A chance to talk to FlatpanelsHD's reviewers.
#12184
Torben Rasmussen wrote:DX900 will remain throughout 2017 as the top LCD model. Coverage was about the same,
stop blaming on the programs used
even a child can see that the DX900 has a wider colorspace than the OLEDs
why is it so hard to admit?
#12196
You are a legend sir. Thanks. I always look out for expert users that pick up on things. What is your work?

I literally have a Hollywood colorist that stalks me online, because he can't admit that so etching like a 900 is worth using on the low end, and everything below $20k is seemingly crap. Some of these guys seem to be types who wound neuroticaly dance around in a single percentage point of difference, been there can pick it, but know 10-20% is also good on low end (not talking about panels but difference between color choices picked). But still a lot of yuck stuff gets released to cinema. :( I prefer the some colourful look of the 90's 00's, and the xfiles and matrix look. But they take things too far I'm the wrong directions often enough. What you have said facinates me, and is what I suspected. Thank you.
#12197
But a question, this 110% dci, what does that mean, is it really usable correctly?

I'm interested in the Rec2020 color space throughout the brightness range (a good graph of saturation changes with brightness would be nice on reviews). But at some high brightness the three colours must be at good performance before the accuracy drops off. That would be a handy graph to have in reviews to calibrate too that point.

Now, I could see the wonderful rendering in the red channel you mentioned, made the panel seem a lot better than it was on those images. But what coverage of usable rec2020 do you think it can achieve in real use Pgice?

I forgot to mention, this HDR issue, is why I was hoping for a new 900 model that used Panasonic new double shuttered LCD technology. I watch in a lot room normally, so the technologies ultra low black level is good enough compared to OLED, which means it would be useful for hdr rec2020 viewing compared to OLED. OLED is pretty lame. Thete is one way they could increase the OLED brightness dramatically, should be guaranteed, but I'll keep that to myself.
#12198
Torben. The difference between some of the gamut coverage results on the 900 (I suspect the 58 inch is a different panel) and 950, could it be light in different review rooms? I've done some design investigation, and controlling light flow differently for opposing directions is tricky. It took me a while to come up with a directional light valve, but was soundly beaten to it anyway. :( Actually, that's not true there was something I came up with years and years ago, but I suspect I probably have been beaten to that too :( :( (but it wasn't really a desirable light valve).

What happens is that a set passes light out, but also passed light in which interacts and bounves back out washing color purity a bit, reducing gamut. Even.the light from.he screen a itself bouncing off objects back into it could do this. Which means the surounding body of the calibration equipment itself could do it. Wow, which is bad.

For an example. The lsservue rear projection TV got around this by using a 3M glasses bead material, where the light was collected on the exposed rear of a bead embedded in black material, and funneled out through the tip of the bead poking out he front of the material. The black surface on the front was much larger than the back, reducing the amount of light that could get in through the beads from the front. If one wanted expand I is a wider spreader mechanism in front of the bead, the nature of optics is that it probably would let a lot more light in through the front. So, the pinprick pixels of that screen must have been a bit uncomfortable up close.

Could this be a factor?

I'm also interested in how development of the rec2100/ HDR test and calibration tools are working out in the industry?

Thanks.
#12202
Steve: We are considering adding color volume measurements (and include a graph in reviews) but we need to invest in new equipment before we can do it. We had hoped that CalMAN 2017 would be a little more flexible but it has quite strict requirements.

Regarding the differences in measurements, it's hard to say for sure but be aware that there are factors to consider here. Brightness on OLED drops if a larger portion of the screen is covered in white (as APL goes up). On the other hand, LCD cannot reproduce the full color space at low luminance levels, simply because they need to filter out too much light and due to reduction in backlight intensity levels. I have in the past referred to this as "inverse APL". See my hands-on articles from CES:
- Samsung "QLED": http://www.flatpanelshd.com/news.php?su ... 1485251457
- LG 2017 OLED: http://www.flatpanelshd.com/focus.php?s ... 1484640539

There are no TVs that cover 100% DCI-P3 color volume yet. This is misleading because it depends on how you actually measure the values. It is easy to measure higher levels simply by using other APL patterns. Using a 80-100% APL pattern to measure on both LCD and OLED, for example, will give you crazy results (and it can be used to mislead, just like Samsung did at CES when it demonstrated live color volume measurements to journalists - many spectators fell for it).

So, these are the facts right now:
- LCDs can go brighter but sacrifize black levels and color volume at the low end to do so. They don't have very high inter-scene contrast / peak brightness either. For example 65" DX900 has 512 zones. If you divide that by 3840x2160 pixels you will realize that each zone covers 16.200 pixels. It's simply not possibly to reproduce bright stars at >1000 nits while retaining a dark night sky. The zone has to either stay dim (faded stars) or get a lot brigther (ruins black levels).
- OLEDs can retain full color volume up to a certain break-off point for peak brightness. However, they cannot go as bright as the brightest LCDs. You have pixel-level controls so each star can sparkle like it was supposed to while the night sky stays black.

In our opinion, HDR looks better on OLED. We will soon introduce a HDR comparison chart that includes all TVs tested in 2016 and 2017 (and onwards). It will give you some clarity on the subject.
#12203
Thanks Rasmus. PGICE used an open source software for his measuments, hardware I don't know.

OLED gets muddy in one way or another. How bright are the stars compared to a 900 LCD value it can achieve without too much of a halo (like 1-5%). Maybe that would be acceptable. But how often do we see stars, or view stars, compared to general well lit images that the 900 seems to be good at. Even on fire based lights in a scene, would the 900 array just add a lustre to them? So the issue is can an LCD do HDR much better at higher brightness without too much compromise? Even on a black background, bright natural deep colored objects may not be bright enough to break the led array in HDR. So, it's a compromise, that OLED suffers due to not only limited HDR, but severely limited depending on amount of screen, and severely limited as PGICE pointed out in color saturation at hdr levels.

But the reason I was hopeful of a new 900 LCD, was not only color space improvements, but hoping for the twin shutter LCD technology bringing black levels down to nominal leveks, which means that the led array will bleed through much less, and even 4000nit arrays might be possible. Now, that would break the TV market, if Panasonic could sell Bluray 4k ready TV with 4000 nit HDR before anybody else could manage to perfect it. Hopefully Panasonic realses a real market advantage to place it back into the market.

Anyway, bring in the post rec2100 hdmi 2.1 4000 nit Dolby vision 10k 16 bit TV's with customisable 3D luts! ?
#12204
I forgot, +100% gamuts can be misleading if it doesn't have full coverage but you add the areas outside of it ontop to give above 100%

The 900's were nice, even though not quiet 100% DCI, it was enough, and the outside color could have added to rec2020 (is it accessible?).
#12206
I don't recall ever seeing an in-depth review of the Vizio Reference TV. Have anyone tested Vizio's claims? I recall that TCL - or maybe another Chinese brand? - claimed something close to 90% Rec.2020, too, but the TV was never released, as far as I'm aware. One of the issues is that to achieve this on you need to use a special type of quantum dots based on cadmium, which is highly toxic to humans.

Of course, both technologies have limitations. HDR is a longterm project that will take years to fully realize. To me, HDR looks better on OLED than the best FALD LCDs, mostly because the OLED can retain pure blacks regardless of the color or luminance. That's essentially what high dynamic range is and based on the fair portion of movies, TV series, and games that I have seen and experienced in HDR, these advantage in contrast shows more often than not. Something as simple as subtitles is enough to show heavy blooming even on the best FALD LCD TVs.

But yeah, fingers crossed for a 4000-nits LCD based on Panasonic's dual shutter. If it happens, it may be revealed at IFA in a few weeks but I doubt it.
#12208
Oh, please, let it be relessed. To me rec2020 and HDR are to musts to check for work. Professional colorist businesses ussually only use the LG OLED at all as a client screen, to show clients, and for set rec709 work etc. So viewing and checking, the 900 offers an edge, but not enough. A surrounding black.might suffer, but this panel can push higher with correct color rendering of detail, high.kesns something (if you can ignore the bleed). As long as you can set it to do HDR at 540-740, that's acceptable fory purposes. The issue is, is the bleed that bad at 540 etc?

Now the VIZIO and the Hisense or TLC I would not be surprised if it was the same panel etc. Remember, the Chinese wanted to do the fiddly THX certification, which might have been an issue, but VIZIO wouldn't face such an issue.

I literally say, standing in front of a 900 in the right mode literally makes me feel good compared to other LCD TV's. The luminance contrast and color off rolling toning is fantastic (better if 100% rec2100) and the brightness must be good for 3D content. I saw velarian and the city of a thousand planets today, and the cruddy dim likely P3 3D projection really diminished a great movie. Oh, if so could have seen it on a bigger laser screen.

A dual shutter 90%+ rec2020 900 would be great. Rather than over 3000:1 contrast you might get a real 1 million to one. I'd have to be in a pitch black room with seating stars to be bothered, and stars are not meant to be that bright. Actually, apart from subtitles, which I'm not concerned about too much) the solution is a mode that pre compresses bright high contrast image elements against large low light areas down so bleed through is reduced or eliminated. Calculations could be done before each frame is displayed. The reality is the eye is going get plenty brightness from such a scheeme against black anyway, and is going to loose detail top or bottom if too bright. Doing so e calculation, you can probably calculate level of acceptable fald bleed near invisibility in the fald zone near an bright area to tweek the image for more brightness before it gets too noticeable. With such a scheeme you can choose to make it unoticable or bleed through not happen at all in settings. So, even the 900 could be further twerked for increased performance by software update.

Rasmus, the 900, what sort of backlight precisely do they use, was it the phosphor coated led? I have forgotten precisely what I read.

Thanks

Again.

Btw: do how far away are the IMAX headquarters from where you are?
#12209
I've just to add. About the blue/violet OLED thing with QD colour filter. I don't know what the peak brightness of theses OLED colors will be, but even if they are as much as white is now, that is still in the 1000 HDR range. I don't know for the LG fits can enhance the color purity of QD but certainly could on the Panasonic and the dual shutter version. However, it might not be suitable for much more. LG is a bit quiet on their RGB filter. They say that wide color gamut primaries make the screen dimmer, and have used white fur that reason. But what are they planning?

There was announcement of QD color filter development, which should have helped reach near 100% rec2020, as most light altering components would be in the oath behind it. But I didnt see this BASF brand a filter come out. Maybe there was an issue, maybe the Chinese were waiting for it. Somebody could have locked away certain technologies from them under licence. Just possible causes of non delivery. However, now there is another company announcing it.

Tantalisingly close.

Now, the hopes of led panels and OLED, printable, flexible. Led panels might just be more expensive. OLED has been more expensive. The hope of cheaper printable OLED and bendable. Bendable OLED has in the past, less color gamut by far. Printable, I forget if it is different, but printable bendable is what I hear about years back. So, will it get any good color, will it be cheap (hope so) will it be reliable longterm)? But unless brightness goes up, you might not get to 1600-4000 HDR, even with a QD colour filter. With micro led, again we are talking half a millimetre pixels so 4 meters for an 8k screen until they can bring them down. So, all these panel types might face limitations. I might like an 4-5 meter 8k 3D seemless tile panelled screen, but it might not be cheap. Again, also reliability of less.

Direct drive quantum dots have suprisingly proved shot lived and fragile, a lot less than even very old OLED tech. Hopefully some solution may come up. So, while they can do 10k nits theoretically, how long, what price? OLED simply took masses amounts of time to over come these, decades. I hope it doesn't.