FSI to Release 5,000 Nit Mastering Monitor this Fall

[jonpais]

FSI will be releasing the XM312U this fall – a 5,000 nit mastering monitor with 2,300 individually controlled LED backlight zones and priced at $22,000. It appears that we’ll be reaching the dizzying 10,000 nit peak brightness of PQ HDR sooner than most people think.

Keep in mind though, that with PQ HDR, anything above 100-200 nits is reserved for spectral highlight detail only – like reflections off the chrome of an automobile, fire, lightbulb filaments, backlighted clouds, etc. After all, HDR is not about increasing the brightness of the entire picture, but increased luminance range. Which is why it shouldn’t be necessary to reach for the remote to adjust brightness when switching between SDR and HDR content on TVs.

The average picture level (APL) of HDR PQ content should not depart radically from that of SDR content for the most part – and in fact, much HDR content actually has lower APL than the SDR version. The sample iPhone Dolby Vision video I posted yesterday is in fact a good example of how not to expose HDR – if watched in a darkened room, you can feel your pupils contract when viewed on an iPhone 12 Pro Max because it is too intense – which is fine for very bright scenes that call for it, but not so cool for an entire clip, let alone an entire film.

Interestingly, when it comes to SDR monitors, grade 3 displays are actually required to be brighter than grade 1 monitors:

Grade 1 Monitor:
70 to at least 100 cd/m2.
Grade 2 Monitor:
70 to at least 200 cd/m2.
Grade 3 Monitor:
70 to 250 cd/m2 or to 400 cd/m2 in adverse conditions.

Fun fact: 100 nits is already 50% of the 10,000 nit PQ HDR peak level and to the human visual system (HVS), the jump from 5,000 nits to 10,000 nits is much smaller than the leap from SDR 100 nits to HDR 1,000 nits.

Picture from LightSpace.

“Technical details aside, the most important thing to understand about HDR is that it doesn’t represent an enhancement as much as the removal of an artificial limitation. In the realm of human vision and physical light, high dynamic range is a default condition, not an added gimmick”. – Cullen Kelly​

700 nits is certainly enough to appreciate the enormous advantage HDR has over SDR, and the real reason most HDR content is not as impressive as it should be is not because consumer TVs can’t reach 10,000 nits, but because the overwhelming majority of shows continue to be lit in an SDR environment, they’re monitored in SDR, and the very first time someone sees the footage on an HDR display is in the grading suite.

“To date, most HDR content has been an emulation of the existing SDR experience. This is a new aesthetic and I’m intrigued when filmmakers will embrace that aesthetic rather than try to just reproduce the existing low-con films onto the HDR format.” – Peter Doyle, Supervising Colorist at Technicolor (2016)​

And it doesn’t end there. While colorists do have access to the tools to see HDR, that alone is no guarantee that the master file will end up preserving all the dynamic range, tonality and color envisioned by the director and the DP. For example, the decision in post-production to constrain the levels in the HDR pass to maintain consistency with the SDR version (for reasons both technical and aesthetic) can prevent HDR from taking wing. This is super common in the industry.

Not infrequently, a project gets the green light for an HDR master after the fact; both the post-production house and producer preemptively rule out a version that dramatically departs from the SDR version; the result being that HDR turns out to be little more than a marketing gimmick.

Another seldom discussed issue is that colorists are resorting to compromising their HDR grades in order to avoid judder artifacts.

HDR means not only lighting differently, but also (1) moving the camera differently, (2) framing differently, (3) exposing differently and (4) cutting differently. Panning more slowly for HDR to avoid judder artifacts. The 7-second rule no longer applies. Framing differently to either include or exclude intense light in the frame. Exposing to take the most advantage of the sensor’s dynamic range. Cutting differently, because juxtaposing bright and dark scenes in HDR is much different from SDR, where the overall brightness differences are insignificant. The editing suite is one place where a consumer OLED TV can actually do some good.

It’s been five years since Peter Doyle challenged filmmakers to begin creating films expressly for HDR, but the industry still lags behind.

 

[ahalpert]

Why is HDR more susceptible to judder artifacts?

 

[jonpais]

Why is HDR more susceptible to judder artifacts?

HDR high contrast scenes shot at 24p are vastly more prone to strobing at higher light levels than are SDR scenes. 24p judder and flicker are also more pronounced in HDR. Higher frame rate acquisition and display reduce the appearance of judder. When it comes to panning, the old 7-second rule (based on traditional theatrical viewing at 24 fps with a 180° shutter angle) no longer applies; twenty or more seconds now being required with the larger, brighter, more contrasty displays found in people's homes. By lowering contrast, colorists reduce the appearance of judder in HDR masters, while at the same time compromising the dynamic range home theater enthusiasts have come to expect.

 

[Tom Roper]

Fun fact: 100 nits is already 50% of the 10,000 nit PQ HDR peak level and to the human visual system (HSV), the jump from 5,000 nits to 10,000 nits is much smaller than the leap from SDR 100 nits to HDR 1,000 nits.

Fun but not a fact. A nit is a measure of screen luminance. It is light intensity, not luma, not gamma. It is linear light and it is absolute. 1 nit = 1 candela/sq-meter. 100 nits is not halfway to 10000. 10000 nits is 100 times brighter than 100 nits. What you are confusing is the PQ-EOTF curve. EOTF= "Electro to optical". In other words, the transfer of IRE input to actual brightness output, unlike conventional 2.2 gamma. 100 nits is at 1/2 the IRE for 10,000. The jump from 5000 nits to 10,000 nits is a linear doubling of actual brightness measured off the screen surface. 100 nits is 1/100th the way to 10,000 nits in light intensity.

 

[jonpais]

Fun but not a fact. A nit is a measure of screen luminance. It is light intensity, not luma, not gamma. It is linear light and it is absolute. 1 nit = 1 candela/sq-meter. 100 nits is not halfway to 10000. 10000 nits is 100 times brighter than 100 nits. What you are confusing is the PQ-EOTF curve. EOTF= "Electro to optical". In other words, the transfer of IRE input to actual brightness output, unlike conventional 2.2 gamma. 100 nits is at 1/2 the IRE for 10,000. The jump from 5000 nits to 10,000 nits is a linear doubling of actual brightness measured off the screen surface. 100 nits is 1/100th the way to 10,000 nits in light intensity.

Both fun and factual, Tom.

Due to the logarithmic response of the human eye to changes in light levels, 100 nits is actually around 50% of PQ based HDR's 10,000 nits peak level - as shown in the diagram that even a child can understand. The human eye does not perceive changes in light level in a linear manner. Humans do not perceive an increase from 100 nits to 200 nits as a doubling of brightness. There are four main HDR flavors in use currently. All of them use a logarithmic approach to retain the maximum amount of information in the highlights. The brighter we go, the more difficult it becomes to tell the differences. I’d have thought all photographers understood that simple concept. Apparently not!

 

[ahalpert]

HDR high contrast scenes shot at 24p are vastly more prone to strobing at higher light levels than are SDR scenes. 24p judder and flicker are also more pronounced in HDR. Higher frame rate acquisition and display reduce the appearance of judder. When it comes to panning, the old 7-second rule (based on traditional theatrical viewing at 24 fps with a 180° shutter angle) no longer applies; twenty or more seconds now being required with the larger, brighter, more contrasty displays found in people's homes. By lowering contrast, colorists reduce the appearance of judder in HDR masters, while at the same time compromising the dynamic range home theater enthusiasts have come to expect.

So you're saying that the same judder was always there, but that it's more pronounced in HDR?

 

[jonpais]

So you're saying that the same judder was always there, but that it's more pronounced in HDR?

Moire patterns are the result of spatial aliasing. In the time-domain, this same source of noise exists. It's known as temporal aliasing, and it's what causes judder in motion imaging. Increased resolution and the increased contrast of HDR make motion aliasing more pronounced.

Scroll down to page 5 of the link to see how brightness impacts judder and the panning speeds necessary to counteract it.
(Figure 2; Judder Equivalence vs. Luminance for Various Camera Pan Speeds)


https://www.pixelworks.com/public/tr...mes_of_Age.pdf

 

[Grug]

What kind of sadist fires 5,000 nits at their audience's eyes?

 

[hojomo]

What kind of sadist fires 5,000 nits at their audience's eyes?

"because we can" seems to be reason enough to subject people to all kinds of horrendous sh*t - so I wouldn't bet on this being any different

also they made live-action Transformers, so unnecessary high luminance-based imagery that permanently scars our retinas seems plausible

 

[Tom Roper]

Samsung has a 2020 model, 55, 65, 75 rtings tested to 5859 nits 10% window.

 

[ahalpert]

Isn't the 5000 nits just for the very brightest part of the screen?

I have a 2800 nits on-camera monitor. It doesn't really seem that bright outside (although bright enough to see in all but the most challenging conditions). During nighttime interiors it is painfully bright at maximum setting

 

[Tom Roper]

They test it in the center. I would think that’s probably the brightest part of the screen, or at least as bright as any other part.

 

[ahalpert]

They test it in the center. I would think that’s probably the brightest part of the screen, or at least as bright as any other part.

But like only the sun or clouds or incident light sources or specular reflections off metal will generate that might light on the screen right? I would think that some scenes would only use 400 nits or less

 

[Tom Roper]

Sustained 100% window, 1875 nits.

 

[rob norton]

Isn't the 5000 nits just for the very brightest part of the screen?

I have a 2800 nits on-camera monitor. It doesn't really seem that bright outside (although bright enough to see in all but the most challenging conditions). During nighttime interiors it is painfully bright at maximum setting

Do you change settings when you go from interior/exterior?

 

[agcohn]

Speaking from a Post-Production perspective when it comes to feature films these days, almost all HDR content is creatively rolled off somewhere between 500-600 nits via the HDR LUT with a hard clip at 998 nits for the very brightest highlights only.

I do not know any big time feature film colorists who would choose to allow the image over 1000 nits, as that is already considered too bright by almost all colorists / filmmakers. Episodic TV / Broadcast might work differently though, but I’m not familiar with that work.

 

[jonpais]

Speaking from a Post-Production perspective when it comes to feature films these days, almost all HDR content is creatively rolled off somewhere between 500-600 nits via the HDR LUT with a hard clip at 998 nits for the very brightest highlights only.

I do not know any big time feature film colorists who would choose to allow the image over 1000 nits, as that is already considered too bright by almost all colorists / filmmakers. Episodic TV / Broadcast might work differently though, but I’m not familiar with that work.

Dolby Cinema is only 108 nits.

 

[ahalpert]

Do you change settings when you go from interior/exterior?

I change settings all the time. Typically full brightness outside, maybe a little less at dusk. Somewhere between full brightness and midway for daytime interiors with windows. Somewhere between midway and probably 10% from minimum for dimly lit evening interiors.

The monitor is touchscreen (Desview R6) and has an absolutely wonderful feature where you simply swipe up and down on the left half of the screen to adjust brightness. I use it all the time, including to dim the monitor to save battery during breaks in shooting.

Crucially, I don't rely on the look of the monitor to judge exposure but toggle false color on and off. It has both Arri and "spectrum" false color. So, dimming the brightness up and down has no bearing on my exposure judgement

 

[agcohn]

Dolby Cinema is only 108 nits.

Dolby Cinema (DV108) is not HDR technically but EDR (Extended Dynamic Range), as it is essentially the same brightness as SDR R709. Dolby Cinema is brighter than typical P3 theatrical content, which is limited to 48 nits / 14fl, but it is not HDR.

 

[jonpais]

Dolby Cinema (DV108) is not HDR technically but EDR (Extended Dynamic Range), as it is essentially the same brightness as SDR R709. Dolby Cinema is brighter than typical P3 theatrical content, which is limited to 48 nits / 14fl, but it is not HDR.

Just as your claim about grading films for theatrical distribution to 1,000 nits was misleading, so is your contention that Dolby Cinema isn't HDR; and it demonstrates your complete lack of understanding about what HDR is. HDR encompasses both brighter highlights as well as deeper blacks. HDR is about a greater range of luminance, not increased overall brightness. That is the reason Ultra HD Premium certification of displays has different standards: 1,000 nits with a black level of less than 0.05 nits (LCD) and 540 nits with a black level of less than 0.0005 nits (OLED). My iPhone 7 is 700 nits but it isn't HDR! APL of both HDR and SDR should be comparable, but in practice, the APL of HDR is not infrequently lower than SDR. Dolby Cinema delivers 500X the contrast ratio and double the brightness of traditional movie theaters.