the TV stay connected for a few days, just normal usage, you can turn it on and off, the bug resolves itself, the color space mode gets stuck on pc graphics mode, even if you change to native/custom. This bug resets if you plug in new devices, or unplug the hdmi, or if it drops out because your cable is dodgy.

This is essentially a bug that fixes another bug. So it allows you to use game mode now without the twist in magenta.

In PC-graphics mode, it's basically perfect out of the box, IFF you like full gamut, 2.2 gamma. Which is TP4's prefered usage mode.

If you want gamut clamping done by the TV itself, and Tp4 can't convince you of how stupid gamut clamping is, then it is slightly more problematic in non-color managed applications. :D

There is a minor drawback to for HDR if you want D65, because D65 uses less power, it pushes the EOTF output up slightly. This is very minor and acceptable for the most part.

I don't know what Vincent is refering to around 10%, It's probably in Filmmaker mode, which is again, in Tp4's opinion a STUPID mode. I didn't buy QD oled just to look at Shitty DCIp3 primaries. I want all the colors this panel is capable of.

And because samsung is really bad at UI design, Filmmaker mode doesn't allow 4:4:4 input, 422 only. This isn't such a big deal on game consoles or Bluray players, but Madvr benefits significantly in terms of rendering textures given 444.

[tp4tissue]

Yes, definitely 100%. That's why I gave up on high end models cause I hated spending all that money on something I wasn't ultimately satisfied with. If I'm not going to be satisfied, might as well go low end and cheap. I'm still dissatisfied, but at least I didn't pay a bunch of money to be dissatisfied. The issues don't seem as bad when you didn't spend a lot of money. "Cheap and cheerful". So maybe I will just go with a B4 or C4 , or maybe even a used C1 as that is the last model with 120hz BFI which looks similar to 240hz sample and hold.

I bought the G3 and had it side by side s90c. G3 does not handle dithering nearly as well as the samsung. It also just wasn't colorful in any bright movies.

G3 sneaks through for live action, because most standard colors exist in a 60% region of rec 709, but you throw a pixar or lego movie on there, WOLED just falls apart.

[flossy_cake]

Ok thank you for the detailed feedback, I will have to think about this some more.

My usage is very different to yours, I will be running SDR 444/RGB 709 gamut. I will not be using HDR or wide gamut. I don't watch Pixar or DreamWorks movies. The closest to that would be some SDR games.

I want to do calibration on the TVs processor, not in applications with ICC profiles or video card gamma table. I don't want to be reliant on that anymore. I want to be free to plug any device into the TV and get relatively accurate balanced stable colours without relying on external devices to correct the colour before it gets to the display.

So let's say I'm inputting SDR 709 to the Samsung, what kind of picture defects can I expect to see on actual content with the magenta issue? Does it affect skin tones? Do those 2 images I posted look okay in SDR mode? Are there any weird greenish/blueish tints in low light skin tones?

[huhn]

if you follow his suggestion you would get:
3:2 judder. they added this as a new feature!
neon color because you need to gamut map at your source.
and a gamma of 2.1 or something.

if you want to calibrate this monster well it starts very well at a grey scale and then you look a bit deeper and ohh boy...

and if you think recall is a problem. samsung Tvs have build in microphones with a hidden of switch! and then i complains that you turned it off and it may still be running.

[tp4tissue]

3:2 issue is not an enthusiast lvl issue. It's for casuals who use apps and devices other than PC.

Neon colorspace is overblown. This is the proper color space. Absolute colorimetric adherence is ideologically flawed. It also doesn't work on OLED at all because of ABL. You might get good measurements in a 10% window test pattern, but as soon as you get real content going, all that colorimetry goes out the window.

Exaggerated colors only happen on the edges in non-memory colors. The result is just people wearing slightly more colorful dresses, and shooting more saturated lasers.

[tp4tissue]

I will be running SDR 444/RGB 709 gamut.


The questions are incorrect.

It's like asking, what if flossy_cake eats raw chicken. While flossy_cake might have the intellectual curiosity to do so. It's not the right approach.

Rec709 is just a container. Viewing creator's intent is a negotiation process between your personal taste and what the creator cooked.

I used to be very narrow minded about containers as well. Absolute colorimetric obsessed. Eventually I realized it looks ugly, and greatly under utilizes the technology at hand.

Best example, tomatos and firetrucks in rec709. They don't look anything like real tomatos or firetrucks. We've all seen significantly more red tomatos / firetrucks in real life than the maximum red in 709.

This demonstrates EXPLICITLY, that rec709 is not the creator's intent, it was only the case that his hands were tied. Why then should we observe this limit, it doesn't look real to us, it wouldn't have looked real to the creator.

You mentioned games. They get eyeballed. There is even less intent where game coloration/ contrast is concerned. Adhering to rec709 for gaming is absolutely ludicrous.

So where do you go from here. Play with the tv, and make the image your own, your taste will change over time, there is no best. Maybe you want to watch a whole movie where people look like smurfs, or donald trump, we can't know what political cult you'll join in the future.

Madvr gives you alot of flexibility though, and it removes most of the pitfalls in terms of color banding, black crush, obvious image defects. And if you want to be a container obsessed luddite, use the 3dlut in SDR.

Getting perfect HDR adherence on OLED is currently impossible. Don't even try. If you want to be a container luddite in HDR, you'll want to go Bravia 9.

[huhn]

3:2 issue is not an enthusiast lvl issue. It's for casuals who use apps and devices other than PC.

Neon colorspace is overblown. This is the proper color space. Absolute colorimetric adherence is ideologically flawed. It also doesn't work on OLED at all because of ABL. You might get good measurements in a 10% window test pattern, but as soon as you get real content going, all that colorimetry goes out the window.

Exaggerated colors only happen on the edges in non-memory colors. The result is just people wearing slightly more colorful dresses, and shooting more saturated lasers.

no these colors even change depending on gamma.

even if a max color is not a representation of the accurate color this is false for everything that is not a max color which will be 99% of all content.

btw. this screen here get's brighter with bigger patches not dimmer. because samsung

[Sunspark]

I prefer relative colourmetric because at least then what the display is able to display it does 1:1 and whatever is outside is only an issue if the source has a much larger gamut which isn't the case with regular bluray which is rec.709 and colours at the edges of the triangle are unlikely to be a significant portion of the scene, they just get pulled to the edge of the srgb triangle. For UHD rec.2020 then I would use perceptual because that one is way larger than srgb. The colours won't be mapped 1:1 but the whole thing will be compressed equally spaced so the distances from each other will be correct.

(The above is talking about displaying content on an sRGB display. Someone with an Adobe RGB display might actually be able to use relative instead of perceptual with rec.2020 content.)

https://upload.wikimedia.org/wikipedia/commons/thumb/b/b3/CIE1931xy_gamut_comparison_of_sRGB_P3_Rec2020.svg/452px-CIE1931xy_gamut_comparison_of_sRGB_P3_Rec2020.svg.png

[flossy_cake]

if you follow his suggestion you would get:
3:2 judder. they added this as a new feature!


Oh yeah I saw in Vincent's review he says to get smooth 24hz you need to use some motion interpolation slider. And then he says after doing that it still has repeat frames on scene changes cause of glitches in the motion processing.

Is that correct? What happens if I feed it 1080p24 — do I get a frame every 41.6ms flawlessly without any repeat frame? Or do I have to use motion processing slider? I just want frame every 41.6ms, I don't care about smoothing it out beyond that.

[huhn]

you feed 120 or give up on this trash tv

[tp4tissue]

btw. this screen here get's brighter with bigger patches not dimmer. because samsung

I don't think so, I believe it's temperature related, it's getting brighter because bigger patches make it go Hot.

If you measure a 10% patch, measurement is brighter when the TV heats up, vs cooled down.

[tp4tissue]

you feed 120 or give up on this trash tv

Why would anyone feed anything other than 119.88 /120. Madvr enthusiasts arn't casuals. :rolleyes:

144 has slight crush but not practical for normal daily use anyway because it's not a power of 60 (online videos).

[tp4tissue]

I prefer relative colourmetric

The above is talking about displaying content on an sRGB display. Someone with an Adobe RGB display might actually be able to use relative instead of perceptual with rec.2020 content.


I think these rendering intents make for useful "taste" profiles. Thanks to Madvr and dispcal, we can just swap them whenever.

It's fine to make some movies Absolute Colorimetric, because it wouldn't change very much, if it's using mostly natural colors.

I'm just saying, the popular rhetoric in ethusiast forums of ONLY accepting Colorimetric is non-sense and certainly inappropriate for SDR.

SDR, just let it go wild. As long as it's not clipping at the end of the RGB values, there's no reason not to just crank it.

With natural colors, they're so unsaturated that even cranking it wouldn't move them very much, they'd still look mostly natural. You'd have to go out of your way to find an irregularly graded movie that's super-saturated to point your fingers at and say, look, look, look, he's orange. It almost never happens. :scared:

[mclingo]

I get this is a pretty niche problem but does anyone else use wireless headphones when watching movies As I'm getting frames skips in madVR every 40 seconds exactly almost, if I turn the headphones off and play through the normal receiver and speakers everything is fine. It does this with all my wireless headphones, I've tried two sets of Bluetooth headphones and also my artist pro headphones using Bluetooth and via the wireless box all produce exactly the same results so it seems to be something on the driver side or the renderer.

I have this when I was Windows 10 as well, i reimaged my PC and replaced the CPU and GPU so the only thing that remains is the RAM and the mainboard so either this is a known issue that cannot be solved or that there is a strange combination of hardware in my machine that causes it.

anyone know for sure, or know of a fix or a workaround?

[Manni]

Given that the level of the discussion on the S90C is appalling and that it’s entirely off-topic, do you mind taking it to another thread? It’s reallly painful to read so much misinformation about this set.

Please create a thread and moan/discuss there, it has nothing to do with madVR. Thanks.

[huhn]

I get this is a pretty niche problem but does anyone else use wireless headphones when watching movies As I'm getting frames skips in madVR every 40 seconds exactly almost, if I turn the headphones off and play through the normal receiver and speakers everything is fine. It does this with all my wireless headphones, I've tried two sets of Bluetooth headphones and also my artist pro headphones using Bluetooth and via the wireless box all produce exactly the same results so it seems to be something on the driver side or the renderer.

I have this when I was Windows 10 as well, i reimaged my PC and replaced the CPU and GPU so the only thing that remains is the RAM and the mainboard so either this is a known issue that cannot be solved or that there is a strange combination of hardware in my machine that causes it.

anyone know for sure, or know of a fix or a workaround?

wireless headset need there own sound card so the sync is different.

[tp4tissue]

If you use the bluetooth on your tv, it shouldn't cause issues since it's transparent to the PC.

[flossy_cake]

wireless headset need there own sound card so the sync is different.

Yeah but 41 sec is 24/1.001 mismatch, I'm struggling to imagine how plugging in headphones would cause that. That's super strange. Maybe it's not 1.001 mismatch and just the audio clock, but in that case ctrlJ would report "1 drop every..."

[flossy_cake]

rec709 is not the creator's intent

Even if the director himself came 20 years later and literally said "I prefer this new look", that's still not necessarily valid either otherwise you would have to agree with what George Lucas did to Star Wars and what James Cameron is doing to his movies

https://a.l3n.co/i/n5HXAx.png

https://www.youtube.com/watch?v=BxOqWYytypg

The artist today is not necessarily the same person as the artist 20 years ago. If all of their brain cells have changed and their decision making process is different due to life experiences - https://en.wikipedia.org/wiki/Ship_of_Theseus

I am not the same person I was 20 years ago and many people I know have changed a lot, some for the better some for worse.

Or, it could be that they are still the same person they were 20 years ago, but the constraints of the medium are what accidentally made them great in the eyes of their fans. Take away those constraints and the art changes.

there's no reason not to just crank it
...even cranking it wouldn't move them very much

So if "not moving them very much" is a good thing, why can't it also be a good thing to just not move any of them?

I don't have a problem with you applying your own palette swap if that's what you like, but to tell others they have to accept your palette swap as the true artistic intent is a different story.

What about matrix - do you insist on using the right matrix at least?

[mclingo]

If you use the bluetooth on your tv, it shouldn't cause issues since it's transparent to the PC.

I am not even sure my TV supports Bluetooth I'll have a look at that the other option I thought about was to buy a headphone Jack to Bluetooth dongle and plug that into my receivers headphone output, I can then adjust the latency in Kodi to match any misgivings.

i just be really interested to know what's going on here because I have noticed since that it doesn't do this when I use MPC renderer in MPC-BE as that also plays 3D movies, So it's definitely something to do with MadVR As I am also using lav filters in MPC BE.

[mclingo]

Oddly it still does it if I use my receivers Bluetooth I thought by using that it would take it out the equation as it wasn't using Bluetooth directly from the pc, Again 40 seconds.

EDIT:Scratch that Using Bluetooth from my receiver it does fix this the latency is terrible but it can be corrected so the issue is only when using wireless direct from the pc.

[flossy_cake]

@mclingo does madvr report "1 frame drop every..." on ctrl+J? Because if yes, then we know it's definitely not related to that other n/1.001 mismatch issue with your receiver which was NOT showing up on ctrl+J.

I had a similar issue recently I was getting a random repeat frame every now and then, and after about half an hour of pulling my hair out trying to fix it realised it's my own calibration app's "twilight transition" feature gradually transitioning between day.cal and night.cal at the video card gamma table and that causes dropped frames. I already knew about this when I released the app and wrote a note about it in the readme file but I totally forgot and was stumped for a good half hour thinking I've got major gremlins in my system. i'll probably make it default twilight transition to disabled state when using separate day/night cals so it doesn't happen to other people.

I don't know if Windows night light does it too, but others like f.lux and Redshift I think use the video card gamma table as well so that could potentially cause the same issue depending how often they update the transition every 1000ms or more. The reason I chose 1000ms is so you can't perceive colour changes during the transition aka "brightness pops" but it can be adjusted to other values

[mclingo]

https://1drv.ms/i/c/6240949ef769c50b/EdXucJDbVdNLlpc1SlqcWv0BID6mondLIz1cIbbBC8k_VA?e=OySZLR

Here is my control J, It does eventually count up to 40 and then just hangs around there.

[huhn]

@mclingo does madvr report "1 frame drop every..." on ctrl+J? Because if yes, then we know it's definitely not related to that other n/1.001 mismatch issue with your receiver which was NOT showing up on ctrl+J.

I had a similar issue recently I was getting a random repeat frame every now and then, and after about half an hour of pulling my hair out trying to fix it realised it's my own calibration app's "twilight transition" feature gradually transitioning between day.cal and night.cal at the video card gamma table and that causes dropped frames. I already knew about this when I released the app and wrote a note about it in the readme file but I totally forgot and was stumped for a good half hour thinking I've got major gremlins in my system. i'll probably make it default twilight transition to disabled state when using separate day/night cals so it doesn't happen to other people.

I don't know if Windows night light does it too, but others like f.lux and Redshift I think use the video card gamma table as well so that could potentially cause the same issue depending how often they update the transition every 1000ms or more. The reason I chose 1000ms is so you can't perceive colour changes during the transition aka "brightness pops" but it can be adjusted to other values

well known to make problem with madVR since flux.

https://1drv.ms/i/c/6240949ef769c50b/EdXucJDbVdNLlpc1SlqcWv0BID6mondLIz1cIbbBC8k_VA?e=OySZLR

Here is my control J, It does eventually count up to 40 and then just hangs around there.
yes just a different clock generator and they can vary quite a lot as i said this could be the same clock generator as your CPU. PC don't have a universal clock this will never be fixed ever.

[flossy_cake]

Maybe try a different audio renderer in MPC, see if that affects it. Can try different audio renderer settings as well like shared vs exclusive.

[Emulgator]

Even if the director himself came 20 years later and literally said "I prefer this new look", that's still not necessarily valid either otherwise you would have to agree with what George Lucas did to Star Wars and what James Cameron is doing to his movies
Well put. I am busy myself to reverse the damage done by modern grading: cyan tint, pale rotten colors of death everywhere.

[flossy_cake]

Well put. I am busy myself to reverse the damage done by modern grading: cyan tint, pale rotten colors of death everywhere.

Have you had any success with that? I have been thinking about that a lot, whether it's possible to colour grade them back to their original palette, or if the damage is permanently done. Warner Bros seem to be especially bad to me - many of their classic movies that I wanted to revisit have been given the teal treatment in their latest bluray releases. Skin tones are often pushed towards magenta, blacks crushed slightly, and crimson reds instead of neutral reds.

If they want to give a movie an exciting new palette for the re-release, fine, but they shouldn't discard the original. If the movie title has the word "remastered" in it, that's a red flag to me - I don't want their crappy "remaster" with weirdo colours and lighting. That's why sometimes I even prefer DVD to bluray if the palette is especially bad on the bluray.

[Sunspark]

With natural colors, they're so unsaturated that even cranking it wouldn't move them very much, they'd still look mostly natural. You'd have to go out of your way to find an irregularly graded movie that's super-saturated to point your fingers at and say, look, look, look, he's orange. It almost never happens. :scared:

With video, madvr is the only one that properly respects my monitor's colour profile on older content. evr-cp and mpc-vr both make skin tones too ruddy. We're not talking grossly wrong, but they do take on a hue shift toward the tomato end of the spectrum. Most people probably won't notice it, but I know it's too much red.

What about matrix - do you insist on using the right matrix at least?

flossy, the colour grading changes that have been going on with re-issues is absolutely out of control. Matrix is an good example, Star Wars: A New Hope is another. I think in some cases, it's not even the director doing it, just some guy at a PC shifting the content to a new format and not understanding what they are doing.

[flossy_cake]

What about matrix - do you insist on using the right matrix at least?
flossy, the colour grading changes that have been going on with re-issues is absolutely out of control. Matrix is an good example

In case there was any confusion I meant colour matrix 601/709/2020

But yes The Matrix should be green, shouldn't it? That's how I always remembered it.

Again, I have no issue with them changing the colours, as long as they don't delete the original from existence.

Maybe in future they could just have the raw ungraded movie and put the colour grading as metadata so the user can select which grading they want.

edit: after some research it seems the matrix was overgreened for the 2k bluray but not for the newer 4k bluray which people are saying is closest to the 35mm scan which is available on channel BT but got poorly scanned with blown highlights & crushed shadows. https://imgur.com/a/matrix-comparisons-iSfSrsM wish I had DVD screenshots to compare

[Sunspark]

The 2018 4K HDR bluray is a new scan from the film negative and was supervised by Bill Pope so it's probably as close as you're going to get to the 1999 DVD in high resolution.

https://dl.dropboxusercontent.com/scl/fi/rp76or4yh9ftc65xd7r4g/GOud4QRXwAAIF9S.jpg?rlkey=gvu3a93orz2yzgfjyng4eupbr&st=axjhp09y&dl=0

[tp4tissue]

The colorimetric specs are just guidelines. There is no right or wrong color. Only different flavors.

All of the releases are in it of themselves CORRECT.

Madvr gives you the option now, you can make it your own. If you want to regrade the whole movie, use davinci resolve, lots of tutorials.

When people get stuck on the idea that THE ORIGINAL is such and such, as if the original is a fixed point, it's not, nothing in the ether is fixed.

[flossy_cake]

@Sunspark thanks for that comparison. Looks like the 2kBD = green, 4kBD = cyan/teal, DVD = yellow, 35mm = neutral. It's a shame the 35mm scan of the film (not trailer) has clipping & crushing as that would have been my preferred version as it was probably closest to what I saw in the cinema in 1999. Whoever did that recent 35mm scan for channel BT must not have set the scanner up properly, such a shame. Just getting the DVD version, will probably just go with that for now as the tints aren't so egregious. It also seems to vary by scene like on some scenes the 4kBD looks quite neutral, on others like that lobby fight scene the stone walls are all tinted cyan/teal compared to the 35mm. So the 4kBD version is still quite off on some scenes.

"Hey let's just make everything teal now" - Hollywood
Me: " yeah no"

edit: nope, looks like there's more than one Matrix DVD version as the remux version I've got is green as heck (https://i.ibb.co/L1JbK4R/title-t00-mkv-snapshot-01-44-14-989.png) and isn't open matte either. There must be multiple DVD versions?

edit: ah yes:

The other main difference are the colours : whereas the 1999 version offers green-brown tones in the Matrix and soft blue tones in the real world, the 2004 version delivers very crude green-white tones in the Matrix and stronger blue-white tones in the real world.

Looks like I'll have to hunt down a 1999 pressing of the DVD :rolleyes:

Just 2024 things :rolleyes:

edit: doh, just realised your screenshot showed the green tint for the 2004 DVD version which I somehow missed.

edit:

I am busy myself to reverse the damage done by modern grading: cyan tint, pale rotten colors of death everywhere.
Have you had any success with that? I have been thinking about that a lot, whether it's possible to colour grade them back to their original palette, or if the damage is permanently done.

It seems there is a person by the name of KK650 who has done exactly that for The Matrix! I am not sure if I trust their judgement though. I suppose if they are targeting the neutral colours of the 35mm film as per what I would have seen in the cinema in 1999 when I was a kid, then as long as it's closer to that, it should be okay. I guess the only way to know for sure is to get all the versions and do your own comparisons on your own display so you can see for yourself. Well, this is going to be quite a journey...:rolleyes:

[flossy_cake]

The colorimetric specs are just guidelines. There is no right or wrong color. Only different flavors.

All of the releases are in it of themselves CORRECT.

Madvr gives you the option now, you can make it your own. If you want to regrade the whole movie, use davinci resolve, lots of tutorials.

When people get stuck on the idea that THE ORIGINAL is such and such, as if the original is a fixed point, it's not, nothing in the ether is fixed.

Forgive me for asking but is it possible you have some form of colourblindness? When you think of colours do you associate them with emotions? For example the colour red is considered bold and attention grabbing / striking. Maybe because it has some association with the sight of blood, maybe that's why we feel a certain way when looking at red. There's a thought experiment that says how could you ever get a blind person to understand what the QUALITATIVE EXPERIENCE is like to look at the colour red - they could never know.

Blue is a colour that can feel cold and sterile, or it can be used in corporate settings to give a "professional" feel, like a bank might use blue as their colour scheme.

Green is a natural colour of trees and grass that makes me feel calm, but darker shades can look like rotten algae or something sickly.

Do you feel these emotions when looking at colours? Because it sounds like you aren't affected much.

[flossy_cake]

Maybe it's not colourblindness but some kind of adaptive perceptual thing where at first the matrix tint looks too green, but then after a few minutes our eyesight adapts to that palette as the new norm and it doesn't seem so green anymore. Maybe the rate or amount of adaptation is different for different people. So you might have adapted to it while I'm still in green hell thinking why is everything so green.

[flossy_cake]

was supervised by Bill Pope

I hope he is more trustworthy than George Lucas and James Cameron! :thanks:

[flossy_cake]

Another psychological aspect is that the altered palette bothers me more when I am deprived of access to the original palette. Then, once I get access to the original palette, I don't feel so bothered by the altered palette, because it's an optional extra, I can take it or leave it. If anything I am more receptive and open minded to see how their new palette might change the look and feel in interesting ways, because it's not being forced on me.

[mclingo]

for me the 2008 bluray seems to be the correct colour but my memory is poor, I do remember this been very green / teal in the cinema.