https://youtube.googleblog.com/2016 /11/true-colors-adding-s

>>57431624
I like the way actual video looks nothing like the "simulated" SDR image and much more like the "actual" HDR image.

https://youtu.be/tO01J-M3g0U?list=PLyqf6gJt7KuGArjMwHmgprtDeY8WDa8YX&t=29
In fact, the actual video on my display looks over saturated in comparison.

>>57431864
Looks like a shitty SweetFX preset.

So this is the next gimmick the industry is going for.
I wondered what the next big thing was going to be after 3D and resolution whoring.
Apparently it's the colors.

>>57431864

This looks like one of those videos they continuously play as a preview on TVs and monitors that are on display at stores.
It's over saturated to hell and back, so it would make even the shittiest of displays look good, which it does.

HDR doesn't actually make the images more colorful though.

I don't understand something. How the shit is HDR supposed to improve contrast by increasing bit depth? You can display finer gradations of color since it's 10b of course, but how the fuck does that affect contrast? How is feeding your screen 30x 0 bits any different than feeding it 24? It's still the darkest value it can go to.

Is it just marketing shit for displays which come with better contrast, which could very well exist with 24bpc as well? I just don't see how a screen's black level or max brightness could possibly depend on the number of color bits you feed it. Is "HDR" basically more colors + some sort of minimum contrast ratio certification which a good-quality SDR screen could get too?

>>57432188
It doesn't affect contrast.
You have to sell it to normies somehow though, so they try to market it like that.

>>57431624
what the fuck they literally just halved the saturation on the left one
what the fuck

>>57431624
It's funny how they advertise HDR. Like they'll show the non-HDR as a shitty, greyed out pícture, while the HDR version is just a standard picture with boosted contrast.

Gotta sell new TVs/monitors somehow I guess, since they have to wait for content creators to catch up to the 4k/8k meme resolution.

>>57431624
>Sanjeev Verma, Product Manager,

>>57432243
I've seen actual HDR TVs at a shop (the super expensive OLED ones), and they do exactly that. It's false advertising

Displays have always been able to display what HDR's main advantage is. Shadow detail and avoiding blown out skies. That's always been an issue with the device recording it. HDR on new displays just add more bits for color gradation and higher contrast ratios due to light output on the display. But for just revealing shadow detail and blown out highlights HDR in recording terms is still something you can see on a standard display.

>>57432244
POOLETE THIS

>>57432212
Yeah that's what I thought. Can 10b color really make that big of a difference though? I haven't ever seen proper 10b content on a 10b screen, but other than potentially decreasing banding issues (which aren't really a huge deal anyway) I can't imagine how it would have some sort of huge impact. Most people apparently can't even tell 720p from 1080p or 1080p from 4K, are they really going to notice 2 extra color bits per channel?

>>57432303
2 bits are a huge fucking deal. It means 4 times the color resolution.
It is apparent mostly in video games, but none of them support it right now apart from shadow warrior 2 (I think).
The ability to display skies is much better, and shadows and dark areas are much more defined.
Thanks to the ps4pro we will see a lot more games with 10bit.
I think the change is necessary and good.
No longer will gradients look fucking retarded.

>>57432361
SW2 does have the option, yeah. I have to see it for myself, I know we're going from 256 to 1024 values/color, but I haven't really seen any glaring issues with shadow detail and skies, even in games, as long as you're running them on a decent monitor which doesn't crush blacks or show clipping in the whites.

What the fuck kind of device was used to take the left picture then? I've seen chink phones with better sensors. Is this even legal?

>desaturate photo
>slightly oversaturate same picture and out it next to it
>LOOG MA INNOVASHUN XDDDDD

File: UHDR.jpg (171KB, 799x450px) Image search: [Google]

171KB, 799x450px

Now presenting to you: Ultra-HDR. Compatible with every display!

>>57432188
10bit doesn't make it HDR.

High contrast makes it HDR.
But if you stick with 8 bit HDR will look like garbage, it's barely enough for regular screens.

ITT /g/ doesn't understand what HDR is and thinks it has anything to do with color.

>>57432243
Pretty much. The 3D thing flopped hard, the 4K things flopped hard, so this is the new buzzword for why you absolutely MUST run out and buy a new TV. It's going to flop hard as well.

File: 1.jpg (334KB, 1920x1080px) Image search: [Google]

334KB, 1920x1080px

I prefer XtremeHDR+
It looks so much better and makes the colors pop-out.

>>57433158
3D and 4K are memes.

But better contrast really makes a difference.
Especially when it's darker blacks.

Yes, the industry needs a new hype every year to sell new TV's.
But at least this is a useful hype unlike most years.

>>57433242
>4k
>meme
Maybe for videos and video games, but certainly not for every day use.

>>57433281
For TV's, ie: movies.

HDR is also not meant for spreadsheets.

File: reee.png (115KB, 799x450px) Image search: [Google]

115KB, 799x450px

>>57433183
I prefer Nvidia G-UHDR desu

>>57433099
HDR in film and photography is about the ability to expose so that bright lights aren't over exposed while dark areas remain visible.

The way HDR TVs are advertised is with saturation levels.

File: my retinas are dissolving.jpg (888KB, 1529x1085px) Image search: [Google]

888KB, 1529x1085px

>>57433352
made it HDR for you

File: ff5hx9hj.jpg (2MB, 2444x1270px) Image search: [Google]

2MB, 2444x1270px

>>57433368

>>57433431
NOW THIS IS HD™

>>57433404
It was already bad enough. either RED has shitty colors or whoever was grading the image was DUDE WEED.

File: c6s17i5f.jpg (772KB, 1922x955px) Image search: [Google]

772KB, 1922x955px

>>57433374

>>57433431
Holy shit, now I know why the matrix dudes wear sunglasses even when using computers

>>57431624
2160p60 HDR porn when?

LCD's still can barely keep up with 8-bit so what the fuck is the point? Where is the monitor that will show this?

>>57433354
>The way HDR TVs are advertised is with saturation levels.

Advertised, yes.
Because you can't actually show HDR on a non-HDR screen so the marketing geniuses decided to show saturation instead.
Sucks but such is life in normie land.

>>57433503
>not 4k 120fps 3D hdr

>>57433354
and in monitors and video coding it's purely about higher contrast levels and wider gamut without banding, dithering pretty much already makes it happen and 99.999% of content is 8 bit so the real world gain will be minimal

>>57433465
fugg :DDD

>>57433565
They'll be here right before Christmas I suspect.

They could make HDR screens 15 years ago, but they had to get on the megapixel race first.

>>57432188
>How the shit is HDR supposed to improve contrast by increasing bit depth?
HDR is not about increasing the bit depth, it's about increasing the dynamic range. They're related but ultimately orthogonal.

>You can display finer gradations of color since it's 10b of course, but how the fuck does that affect contrast?
HDR is *not* the same thing as 10-bit SDR (as was used with UltraHD or Hi10P content). HDR is about increasing the brightness range you can represent, from 0.1-120 to more like 0.001-10,000 nits

>Is it just marketing shit for displays which come with better contrast, which could very well exist with 24bpc as well?
Pretty much. Brighter whites and darker blacks. That's what HDR is all about.

Note that youtube adding HDR is actually a bigger benefit than just increasing contrast: Since HDR sort of implies BT.2020, we're also getting a much wider color gamut out of the box.

>>57433601
and what is the technology? not lcd. oled? enjoy paying $10000 for that TV

File: mpv-shot0006.jpg (2MB, 4096x2160px) Image search: [Google]

2MB, 4096x2160px

>>57433603
btw, mpv supports this but it misdetects the gamma (it detects it as bt.1886 but it should be st2084, I think)

File: het.png (839KB, 600x876px) Image search: [Google]

839KB, 600x876px

>>57433580
>3D porn

File: Foveon-Testfoto-Sigma-DP1-0246[1].jpg (289KB, 1200x800px) Image search: [Google]

289KB, 1200x800px

>>57433442
>tfw no foveon digital film camera

>>57433624
Just improves LCD and OLED screens, yes.

>enjoy paying $10000 for that TV

I doubt it will even be more expensive than 2016 models.
It's just improvements to be able to keep selling new TV's.

Since I never fell for the full-HD meme, let alone the 4K meme I've saved plenty of money to replace my good old 720p TV.

>>57433663
that looks so comfy

>>57433567
I understand that but it is misrepresenting current state of devices and the feature.

>>57433700
I feel your autism, bro.

>>57433663
>Foveon
Where is that? looks a lot like where I live

>>57433624
You can buy HDR OLED TVs for like $5000 right now.

>>57433242
It's irrelevant whether it makes a difference or not. People aren't going to throw away their TV and buy a new one to get it.

>>57432227
>>57432243
>>57432464
Is this your first time seeing a marketing picture?

>>57433663
Meh, my D800 has a better sensor.

>>57433738
>Sony sensor
>good

>>57433747
Uhm....yes, Sony sensor good.

>>57433729
they usually at least try

>>57433685
>improving LCD

how? they're at their contrast limits already, if display manufacturers could push more inflated native contrast numbers than 1:2000~ for the very best they would have done so long ago

File: 10489888_695440023838520_3755568354881658043_n.jpg (42KB, 640x480px) Image search: [Google]

42KB, 640x480px

>>57431624
I have a Dell Ultrasharp 2408wfp monitor which supports a color gamut outside the standard 16 million colors.

Will I be able to see any benefit from this?

The spec sheet says it has

>110%2 Color Gamut (CIE 1976) - With Dell TrueColor Technology, you’ll see more color than average monitor of 72% color gamut.

>>57433797
>if display manufacturers could push more inflated native contrast numbers than 1:2000~ for the very best they would have done so long ago

No, because they were focused on making higher resolution panels with shitty contrast because nobody gave a shit about anything except how many "K's" it had.

>>57433798
>which supports a color gamut outside the standard 16 million colors.
You're thinking of bit depth. A monitor can support 10bpc color but that doesn't mean it has a wide gamut. Likewise there are plenty of 8-bit monitors with a wide gamut.

In either case, no, you won't notice any real world benefits.

File: by doing this.png (340KB, 736x829px) Image search: [Google]

340KB, 736x829px

>>57433797

>>57433834
yeah the backlight thing, which isn't native and looks like shit, it's just localized dynamic contrast

>>57433829
Are there any videos that exist with a wide color gamut that I could play on my monitor see it using its full potential?

>>57433798
windows assumes srgb, and performs no correction for dekstop elements, so you're probably already seeing it, just stretched from srgb to your monitors native

>>57433874
Yes, the colors look very over-saturated on everything unless I put the monitor in sRGB mode, and then it looks dull compared to my other normal monitor.

The overall image is still crazy good. I wish there was a way I could use it as my main desktop monitor without funky colors where I only see things with more color when they actually have more color.

>>57433729
It's exactly the same with actual HDR monitors/TVs. I've seen a few. They grey out the normal picture to make it appear worse.

>>57433770
No they don't. Show me literally a single video tech or monitor marketing picture that even remotely tries

>>57433857
get OLED and you have it native

>>57434047
the comparisons are mostly overdone because they compare it to the oldest tech grandpa possible.. if you compare a new HDR IPS/MVA Monitor with an OLED one, OLED would still look better.

but compare that with TN, it most likely wouldn't.

i really hope quantum dot will make it's appearance in the same momentum

>>57432188
You're confusing HDR with 10bit video.
They're two different things.

>>57434330
>the comparisons are mostly overdone because they compare it to the oldest tech grandpa possible..
No, the comparisons are always completely and absolutely made up. There is no technological basis whatsoever in them

>>57433431
SUCH VVVVVVVVVVVVVVVVVVVVVVVIBRANCE

>>57432188
>some sort of minimum contrast ratio certification which a good-quality SDR screen could get too?

Unfortunately I don't think this is the case. An HDR rated display can have any contrast ratio.

But yeah it is confusing and the facts are fudged. But what consumers need to know is that HDR video and good contrast ratios go hand in hand, because you need good contrast to take advantage of HDR video.

>>57434655
HDR is basically a marketing label used to mislead the public into buying “HDR” LCD panels. It's a bait-and-switch: Show people fancy OLED tech at conventions, get them hyped about OLED, sell them shitty “dynamic HDR” LCD panels instead.

>>57433603
>HDR is not about increasing the bit depth, it's about increasing the dynamic range. They're related but ultimately orthogonal.

Yes it is, depending on which technology named HDR you're talking about.

>Pretty much. Brighter whites and darker blacks. That's what HDR is all about.

That's static contrast ratio.

>>57432996
HDR video is 10 bit
We've always had high contrast displays, and we didn't call them HDR. You could for example, have an OLED display that can show perfect blacks and whites as bright as the sun, and it only displays 8 bit video.

https://en.wikipedia.org/wiki/High-dynamic-range_video

>>57432243
That's the best simulation you can present on an SDR display though. Black is actually dark grey, and white is bright grey. Your eyes just adjust.

>>57433857
>yeah the backlight thing, which isn't native and looks like shit, it's just localized dynamic contrast

Compared to what? OLED is a good while away from being cheap and long lasting.

Then again the active matrix backlight with a decent resolution is a long way from being cheap too.

>>57432303
>but other than potentially decreasing banding issues (which aren't really a huge deal anyway)
Really? It's a pretty huge deal.

>>57434736
>Yes it is, depending on which technology named HDR you're talking about.
I'm talking about the term `HDR` which literally stands for `high dynamic range`, and in this context refers to SMPTE ST.2084 and ARIB STD-B67

>That's static contrast ratio.
Sort of. The point of HDR is not just pushing the static contrast number up, but also increasing the overall envelope (i.e. brighter brights, not just darker darks)

In a way, static contrast ratio is a number that tells you almost nothing about the real black/white points. For the purpose of clarity it's best to focus on the concrete numbers rather than their ratio.

A typical, high-quality SDR display has a black point of like 0.1 cd/m2 and a white point of like 100 cd/m2, giving you a contrast of about 1000:1

A high-contrast SDR display might have a black point of 0.001 and a white point of 100 cd/m2, giving it a static contrast of 100,000:1. But it's still not HDR

A HDR display might have a black point of 0.01 and a white point of 2,000 cd/m2, giving you a static contrast of 200,000:1 but a greatly increased dynamic range (meaningful brightness spectrum).

>HDR video is 10 bit
That's an implementation detail, really. You could have an 8 bit HDR video, it would just be lower quality. You could also have a 12 bit HDR video. (In fact, the spec allows for the latter iirc)

>>57432303
You can make up for minor differences in bit depth by dithering heavily, but the problem with dithering is that it's extremely expensive to encode.

(This is why 10-bit encoding is more efficient than 8-bit encoding on anything with gradients - you don't have to dither)

So really, while 10-bit isn't strictly necessary, it's simply more efficient bit-wise.

File: 8bit_motion.gif (880KB, 500x500px) Image search: [Google]

880KB, 500x500px

>>57432303
>8 bit ought to be enough for everybody

>>57431624
This is the same thing I got when I set the display output range in my video player from 16-235 to 0-255.

11 years ago.

/p/ here

i want to murder you all

this is some atrocious dunning krugering

>>57434931
That's from shit encoding though.

>>57434987
The image literally displays a lossless 8 bit grayscale ramp

File: Untitled.png (11KB, 1920x1080px) Image search: [Google]

11KB, 1920x1080px

>>57434987

>HDR
>marketing alphabet soup that has nothing to do with the photographic technique it's named after
>there are two wildly different competing standards - one open and easily implemented, the other a Dolby abortion that requires your display include a proprietary image processing chip
Not only is it meaningless marketing speak, it's three or four separate and arguably unnecessary improvements. Increased bit depth, a different color space, vastly increased brightness, and contrast-boosting techniques like local dimming.

>>57435147
>marketing alphabet soup that has nothing to do with the photographic technique it's named after
Fuck off, the `HDR` from photography is pure cancer and has nothing to do with actual dynamic range. That's just shitty faux-HDR tone mapping.

This is about true HDR, end-to-end; which requires a HDR camera, HDR encoding and HDR display. No tone mapping involved.

>>57435147
>the other a Dolby abortion that requires your display include a proprietary image processing chip
Wait, why? Do they have patents on the algorithm or what?

>>57434975
If it were up to /g/ we'd all be using 480i CRT eye cancer crates.

>>57434975
/g/ here

I implement the HDR support in your image editors and media players

feel free to fuck off

>>57431864
I watched it in 144p for the luls

File: mpv-shot0007.jpg (480KB, 4096x2128px) Image search: [Google]

480KB, 4096x2128px

>>57435320
you madman

>>57435147
>photographic technique it's named after
Im getting second-hand embarrassment here, kill yourself. HDR tech in TV is not derived from the tonemapping techniques used in photography. If anything it attempts to do the opposite: actually displaying a greater range of contrast, rather than compressing it into a lower dynamic range.

>arguably unnecessary improvements
You're a giant faggot

>>57433324
But AMD has FreeHDR which does the same shit you fanboy faggot

File: 1478462305311.jpg (133KB, 1367x1692px) Image search: [Google]

133KB, 1367x1692px

>youtube

itt:
>view HDR material on non-HDR display
>"wow, this looks the same/worse"

>>57436419
More like "wow it looks nothing like what they claim sdr content looks like"

>>57436442
there's no point in viewing a real side-by-side comparison on an SDR display, it makes no sense

it's just a marketing simulation, it's like advertising 3D displays with an image of something actually coming out of a screen, of course that doesn't happen, but it's just to convey the idea

>>57435031
>>57434931
>what is dithering

>>57436465
The problem is that the simulation doesn't reflect the reality in any way.

>>57431864
This video looks amazing in my Nexus 6 holy shit.

>>57431948
Real life (your eyes) will be probably like the right one desu.

>>57436537
Not enough to truly fix the problem. Also that's using a linear gradient which doesn't have linear light output. Using a gamma corrected gradient has even worse banding in the dark areas.

>>57436537
good job wasting 100 mbps on what could be 10 mbps with a 10-bit image instead

>>57436537
difficult to compress, is what it is

a smooth, high bitdepth gradient is easier to handle than a noisy, low bitdepth gradient

LCD's are able to display higher nits than OLED at present. But because OLED's inherently have supreme contrast ratios to begin with they got the HDR label anyhow. I think the intended nits is something like 5000 nits which is like sunlight blinding. But all LCD's out there can only go up to 2000 at the most I believe. A prototype had 4000.

>>57438189
Oh! The offshoot of this is that the power usage of TV's has to be kept in check (especially in countries trying to get power usage down like those in the EU). Generating brighter images requires more power obviously so it's a push/pull problem to get TV's with low power usage to meet the requirements in certain countries vs. trying to make the image brighter.

>>57438189
The problem with LCDs is that the brighter the backlight gets the more difficult it is for the pixels to block the light completely which can ruin your blacks. Even if the backlight was managed dynamically in scenes that have both very bright and very dark areas you would need the backlight to be sectioned into grids to get the effect, and even that wouldn't be perfect in some situations.

OLED has the benefit of producing light from the pixels themselves, which means blacks are pure, but the added problem is that it is more difficult to create much brighter pixels. I think some panels have a dedicated white subpixel to help in this.

We are probably going to see more clever subpixel designs in OLED to overcome the brightness issue before LCD can control its black issues.

So, if I got it correctly to display HDR both the screen and the media itself has to be compatible.

Doesn't this imply regular content would look relatively dull on a HDR screen? And wouldn't this mean HDR would look blown out on a regular screen?

Why can't they make HDR screens natively convert media into the HDR range?

>>57438421
>Why can't they make HDR screens natively convert media into the HDR range?
But they will. It just won't be perfect because it will be done in software.

Also, what about regular images?
Is there an HDR image format for the internets on its way?

>>57438445
Not perfect how?

>>57438461
Most likely it would just adjust the levels dynamically in an attempt to expand the range but it will be limited in how much detail was lost in the storage of the video and the algorithm.

In proper HDR video effort will be made to retain details in the very dark and very bright areas so that they can be displayed correctly.

>>57438451
It's called PNG.

i sure am glad i dont have autism and don't know what any of you are talking about and thus do not get triggered on a regular basis based on what i watch.

>>57438770
>posting on this board while claiming you dont have autism

>>57437785
>>57437887

okay so store it in high bit and dither at runtime, solved

>>57436874
It somehow looks good on my shitty (color-wise) monitor running redshift. I think it doesn't have any flat areas where the oversaturation becomes apparent.

>>57432243
>catch up

there's nothing to catch up to, you literally can't see more than 1080p on an average TV

you need 60"+ TV's to notice 4K, 8K is large format cinema only, direct competitor to 70mm

File: hdr-comparison.jpg (1MB, 2046x3186px) Image search: [Google]

1MB, 2046x3186px

>>57438421
>So, if I got it correctly to display HDR both the screen and the media itself has to be compatible.
Pretty much, yes

>Doesn't this imply regular content would look relatively dull on a HDR screen?
Yes.

>And wouldn't this mean HDR would look blown out on a regular screen?
Well, fundamentally this is impossible to display. So it depends on what adjustments you do to turn HDR into SDR:

- no adjustment, just display the HDR image as if it was SDR: image will look extremely gray and bland
- hard-clip out-of-range values: image will be will of awful clipping artifacts as if it was constantly blowing out the sensor
- tone-map the brightness with a nonlinear curve: will look sort of like SDR content normally looks, with not much loss. (but slightly distorted colors)

If you use a media player like mpv which is HDR-aware, it will do the third by default. If you use any other media player, it will look like the first.

Pics related (in order)

>>57440078
That's how it's done in practice. Good job figuring out what we knew 50 years ago.

File: hdr-comparison-2.jpg (922KB, 2046x2124px) Image search: [Google]

922KB, 2046x2124px

>>57440530
Oh, also, if you know the HDR signal peak then you can do a much better job of adapting the brightness range. Normally this is known from the image metadata (including on youtube), but this clip in particular had broken metadata so I had to tag it manually. After doing so, the top is what the “tone-mapped” result looks like.

Also, there's a third possibility for what you can do with the values: You can linearly stretch them to your device's brightness range instead. The bottom pic shows what that would look like.

>>57432160
It doesn't? So going from 8-bit to 10-bit does not add more colours?

>>57433747
Best.

>>57440636
>So going from 8-bit to 10-bit does not add more colours?
Going from 8-bit to 10-bit pretty much just makes gradients smoother. Bit depth and colorspace have almost nothing to do with each other.

That said, HDR, in practice, *does* make the images more colorful because HDR is usually paired with BT.2020 which is a much wider colorspace than BT.709, allowing you to represent significantly more saturated colors than before.

File: world-in-hdr.jpg (553KB, 2046x2124px) Image search: [Google]

553KB, 2046x2124px

>>57440668
>>57440597
Here's a comparison on one of the youtube clips. Top is the tone-mapped HDR image, bottom is the `SDR` image for comparison.

Btw, these images will most likely not look correct on your monitor because they're rendered for my calibrated wide-gamut monitor in particular. So use them for a comparison between the two, but don't use them as an objective reference.

>>57440535
50 years ago shit was analogue and didn't have banding issues at all nigger

>>57440727
just boost the saturation of the skateboard got it

http://www.samsung.com/us/computing/monitors/curved/34--cf791-wqhd-monitor-lc34f791wqnxza/


pre-ordering this right now folks

>>57440739
Digital signal processing is 50 years old nigga

>Simulated SDR vs HDR
Nice marketing.

>>57440758
If you don't know what you're talking about, it would behoove you to stop commenting on it

>>57440758
>taking it as an objective reference as advised

>>57440769
>21:9
enjoy your black bars
>VA
enjoy your shit off-angle gamma
>3000:1 SCR
not even good for a VA display. I don't know why they're marketing this as if it's supposed to be good
>curved
you just fell for the ultimate meme

>>57440828
i watch movies on it 70% of the time so which blackbars :-)

>>57440828
also, it's quantum dot

File: saturation.jpg (413KB, 2046x2124px) Image search: [Google]

413KB, 2046x2124px

>>57440807
>>57440819

right back at you

bottom saturation matched to top, it's the exact same image with the board artificially boosted

>>57440727
>youtube clips
The OP only links to HDR playlists. Wheres the SDR video?

just in theory,

wouldn't a normal dynamic range video look like a HDR video and vice versa when the monitor doesn't support HDR?

>>57440828
>enjoy your black bars
b8

>>57440839
>quantum dot
wew lad, I thought consumers were smarter than this

guess not

>>57431624
>this meme again
Most people don't even have HDR-capable screens, so how the fuck are they supposed to tell the actual difference without spending $300+ on an HDR TV/monitor?

>>57440857
>artificially boosted
Dude just stop. The one in the bottom is artificially clipped (BT.709 gamut)

look it up

>>57440977
then why is nothing else """""""clipped""""""""

File: gamut.png (660KB, 1352x1907px) Image search: [Google]

660KB, 1352x1907px

>>57440990
Because nothing else was out of gamut? Do you even understand what a gamut is? Here is a page from the ITU-R Report BT.2246-2 giving you an example of some real-life objects and their associated color spectrums.

Everything inside the smaller triangle will look the same in both pictures. Everything outside the smaller triangle has to be clipped for the SDR (BT.709/HDTV) release.

Have you never noticed how artificially limited the colors of your display are compared to nature? Go find the reddest object you can find and compare it against #FF0000 on your display - it will look extremely orange in comparison.

Same is happening with the skateboard. It's red in real life, but BT.709 can only represent up to deep orange. That's why it has to be clipped for the SDR release (bottom image), but it can have its true color in the wide gamut version (top image)

>>57440219
>you need 60"+ TV's to notice 4K, 8K is large format cinema only, direct competitor to 70mm
get fucked nigga
You WOULD notice the difference if you were given a 1080p tablet and a 4K tablet.

>>57440948
>being on 4chan
>"WHY DOES THIS TOPIC NOT BELONG TO MY AUDIENCE"

it's because you're poor anon

>>57440948
>people don't even have HDR-capable screens,
Yeah they do. Because they're not poorfags and buy Apple

>>57441077
yeah? and you hold tablets inches from your face, not across the living room from your couch. even then most of this retina bullshit is placebo

>>57441125
>>57441135
You dumb niggers. I'm saying how are people without HDR screens supposed to tell the difference between the two footages if they were actually comparing 8-bit to 10-bit/12-bit color. Their monitors would only display 8-bit colors and have color banding for both footages.

>>57441137
have you actually ever used a HiDPI device or are you trolling?

>>57441039
>Because nothing else was out of gamut?

yeah, because nothing else has been boosted. great explanation satya

>>57441167
yes. it's a complete meme

>>57441164
how are people without virtual reality devices supposed to tell the difference between that and a monitor?

pro-tip, they can't because they don't own a virtual reality headset

now that is clear.

what about a people that do indeed own a HDR capable monitor? now is the chance to try it out!

>>57441184
instead of monitor i wanted to say TV.. there currently are no fully HDR certified/capable monitors

>>57441039
>>57441170
Btw, pic related is a 3D comparison (intersection) of the gamut in the HDR test frame I presented here in >>57440727 (top version, solid) versus the gamut of the BT.709 color space (wireframe)

Everywhere where the original image exceeds the gamut of the wireframe, you would get clipping in the SDR version since the gamut has to be desaturated to fit. In the BT.2020 version, no such restriciton applies.

a nice article for some stupidos

https://www.rockpapershotgun.com/2016/09/22/hdr-gaming-and-the-pc-its-complicated/

File: test-gamut.png (390KB, 1267x1375px) Image search: [Google]

390KB, 1267x1375px

>>57441257
Forgot pic

Anyway, you can notice how the red parts are significantly out-of-gamut. This is not “artificially boosted”, it's just out of gamut. The red of the skateboard is not representable by BT.709, so it gets clipped in the SDR version.

>>57441272
so? the board has red and yellow on it, you're still not proving anything, the rest of the image is the exact same

>>57441170
Nothing in >>57441039 was “boosted”

The measurements were derived by using a carefully tuned xenon lamp with the spectral energy distribution close to that of the midday sun, and measured using a high quality scientific imaging colorimeter worth many times your house.

The values displayed are nothing less than an accurate indication of the gamut of real life. This was known 70 years ago, the original television sets were all wide gamut.

Then the jew came along and decided to desaturate all of the primaries so they could manufacture TVs for cheaper instead, and decided everything being bland and orange would be fine for the average goyim. Thus originated the BT.601 colorspace.

What you're seeing is nothing but the colors of reality finally being unlocked, which were stolen from you for 50+ years by greedy phosphor manufacturers.

>>57441311
Well you're free to wallow in your ignorance and think that reds are supposed to look orange. Whatever, this is my last reply.

>>57441321
>Nothing in >>57441039 was “boosted”

okay? you're still not proving anything because we're not talking about the ITU pics we're talking about >>57440857 >>57440727

Just one question: when are the HDR monitors going to come out? It's been almost a year since AMD first made a huge deal about it and yet there are still less non-richfag HDR supporting monitors than the fingers I have on one hand.

>>57441558
HDR monitors require DisplayPort 1.4 that supports transmitting the proper HDR metadata and GPUs that support DisplayPort 1.4 like Nvidia Pascal

http://www.geforce.com/whats-new/articles/pascal-video-playback

>>57441558
>>57441634
Note that as long as your OLED or whatever display can represent the brightness range required for HDR, it doesn't really matter what signal format you use to transmit to the monitor since you can do the necessary conversions in software as well.

>>57441400
Well feel free to buy a copy of the skateboard in question and hold it up to your monitor

>>57441039
is that image really putting every color dot where it belongs in the graph or did someone just spilled some colored dust and took a pic

>>57433431
Holy shit this is truly progressive.

>>57440917
dude, quantum, lmao

>>57442292
The graphs look reasonable to me.

>>57431624
Looks like oversaturated shit. I was expecting the same result as the HDR feature on cameras..

>>57443334
But that's literally exactly what this is about

>>57443681
What? No. HDR pictures taken with my phone don't look oversaturated.

File: 1477496378304.jpg (16KB, 172x233px) Image search: [Google]

16KB, 172x233px

>views on a non hdr display
>"this looks fucking oversaturated, HDR is truly a meme"

>>57443715
Your phone almost certainly doesn't have a HDR camera

Also, are you viewing these videos on a HDR monitor?

>>57433281
its only not a meme if you

1) have a 40-50 inch display and use it for productivity
2) work with 4k video and need a 5k so you can see the full video and tools at the same time
3) have a massive fuck off screen to the point 4k has a detail advantage.

>>57433720
when you see how good the contrast is on oleds, there are more then a few people who would junk their current tv for that.

>>57441167
have access to a first gen ipod touch, current gen ipod touch, and a prior gen ipad.

to see a difference you have to put it uncomfortably close to your face or get out a jeweler's loupe

the only person I know with a higher then 1080p monitor sits less then 10 inches from the display and bitches about jaggies

dpi is best between 80 and 120 dpi, most comfortable without any scaling shit.

>>57443746
A camera can do actual HDR, it helps with getting more detail out of dark and light areas like what HDR tvs do

the video is not HDR, its just over saturation

>tfw my "wide gamut" thinkpad screen displays everything oversaturated by default

>>57440948
fun thing is, there is a comparison video online, of normal vs hdr screens with the same video

shot in a shitty video camera
compressed by shitty youtube,

the only difference is they fucked with the contrast and saturation to he point the video looks like hell

>>57433917

> I wish there was a way I could use it as my main desktop monitor without funky colors where I only see things with more color when they actually have more color.

It's called macOS.

>>57432451
the difference isn't massive, you are getting 3 more in between colors for every color, unless banding was an issue for you its not a massive deal.

>>57443801
32" screen here and 4K is a _big_ step up from 1080p

I just wish they'd increase the compressed bitrate on 1080p videos.

>>57438384
Panasonic uses a honeycomb mesh with an array of local dimming LED's to achieve this. The mesh prevents halo's around bright objects apparently and supposedly works rather well. I dislike LCD because of the ghosting I sometimes see on fast moving objects. I have a Panasonic plasma (A midrange G20 from 6+ years ago). It has issues and I am looking to get an OLED as my next purchase once the prices drop to a more reasonable level. Currently the latest LG 55" model is £1999 in the UK but I am sure it will come down further (hopefully in the sales).

>>57443893
>the video is not HDR, its just over saturation
Yes it is, HDR video is literally made to fully capture the output of a HDR camera.

>A camera can do actual HDR, it helps with getting more detail out of dark and light areas like what HDR tvs do
Yes, which is exactly the effect you get out of HDR videos. Full details preserved in both bright and dark areas

>>57433720

Flat screen announced
>It's irrelevant whether it makes a difference or not. People aren't going to throw away their TV and buy a new one to get it.

720p announced
>It's irrelevant whether it makes a difference or not. People aren't going to throw away their TV and buy a new one to get it.

1080p announced
>It's irrelevant whether it makes a difference or not. People aren't going to throw away their TV and buy a new one to get it.

etc ...

Yeah sure buddy. Don't innovate because people aren't going to buy it.

>>57433459
A mysterious rare pepe is emerging from the trees

>>57440078
we already do that

File: SAM_0126.jpg (3MB, 4320x3240px) Image search: [Google]

3MB, 4320x3240px

I bought an aoc g2460pf. I'm not quite a fan of the colours. 'scuse the mess.

>tl;dr:
>increasing saturation

>>57436880
t. samsung developer

>>57444060
What's the point in local dimming if it doesn't let you actually have something dark next to something bright?

>>57444621
Have you calibrated it?

>>57444629
saturation != dynamic range

>>57434707
Actually HDR LCDs are pretty fucking decent, they get bright as hell and they use VA instead of IPS to achieve stupidly high static contrast ratios. In fact I haven't seen a single IPS HDR display, probably because shit contrast and IPS glow.

>>57438384
Samsung's Quantum Dot tech is pretty cool too. It's a normal IPS LCD display, but instead of a giant backlight on the back, each pixel is being illuminated individually by quantum dots.

>>57446590
VA crushes blacks though

>>57444621
how about a noose?

>>57433183
new Arcade Fire promos are looking great

>>57440739
>>57440535
Also the idiot poster is assuming that analog means == infinite bandwidth.

Color TV for example was shot and stored with chroma 1:1:! (4:4:4) to luma, but was broadcast at 4:2:0 - we've been doing 'this' since longer than digital data even existed.

Whats going to be the next meme to sell new tv's after 4k and "HDR"

>>57446698
>VA crushes blacks though
It doesn't.

>>57446590
See: Most every professional monitor to exist.
Even my old U2410 is 'HDR' (HDR is DCI P3/AdobeRGB or better - good luck finding even an OLED display that can do 100% of Rec.2020)

So, how is HDR stored in files?
Compressed down to 8 bits, the bits increased or some dynamic curve shit?

>>57446619
>It's a normal IPS LCD display, but instead of a giant backlight on the back, each pixel is being illuminated individually by quantum dots.
False.

Samsung has shown interest in potentially, for the TV market only, making individually addressable QDLED backlights.
For now, EVERY samsung QD display is using a single plane - not individually addressable backlight 'pixels'
Technically, even the older 9 zone 'local dimming' backlights of older Samsung LCDs is better.

Sony and LG are still to date the only companies who have made displays with individually addressable backlighting
>Sony on a few highend HX series TVs - all over 50", very expensive since it used 'regular' smd LEDs
>LG with their W-OLED TV's which have a full resolution white OLED panel fused directly behind an IPS panel.

>>57446744
HDR's requirement is:
1,000+ Nits for Whites and 0.05 Nits for Blacks. So no, your Dell monitor doesn't meet the requirements.

>>57446590
>VA instead of IPS
I prefer 1000:1 and accurate colors honestly

>>57446742
8K along with what ever else we think we need back from CRT that we used to have...

Maybe we'll even steal ideas from Plasma - 8K TV with 600hz sub-field drive?

>>57446744
>Even my old U2410 is 'HDR' (HDR is DCI P3/AdobeRGB or better
What the FUCK are you smoking?

>>57446816
For computer monitors and tablets sure, but for media consumption and phones, I rather have something like VA instead of IPS. Contrast is better than color accuracy for entertainment imo and VA consumes much less power compared to IPS and doesn't lose as much brightness when viewed at an angle.

>>57446798
There are two competing standards

One is SMPTE ST 2084, which is by far the more common one (used in UHDBD, youtube and pretty much every test clip I've seen so far). It's being pushed by Dolby, Samsung etc.

The other is ARIB STD-B67, which is doesn't seem to be as widespread, but it was mainly being pushed by ARIB and BBC.

ST 2084 introduces the PQ (perceptual quantizer) transfer function, which is a fairly complex, multi-segment function tuned to match the human visual response. This allows it to get the maximum amount of “quality” out of any X number of bits, but it requires a more complex decoder logic; and it also means that if you play it back on a ‘normal’ screen without appropriate conversion logic it will look like cancer. PQ is also defined on an “absolute scale”, meaning that a PQ-encoded value of Y will *always* map to an absolute brightness value.

STD-B67 introduces the HLG (hybrod log-gamma) transfer function, which is a much simpler, two-segment function. Below 0.5 it's a normal gamma curve (as with legacy systems), and above 0.5 it's a log curve (as often used by cameras, see e.g. Panasonic V-Log). Unlike PQ, it's not mapped against an absolute scale but against a relative scale (so a HLG-coded value of 0.5 will always map to your subjective reference brightness, which depends on the display), with everything above that being considered a super-highlight. Also unlike PQ, if you naively display it on a ‘normal’ screen, it won't look nearly as bad. It's also easier to implement.

The main benefit to PQ is the increased coding efficiency. The main benefit to HLG is the fact that it's intrinsically scaled to your subjective reference brightness, which makes implementation easier still.

>>57446837
Well I consume media on my computer monitor

Also for phones, viewing angles matter

>>57440636
Human eyes can only see 8-bit colors

>>57446877
Thanks anon, that's incredibly helpful.

To my dipshit brain, that means ST 2048 takes a set number of bits and spans the image over them, but super clever like, and STD-B67 is as if you took 8 bit, 0-255, but also let it go into the negatives and further than 255, and all the cheeky beaky stuff goes on in there.

File: maxresdefault.jpg (191KB, 1920x1080px) Image search: [Google]

191KB, 1920x1080px

i don't get it.

I can understand the deeper blacks, but like wtf why can't your tv just display the colors that are 'missing' ?

I can understand the need for a better camera to capture all the light, but why would I need a more expensive tv to display the video from the camera ?

>>57446949
Same reason you need a 64-bit processor and OS to support more than 4GB of RAM.

>>57446929
>To my dipshit brain, that means ST 2048 takes a set number of bits and spans the image over them
I don't think I understand what you mean by this

>STD-B67 is as if you took 8 bit, 0-255, but also let it go into the negatives and further than 255, and all the cheeky beaky stuff goes on in there.
Pretty much, although it only goes into the super-brights. Another way of thinking about HLG is like the brightness equivalent of scRGB, by simply allowing you to go “above” 255, you can encode super-brights and out-of-gamut colors for display devices capable of representing them.

(Meanwhile, standard gamut and standard-brightness displays would just soft-clip to the range they can represent)

>>57446949
Well, you're being strongly misled in that picture. If you actually want to understand any of this, I recommend staying as far away as possible from any consumer presentations (especially these nonsensical “comparisons”)

HDR vs SDR affects _nothing_ but the image brightness range (“dynamic range”)

What you are confusing for HDR here is the fact that the HDR generation of TVs *also* comes with improvements to the gamut, bit depth refresh rate and so on. In particular the “HDR10” standard, besides ST2084 (“HDR”) gamma, also mandates the rest of BT.2020 specifications

But just as you can have a HDR standard-gamut display, you can also have a SDR wide-gamut display. It's very important to distinguish between these two concepts for that reason.

>I can understand the deeper blacks, but like wtf why can't your tv just display the colors that are 'missing' ?
The reason why standard gamut displays can't display wide gamut colors is simply because they can't physically generate the required frequencies of light. The LEDs in the backlight itself simply don't emit those parts of the spectrum. When you buy a wide gamut device, you're buying a device that uses different (more expensive) phosphors that enable it to represent more frequencies.

The reason why SDR displays can't display HDR images is because of simple mechanical limitations on the output brightness range. Even if you turn up the backlight brightness, your dynamic range won't go up because of imperfections in the liquid crystal filtering. You can't have a 0.01 cd/m2 black pixel and a 1000 cd/m2 white pixel on the same screen with conventional LCD tech. HDR displays use more expensive backlight solutions that are capable of doing this.

>>57447046
That has absolutely fuck-all to do with anything

>>57446949
>I can understand the need for a better camera to capture all the light, but why would I need a more expensive tv to display the video from the camera ?
Imagine trying to display the output of a full-color camera on a black-and-white TV.

At best, you can invent a good algorithm for mapping the output of the camera into black and white hues, but at the end of the day, you're limited by what your display can output just as much as you're limited by what your camera can capture.

Without an upgrade to the TV set, you're simply never going to be able to experience a full color picture on your black-and-white screen. It's fundamentally impossible.

>>57447164
Don't worry, first thing was stupid.

That makes sense, thanks again.