NEW JVC RS3000/NX9 RS2000/N7 RS1000/N5 Native 4K Projectors Anticipation Thread - Page 47 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 13017Likes
Reply
 
Thread Tools
post #1381 of 13653 Old 08-31-2018, 03:50 AM
AVS Forum Special Member
 
christoffeldg's Avatar
 
Join Date: Mar 2012
Posts: 1,375
Mentioned: 19 Post(s)
Tagged: 0 Thread(s)
Quoted: 1310 Post(s)
Liked: 527
Quote:
Originally Posted by Highjinx View Post
Thought that might be the case with the chip polishing.

Now for the low down on how/which methods are being used for the contrast hike.
What do you mean with chip polishing?

Video: Sony VPL-VW760ES, Elite screen Aeon 135" Cinewhite + JVC X7900, Magicscreen Reference ALR 120"
Speakers: Bowers and Wilkins 802 D3 front, JBL 580, JBL 520c, JBL 550p
Amplifiers: Lyngdorf stereo TDAI 2170, Lyngdorf SDA 2400, Denon 4300H Home Theatre
Equipment: PC/PS4/Xbox One/Switch/Synology 2415+
christoffeldg is offline  
Sponsored Links
Advertisement
 
post #1382 of 13653 Old 08-31-2018, 03:55 AM
AVS Forum Special Member
 
Highjinx's Avatar
 
Join Date: Dec 2002
Location: Australia
Posts: 3,546
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 1320 Post(s)
Liked: 535
Quote:
Originally Posted by christoffeldg View Post
What do you mean with chip polishing?
See from Javs post below it's a mechano/chemical polishing method if I'm not mistaken:


Quote:
Originally Posted by Javs View Post
I dont have the meters but:

  • THX 4K DISPLAY Certification for DLA-NX9 - in the process of being obtained
  • With Clear Motion Drive motion control technology, 4K60P (4: 4: 4) signals are supported, improving image flow and reducing blur. For its part, the "Motion Enhance" attenuates the kinetic blur by optimizing the training of D-ILA matrices. Combined together, these two technologies provide a high level of detail.
  • "Installation mode" storing up to 10 installation configurations (lens position memory, pixel adjustment, screen mask, etc.)
  • Compatible with auto-calibration functions, optimizing image quality according to installation conditions and compensating for any breaks in color balance when using the projector for a long time.
  • "Color optimization mode", corrects the color balance that may vary depending on the screen used
  • "Low latency mode", reducing offsets from the source
  • Optimized configuration of circuits that accelerate the time required to recognize the signal compared to previous models.
Also:

New generation of D-ILA matrix
The latest of the D-ILA matrices in native 4K of 0.69 "with a pixel pitch of 3.8 μm, allows to display a resolution of 8.8 million pixels (4096 x 2160 pixels). Defects such as scattering and diffraction of light are greatly reduced by the combination of optimized planarization technology and the matrix.



http://www.passionhomecinema.fr/blog...-natif-17999e/

May the success of a Nation be judged not by its collective wealth nor by its power, but by the contentment of its people.
Hiran J Wijeyesekera - 1985.
Highjinx is offline  
post #1383 of 13653 Old 08-31-2018, 03:55 AM
AVS Forum Special Member
 
Wookii's Avatar
 
Join Date: Aug 2008
Posts: 3,574
Mentioned: 60 Post(s)
Tagged: 0 Thread(s)
Quoted: 3024 Post(s)
Liked: 2228
Quote:
Originally Posted by Manni01 View Post
Or the Z1 chassis might disappear entirely if they can fit both the lens and laser in the new one.

Or they could keep both until they bring a unified third chassis for the whole range.

There is no way to know, and it’s pretty much irrelevant. We know what we can get now, they won’t tell us what we’re going to get in the next 1-5 years.
I hope they do lose the Z1 chassis, its just too big and bulky. The Z1 was well out of budget for me, but even if I could have afforded it, it would have been too big to house in my room properly.
Highjinx, phara and Manni01 like this.
Wookii is offline  
Sponsored Links
Advertisement
 
post #1384 of 13653 Old 08-31-2018, 04:00 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,117
Mentioned: 329 Post(s)
Tagged: 0 Thread(s)
Quoted: 5443 Post(s)
Liked: 5649
Quote:
Originally Posted by Highjinx View Post
See from Javs post below it's a mechano/chemical polishing method if I'm not mistaken:
javs post is an automatic translation of a French blog post that translated an english brochure that was itself most likely translated from Japanese

I really wouldn’t pay any attention to the words in this post. Or even the information. The blogger has simply copied and pasted generic marketing info and added some of his own poetic license. It’s not an accurate/reliable source of information (nothing to do with Javs of course, who usually is). I speak French and it’s not a technical article, just a generic news rewording the specs sheet and some marketing brochure with pretty much zero new info except the horizontal/vertical lens shift range, that I asked Ran/Nigel to confirm with JVC in our list of questions because it conflicts with the values in the specs sheet.
Highjinx likes this.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders

Last edited by Manni01; 08-31-2018 at 04:18 AM.
Manni01 is online now  
post #1385 of 13653 Old 08-31-2018, 04:09 AM
AVS Forum Special Member
 
Highjinx's Avatar
 
Join Date: Dec 2002
Location: Australia
Posts: 3,546
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 1320 Post(s)
Liked: 535
Quote:
Originally Posted by Wookii View Post
I hope they do lose the Z1 chassis, its just too big and bulky. The Z1 was well out of budget for me, but even if I could have afforded it, it would have been too big to house in my room properly.
Don't know if you remember Projection Design, they had the right idea, a small head unit(magnesium) and the light engine was separate connected via an optical cable, so did RED with their laser proto, great idea, but would not likely happen.

May the success of a Nation be judged not by its collective wealth nor by its power, but by the contentment of its people.
Hiran J Wijeyesekera - 1985.
Highjinx is offline  
post #1386 of 13653 Old 08-31-2018, 04:15 AM
AVS Forum Special Member
 
Wookii's Avatar
 
Join Date: Aug 2008
Posts: 3,574
Mentioned: 60 Post(s)
Tagged: 0 Thread(s)
Quoted: 3024 Post(s)
Liked: 2228
Quote:
Originally Posted by Highjinx View Post
Don't know if you remember Projection Design, they had the right idea, a small head unit(magnesium) and the light engine was separate connected via an optical cable, so did RED with their laser proto, great idea, but would not likely happen.
Yes, I remember those. It would be a great idea in theory - pick the level/quality of head unit you want with the lens, chips and processing in it, and then select the level of laser light source you want, 2000 lumen, 3000 lumen, 4000 lumen etc in a separate box you can install in the garage 20m away lol
Highjinx likes this.
Wookii is offline  
post #1387 of 13653 Old 08-31-2018, 04:21 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,117
Mentioned: 329 Post(s)
Tagged: 0 Thread(s)
Quoted: 5443 Post(s)
Liked: 5649
Quote:
Originally Posted by ARROW-AV View Post
Heres a titbit... ANSI contrast is increased by circa 50%

Great news. Is this what you measured, or is it as per JVC? Is it for all 3 models or just for the flagship N9?
Mike_WI and cemo62 like this.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #1388 of 13653 Old 08-31-2018, 04:32 AM
aka jfinnie
 
Join Date: Aug 2015
Location: Norwich, UK
Posts: 3,271
Mentioned: 54 Post(s)
Tagged: 0 Thread(s)
Quoted: 2642 Post(s)
Liked: 1740
Quote:
Originally Posted by Wookii View Post
I don't thinks its as clear cut as that. Sure there are many in the latter camp, and a handful in the former, but I would imagine a larger proportion of users on this forum would bridge both camps. I calibrate my displays within an inch of their lives, and I have invested perhaps a disproportionate amount of money on a meter to ensure I can achieve that easily, accurately and repeatably.

That however doesn't stop me making judicious use of any image processing that provides a perceived enhancement to the viewing experience. I am a huge fan of Darby processing (and its such a shame they don't seem able to produce a 4K solution) and use that on 1080p content in my Lumagen Pro. The sharpening settings in the Panny blu-ray players add a noticeable improvement in perceived sharpness on UHD discs, and you will find most self-calibrating owners on this forum turning them on to a greater or lesser extent. Likewise RC on the Sony PJ add a noticeable improvement in perceived sharpness and fine detail, so I use that also as long as it doesn't add visible artefacts. You'll also find many madVR users here producing full 3D LUT's but also taking advantage of the softwares advanced processing.

At the end of the day, to me a calibration should involve utilising all the tools available to optimise the perceived image to give the most compelling and engaging viewing possible in the system.
I perhaps have a slightly different viewpoint to many as I live day to day in the professional world in test and measurement and so shy away from all these things which are never present on professional product. In my world calibration to me means the quest for absolute accuracy so I can discover the source fully.

Anyone who has calibrated themselves has probably done the same as you at some point - I know I have and still do! I have had Darbee courtesy of various Lumagen units (and also the PrismaVue processing from the Prisma LUT box) for the last 3 years or so; and I always go through the same process with it. I like the effect as I turn it on - it clearly changes the image in a way which at that point of switching it in circuit I seem to like. But it only seems to survive for a little while before I get tired of the effect and decide that while it is doing "something" I've no indication it is getting me closer to the original intent and so it gets turned off again.

I've seen some good pics out of MadVR and some that look like a dog's dinner thanks to too much processing (there is one (in)famous poster with an old JVC who's name I shall not mention as they're a bit like Beetlejuice... lol... but he'll probably be along as soon as a screenshot emerges :-p).

Quote:
Originally Posted by christoffeldg View Post
IMO this is very unfair and makes people who like good software processing solutions look like bumbling idiots.
I'm sorry you feel that way but it is not the intention; for me it is merely a truism that the use of these algorithms can only reflect preference - which is often not in the interest of accuracy (see torch mode TV retail example). It's clearly different but have you seen the original source material to know if it is closer to it? Your reference when you enable it an look on the screen is a self-reference not to an original master.

I would myself love to see some real science debate and unfortunately the world of these algorithms is almost completely devoid of it; until such research happens I'll continue to call them out for unsubstantiated claims; you just have to look at the kind of BS pixelated improvement images we see in the advertising for all these algos. Sony aren't alone in this; everyone does it.

An example of testing that would be really interesting (in the vein of the video I linked to); take some >4K images, process both direct to 4K and also as if via a 2K DI; pack it for transmission (ie encode at 4:2:0 with UHD compression). Then process via choice of image processing algorithms (eg MPC/Reality Creation/MadVR) and then compare the processed and unprocessed results with the original pre-compression masters sent to the display at full resolution 4K 4:4:4 and see how they fair. This would be really interesting and would further the understanding of the capabilities and limitations of such processing no end.
Gary Lightfoot and CrazyEddie like this.
bobof is online now  
post #1389 of 13653 Old 08-31-2018, 04:42 AM
AVS Forum Special Member
 
markmon1's Avatar
 
Join Date: Dec 2006
Posts: 6,223
Mentioned: 106 Post(s)
Tagged: 0 Thread(s)
Quoted: 5227 Post(s)
Liked: 3400
Quote:
Originally Posted by bobof View Post

An example of testing that would be really interesting (in the vein of the video I linked to); take some >4K images, process both direct to 4K and also as if via a 2K DI; pack it for transmission (ie encode at 4:2:0 with UHD compression). Then process via choice of image processing algorithms (eg MPC/Reality Creation/MadVR) and then compare the processed and unprocessed results with the original pre-compression masters sent to the display at full resolution 4K 4:4:4 and see how they fair. This would be really interesting and would further the understanding of the capabilities and limitations of such processing no end.
I dont care if it looks as intended or not. The director that created the content isn't sitting in my room, I am. Reality creation adds a ton of detail that I'd otherwise miss. I love it. I find it far superior to Darby, which was more of a sharpener and even on low was much more extreme and looked fake. Like I said, if Sony would sell me an HDMI Reality Creation device that plugged inline in my HDMI and did only Reality Creation, I'd pay up to $2k for it. I like it that much.
farsider3000 and Archibald1 like this.

Video: JVC RS4500 135" screen in pure black room no light, htpc nvidia 1080ti.
Audio: Anthem mrx720 running 7.1.4, McIntosh MC-303, MC-152, B&W 802d3 LR, B&W HTM1D3 center, B&W 805d3 surround, B&W 702S2 rear, B&W 706s2 x 4 shelf mounted for atmos, 2 sub arrays both infinite baffle: 4x15 fi audio running on behringer ep4000 + 4x12 fi audio running on 2nd ep4000.
markmon1 is offline  
post #1390 of 13653 Old 08-31-2018, 04:48 AM
703
Senior Member
 
703's Avatar
 
Join Date: Mar 2006
Location: New Zealand
Posts: 354
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 152 Post(s)
Liked: 49
Quote:
Originally Posted by Manni01 View Post
Quote:
Originally Posted by 703 View Post
I reckon JVC will continue selling their current line-up, alongside the new ones, simply because some people will prefer the higher contrast ratios, and the smaller chassis.

Mike has confirmed the current range was end of life, and stocks are running low (at least for 540), so if you want a new one, don’t wait. There will be plenty of used ones soon though [IMG class=inlineimg]/forum/images/smilies/smile.gif[/IMG]
The only issue is that if the new projectors dont bring out a better picture compared to the old on everyday 1080p material, then it will hurt their sales as people who have been sitting on the fence wanting to upgrade to the “JVC” experience cannot buy a ultra high contrast projector anymore.

So they will wait or just buy second hand JVCs.
703 is offline  
post #1391 of 13653 Old 08-31-2018, 04:53 AM
aka jfinnie
 
Join Date: Aug 2015
Location: Norwich, UK
Posts: 3,271
Mentioned: 54 Post(s)
Tagged: 0 Thread(s)
Quoted: 2642 Post(s)
Liked: 1740
Quote:
Originally Posted by christoffeldg View Post
That puts it firmly in the ~500 range, that's much better than what the Sony is capable of.
I think the max I've seen is 290:1 under optimum conditions on x7x00 (iris full open, max zoom). I don't think X9x00 had any advantage but can't find numbers... http://cine4home.de/heute-bei-uns-di...a-x5500-x7500/
290 +50% = 435:1
bobof is online now  
post #1392 of 13653 Old 08-31-2018, 04:55 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,036
Mentioned: 490 Post(s)
Tagged: 0 Thread(s)
Quoted: 6922 Post(s)
Liked: 6629
Quote:
Originally Posted by Bytehoven View Post
500-600 range?

Edit... I'd guess 543 ANSI, for the win.
More like 450.
David Mathews, Manni01 and bobof like this.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #1393 of 13653 Old 08-31-2018, 04:55 AM
Advanced Member
 
12GAGE's Avatar
 
Join Date: Nov 2005
Posts: 737
Mentioned: 24 Post(s)
Tagged: 0 Thread(s)
Quoted: 351 Post(s)
Liked: 394
Sony did the owners a great service with this campaign. I agree it was excellent customer service from them in this instance. Technology is moving so quickly it may not be realistic to expect these type of upgrades in the future. It would however be quite nice if designs were more modular where these type of upgrades were easier. Similar to what projectiondesign had on some of its high end dlp units.


Quote:
Originally Posted by HoustonHoyaFan View Post
Quote:
Originally Posted by Jmouse007 View Post
...
Another example are those who plopped down over $25,000 dollars for the 4K Sony 1000ES, only to discover it wasn't ready for prime time and couldn't even play native 4K without shipping it back to Sony for a "hardware update"! Hence the need for the Sony 1100ES...
That is a very bad example to use for what was a stellar piece of customer service by Sony.

The VW1000es shipped before widespread implementation of HDMI2.0/HDCP 2.2. When the more expensive VW1100es shipped, Sony offered VW1000es owners a $2500 upgrade to the VW1100es. Sony sent a tech to your house to basically do a motherboard swap which turned your VW1000es into a VW1100es with not only HDMI2.0 but also improved processing chips for stuff like RC.

Lets see if in 18 months when HDMI 2.1 becomes widely available if JVC or Sony offers current customers a similar upgrade path.
Archibald1 likes this.
12GAGE is online now  
post #1394 of 13653 Old 08-31-2018, 04:58 AM
AVS Forum Special Member
 
Highjinx's Avatar
 
Join Date: Dec 2002
Location: Australia
Posts: 3,546
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 1320 Post(s)
Liked: 535
Quote:
Originally Posted by 703 View Post
The only issue is that if the new projectors dont bring out a better picture compared to the old on everyday 1080p material, then it will hurt their sales as people who have been sitting on the fence wanting to upgrade to the “JVC” experience cannot buy a ultra high contrast projector anymore.

So they will wait or just buy second hand JVCs.
Only those who use the longest throw and -15 will get the best out of the current projectors. The new ones won't have the stellar contrast of the outgoing ones, but already ANSI is 50% higher. I do doubt their claim of the massive increase in contrast compared to the RS4500 for example, I have a niggling feeling they are using some lamp light blocking technique, but we'll see.

On the whole the new series on balance is likely to be a step up for most. As time goes by no doubt they will catch up to the outgoing generation contrast wise.

May the success of a Nation be judged not by its collective wealth nor by its power, but by the contentment of its people.
Hiran J Wijeyesekera - 1985.
Highjinx is offline  
post #1395 of 13653 Old 08-31-2018, 05:05 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,036
Mentioned: 490 Post(s)
Tagged: 0 Thread(s)
Quoted: 6922 Post(s)
Liked: 6629
Quote:
Originally Posted by bobof View Post
An example of testing that would be really interesting (in the vein of the video I linked to); take some >4K images, process both direct to 4K and also as if via a 2K DI; pack it for transmission (ie encode at 4:2:0 with UHD compression). Then process via choice of image processing algorithms (eg MPC/Reality Creation/MadVR) and then compare the processed and unprocessed results with the original pre-compression masters sent to the display at full resolution 4K 4:4:4 and see how they fair. This would be really interesting and would further the understanding of the capabilities and limitations of such processing no end.
In the case of MadVR to see how close it got to the real 'truth' I have done exactly this...

I took a control 5.7k holiday image and shrunk it down to 2160p, then I output a specific 1080p version of the same file using the same 'master' as the source.

Then what I did is compared them all using MadVR and essentially vanilla up-scaling from Kodi, which I think is nearest neighbour or something fairly default, then I also up-scaled using the MadVR NGU settings I use for 99% of my viewing and compared them all against the original master image. Doing this of course we know what the original would look like in true 4k as it were, and we can see how good/close up-scaling can get to that original.


Control UHD Master vs 1080p Vanilla Up-scale

http://screenshotcomparison.com/comparison/119475


1080p Vanilla Up-scale vs MadVR NGU Up-scale

http://screenshotcomparison.com/comparison/113573


MadVR NGU Up-scale vs Control UHD Master

http://screenshotcomparison.com/comparison/119474


Indeed NGU comes very close to the original, but not quite there yet. I should note the Control UHD is still passing through MadVR, but all its doing is Chroma up-sampling, no sharpening.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #1396 of 13653 Old 08-31-2018, 05:10 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,036
Mentioned: 490 Post(s)
Tagged: 0 Thread(s)
Quoted: 6922 Post(s)
Liked: 6629
Quote:
Originally Posted by markmon1 View Post
I dont care if it looks as intended or not. The director that created the content isn't sitting in my room, I am. Reality creation adds a ton of detail that I'd otherwise miss. I love it. I find it far superior to Darby, which was more of a sharpener and even on low was much more extreme and looked fake. Like I said, if Sony would sell me an HDMI Reality Creation device that plugged inline in my HDMI and did only Reality Creation, I'd pay up to $2k for it. I like it that much.
Why dont you just buy the Sony UHD player, its built into that is it not?

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #1397 of 13653 Old 08-31-2018, 05:13 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,757
Mentioned: 530 Post(s)
Tagged: 0 Thread(s)
Quoted: 2620 Post(s)
Liked: 3530
Quote:
Originally Posted by ARROW-AV View Post
Heres a titbit... ANSI contrast is increased by circa 50%
I was secretly hoping for that, but not really believing in it. So that sounds like great news! However, IIRC JVC has promised ANSI contrast improvements in the past which didn't really play out. So I'll keep my hat on for now, until measurements actually confirm this. But I'm hopeful!

Hmmmm... As Manni asked, though, is this only for the NX9? E.g. I can imagine the better lens helping with ANSI contrast. Or is this also for N5 and N7?

Quote:
Originally Posted by bobof View Post
I have had Darbee courtesy of various Lumagen units (and also the PrismaVue processing from the Prisma LUT box) for the last 3 years or so; and I always go through the same process with it. I like the effect as I turn it on - it clearly changes the image in a way which at that point of switching it in circuit I seem to like. But it only seems to survive for a little while before I get tired of the effect and decide that while it is doing "something" I've no indication it is getting me closer to the original intent and so it gets turned off again.

I've seen some good pics out of MadVR and some that look like a dog's dinner thanks to too much processing (there is one (in)famous poster with an old JVC who's name I shall not mention as they're a bit like Beetlejuice... lol... but he'll probably be along as soon as a screenshot emerges :-p).
FWIW, "processing" as in sharpening is something that definitely is a matter of taste. However, there are some sorts of processing which you cannot technically avoid. For example, consumer video is encoded as YCbCr 4:2:0, but displays are RGB 4:4:4, so someone somewhere has to upscale chroma from 4:2:0 to 4:4:4 and then convert that to RGB. Furthermore, if you play a 1080p Blu-Ray on a 4K display, again someone has to upscale the whole video from 1080p to 4K. So the question is not *if* you do that kind of processing, the question is only how high the quality of the algorithm you're using is. There are good and bad ones.

Quote:
Originally Posted by bobof View Post
An example of testing that would be really interesting (in the vein of the video I linked to); take some >4K images, process both direct to 4K and also as if via a 2K DI; pack it for transmission (ie encode at 4:2:0 with UHD compression). Then process via choice of image processing algorithms (eg MPC/Reality Creation/MadVR) and then compare the processed and unprocessed results with the original pre-compression masters sent to the display at full resolution 4K 4:4:4 and see how they fair. This would be really interesting and would further the understanding of the capabilities and limitations of such processing no end.
This is actually very near to how madVR's "NGU Sharp" algorithm was designed: It tries to undo/revert a 4K -> 2K downscale in the best possible way. There's zero artificial sharpening going on. The algo is just looking at the 2K downscale and then tries to take a best guess at how the original 4K image might have looked like, by throwing lots and lots of GLOPS on the task. The core part of the whole algo is a neural network (AI) which was carefully trained to "guess" the original 4K image, given only the 2K image. The training of such a neural network works by feeding it with both the downscaled 2K and the original 4K image, and then the training automatically analyzes what the neural network does and how much its output differs from the original 4K image, and then applies small corrections to the neural network to get nearer to the ideal results. This training is done hundreds of thousands of times, over and over again.

Sadly, if a video wasn't actually downscaled from 4K -> 2K, but is actually a native 2K source, the algorithm doesn't produce as good results as otherwise, but it's usually still noticably better than conventional upscaling algorithms.
madshi is online now  
post #1398 of 13653 Old 08-31-2018, 05:19 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,036
Mentioned: 490 Post(s)
Tagged: 0 Thread(s)
Quoted: 6922 Post(s)
Liked: 6629
Quote:
Originally Posted by madshi View Post
This is actually very near to how madVR's "NGU Sharp" algorithm was designed: It tries to undo/revert a 4K -> 2K downscale in the best possible way. There's zero artificial sharpening going on. The algo is just looking at the 2K downscale and then tries to take a best guess at how the original 4K image might have looked like, by throwing lots and lots of GLOPS on the task. The core part of the whole algo is a neural network (AI) which was carefully trained to "guess" the original 4K image, given only the 2K image. The training of such a neural network works by feeding it with both the downscaled 2K and the original 4K image, and then the training automatically analyzes what the neural network does and how much its output differs from the original 4K image, and then applies small corrections to the neural network to get nearer to the ideal results. This training is done hundreds of thousands of times, over and over again.


I did not know that.
kingwiggi likes this.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #1399 of 13653 Old 08-31-2018, 05:19 AM
aka jfinnie
 
Join Date: Aug 2015
Location: Norwich, UK
Posts: 3,271
Mentioned: 54 Post(s)
Tagged: 0 Thread(s)
Quoted: 2642 Post(s)
Liked: 1740
Quote:
Originally Posted by Javs View Post
In the case of MadVR to see how close it got to the real 'truth' I have done exactly this...

I took a control 5.7k holiday image and shrunk it down to 2160p, then I output a specific 1080p version of the same file using the same 'master' as the source.

Indeed NGU comes very close to the original, but not quite there yet. I should note the Control UHD is still passing through MadVR, but all its doing is Chroma up-sampling, no sharpening.
Nice work. I think some of the links are borked...

It is similar but not quite the same. Your test is for upscaling prowess between algorithms. I'm particularly interested in claims that processing algorithms get a 2KDI'd piece of content delivered in 4K (ie upscaled to 4K "professionally") closer to what it should have looked like prior to the 2KDI intermediary process. (in processing for release I'm sure the 2KDI goes through a better upscaler than Kodi's nearest neighbour or whatever). The idea is you take content processed via the professional 2K -> 4K upscaler, process it again via whatever processing you like (which should "recover / add detail") and then compare that to what it would have been had it not been 2K DI'd.

You could also do this to 4K content which hadn't been 2KDI'd to see if the algorithms brought some advantage in dealing with chroma upsampling and artifacts introduced by compression.
bobof is online now  
post #1400 of 13653 Old 08-31-2018, 05:21 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,036
Mentioned: 490 Post(s)
Tagged: 0 Thread(s)
Quoted: 6922 Post(s)
Liked: 6629
Quote:
Originally Posted by bobof View Post
Nice work. I think some of the links are borked...
Try again you clicked too quickly.
bobof likes this.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #1401 of 13653 Old 08-31-2018, 05:28 AM
Member
 
scrowe's Avatar
 
Join Date: Feb 2011
Posts: 40
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 29 Post(s)
Liked: 27
Minor gripe …

AVF is speculating no motorised lens on any models, as the spec sheets list Lens Cap as accessory on all models?
scrowe is offline  
post #1402 of 13653 Old 08-31-2018, 05:30 AM
AVS Forum Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 7,757
Mentioned: 530 Post(s)
Tagged: 0 Thread(s)
Quoted: 2620 Post(s)
Liked: 3530
Quote:
Originally Posted by Javs View Post
In the case of MadVR to see how close it got to the real 'truth' I have done exactly this...

I took a control 5.7k holiday image and shrunk it down to 2160p, then I output a specific 1080p version of the same file using the same 'master' as the source.

Then what I did is compared them all using MadVR and essentially vanilla up-scaling from Kodi, which I think is nearest neighbour or something fairly default, then I also up-scaled using the MadVR NGU settings I use for 99% of my viewing and compared them all against the original master image. Doing this of course we know what the original would look like in true 4k as it were, and we can see how good/close up-scaling can get to that original.
Cool beans, thanks! There seems to be a sub-pixel shift in the image, though? That shouldn't be there - unless you used NGU Anti-Alias? And how did you downscale the image? Catmull-Rom or Nearest Neighbor?

Quote:
Originally Posted by bobof View Post
It is similar but not quite the same. Your test is for upscaling prowess between algorithms. I'm particularly interested in claims that processing algorithms get a 2KDI'd piece of content delivered in 4K (ie upscaled to 4K "professionally") closer to what it should have looked like prior to the 2KDI intermediary process.
Such a kind of "upscaling correction" is not something madVR supports atm. Technically, it would not differ a lot from how NGU Sharp upscales atm, provided the "professional" upscaling algorithm which was used to upscale the 2KDI to 4KDI is a known standard algorithm. If the "professional" upscaling algorithm is done with an unexpected (and very different looking) algorithm, such kind of "upscaling correction" cannot work properly.
Manni01 and bobof like this.
madshi is online now  
post #1403 of 13653 Old 08-31-2018, 05:30 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,117
Mentioned: 329 Post(s)
Tagged: 0 Thread(s)
Quoted: 5443 Post(s)
Liked: 5649
Quote:
Originally Posted by scrowe View Post
Minor gripe …

AVF is speculating no motorised lens on any models, as the spec sheets list Lens Cap as accessory on all models?
Lens cap has nothing to do with motorized lens, it’s just the lens cover that isn’t motorized.
Mike_WI and scrowe like this.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders
Manni01 is online now  
post #1404 of 13653 Old 08-31-2018, 05:32 AM
aka jfinnie
 
Join Date: Aug 2015
Location: Norwich, UK
Posts: 3,271
Mentioned: 54 Post(s)
Tagged: 0 Thread(s)
Quoted: 2642 Post(s)
Liked: 1740
Quote:
Originally Posted by madshi View Post
FWIW, "processing" as in sharpening is something that definitely is a matter of taste. However, there are some sorts of processing which you cannot technically avoid. For example, consumer video is encoded as YCbCr 4:2:0, but displays are RGB 4:4:4, so someone somewhere has to upscale chroma from 4:2:0 to 4:4:4 and then convert that to RGB. Furthermore, if you play a 1080p Blu-Ray on a 4K display, again someone has to upscale the whole video from 1080p to 4K. So the question is not *if* you do that kind of processing, the question is only how high the quality of the algorithm you're using is. There are good and bad ones.

This is actually very near to how madVR's "NGU Sharp" algorithm was designed: It tries to undo/revert a 4K -> 2K downscale in the best possible way. There's zero artificial sharpening going on. The algo is just looking at the 2K downscale and then tries to take a best guess at how the original 4K image might have looked like, by throwing lots and lots of GLOPS on the task. The core part of the whole algo is a neural network (AI) which was carefully trained to "guess" the original 4K image, given only the 2K image. The training of such a neural network works by feeding it with both the downscaled 2K and the original 4K image, and then the training automatically analyzes what the neural network does and how much its output differs from the original 4K image, and then applies small corrections to the neural network to get nearer to the ideal results. This training is done hundreds of thousands of times, over and over again.

Sadly, if a video wasn't actually downscaled from 4K -> 2K, but is actually a native 2K source, the algorithm doesn't produce as good results as otherwise, but it's usually still noticably better than conventional upscaling algorithms.
Thanks @madshi ; I get all that. Chroma upscaling has to happen for most video, and I think if you are upscaling 1080p -> 4k yourself all bets are off; you do whatever to attempt to get the best result. Sounds like your approach has paid dividends.
However, this is not the specific case being argued as I tried to explain in subsequent message to @Javs .

The specific case being argued is the proof of benefit of algorithms such as Reality Creation and MPC when the input source is already 4K. ie - there has been a 2K process in the intermediate stages of the content creation (eg 2KDI movie), this has professionally been then mastered to 4K. The claim that I think merits investigation is that such algorithms are capable of operating on this already upscaled content and getting it to closer to what it should have been had it not been for this meddling processing having been done in the middle.
bobof is online now  
post #1405 of 13653 Old 08-31-2018, 05:33 AM
AVS Forum Special Member
 
markmon1's Avatar
 
Join Date: Dec 2006
Posts: 6,223
Mentioned: 106 Post(s)
Tagged: 0 Thread(s)
Quoted: 5227 Post(s)
Liked: 3400
Quote:
Originally Posted by Javs View Post
Why dont you just buy the Sony UHD player, its built into that is it not?
My use would be for games. I don't want another player.

Video: JVC RS4500 135" screen in pure black room no light, htpc nvidia 1080ti.
Audio: Anthem mrx720 running 7.1.4, McIntosh MC-303, MC-152, B&W 802d3 LR, B&W HTM1D3 center, B&W 805d3 surround, B&W 702S2 rear, B&W 706s2 x 4 shelf mounted for atmos, 2 sub arrays both infinite baffle: 4x15 fi audio running on behringer ep4000 + 4x12 fi audio running on 2nd ep4000.
markmon1 is offline  
post #1406 of 13653 Old 08-31-2018, 05:35 AM
Member
 
scrowe's Avatar
 
Join Date: Feb 2011
Posts: 40
Mentioned: 4 Post(s)
Tagged: 0 Thread(s)
Quoted: 29 Post(s)
Liked: 27
Quote:
Originally Posted by Manni01 View Post
Lens cap has nothing to do with motorized lens, it’s just the lens cover that isn’t motorized.
Thanks for correction, I did of course mean motorised lens cover.
Manni01 likes this.
scrowe is offline  
post #1407 of 13653 Old 08-31-2018, 05:36 AM
AVS Forum Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 9,117
Mentioned: 329 Post(s)
Tagged: 0 Thread(s)
Quoted: 5443 Post(s)
Liked: 5649
Interesting write-up here (while we wait for more news from Ran/Nigel):

http://lbtechreviews.com/news/home-t...ater-projector

I think the extended lens shift range (+/- 100% V +/- 43% H) might be for the N9 with the better (100mm) lens.

Maybe the lower models (with the 65mm lens) get the +/- 80 V +/- 34 H from the specs sheet we have?

Also seems to confirm that as expected the new auto tone mapping is based on HDR10 static metadata and not real-time dynamic processing based of the peak brightness of each frame/shot.

So it’s automatic per title (think one custom curve optimized according to static HDR10 metadata when playback starts, this single static curve being selected amongst many curves rather than a max of theee with the Vertex) rather than proper dynamic frame-based processing such as in MadVR.

Batch Utility V4.02 May 16 2019 to automate measurements files for madVR with support for BD Folders

Last edited by Manni01; 08-31-2018 at 05:45 AM.
Manni01 is online now  
post #1408 of 13653 Old 08-31-2018, 05:40 AM
aka jfinnie
 
Join Date: Aug 2015
Location: Norwich, UK
Posts: 3,271
Mentioned: 54 Post(s)
Tagged: 0 Thread(s)
Quoted: 2642 Post(s)
Liked: 1740
Quote:
Originally Posted by madshi View Post
Such a kind of "upscaling correction" is not something madVR supports atm. Technically, it would not differ a lot from how NGU Sharp upscales atm, provided the "professional" upscaling algorithm which was used to upscale the 2KDI to 4KDI is a known standard algorithm. If the "professional" upscaling algorithm is done with an unexpected (and very different looking) algorithm, such kind of "upscaling correction" cannot work properly.
Our messages are slightly out of sync, sorry.
Of course the point you make is exactly the reason this can't really be doing what is being claimed. There isn't one algorithm or set of settings which would be used in the processes.

A different result, that might appear "sharper", and might be preferable when enabled via remote vs disabled? Sure, why not, may happen by twiddling the buttons. But to me it seems pretty clear this won't be more accurate unless you could very closely constrain or in great detail know the input conditions.
bobof is online now  
post #1409 of 13653 Old 08-31-2018, 05:43 AM
AVS Forum Special Member
 
Javs's Avatar
 
Join Date: Dec 2012
Location: Sydney
Posts: 8,036
Mentioned: 490 Post(s)
Tagged: 0 Thread(s)
Quoted: 6922 Post(s)
Liked: 6629
Quote:
Originally Posted by madshi View Post
Cool beans, thanks! There seems to be a sub-pixel shift in the image, though? That shouldn't be there - unless you used NGU Anti-Alias? And how did you downscale the image? Catmull-Rom or Nearest Neighbor?
No idea mate, was so long ago now The downscale was in photoshop though, or maybe lightroom.. really cant remember..

I would love to give you the image if you wouldn't mind doing the comparison the best you can on your end?

https://www.dropbox.com/s/s2tvsfmpxn..._6009.tif?dl=1
Manni01 likes this.

JVC X9500 (RS620) | 120" 16:9 | Marantz AV7702 MkII | Emotiva XPA-7 | DIY Modular Towers | DIY TPL-150 Surrounds | DIY Atmos | DIY 18" Subs
-
MadVR Settings | UHD Waveform Analysis | Arve Tool Instructions + V3 Javs Curves
Javs is offline  
post #1410 of 13653 Old 08-31-2018, 05:47 AM
AVS Forum Addicted Member
 
ack_bk's Avatar
 
Join Date: Jan 2007
Location: New England
Posts: 11,363
Mentioned: 26 Post(s)
Tagged: 0 Thread(s)
Quoted: 1343 Post(s)
Liked: 1503
Just reading some of the first impressions of the new JVC lineup now online and they seem very favorable. Especially the automatic tone mapping for HDR.
I am waiting this out a few more years since my 770R is not even a year old, but that could be perfect timing for them to work out bugs, increase lumens and contrast even more, and see the prices go down

Very interested to see how Sony counters and what Epson does next.
ack_bk is online now  
Sponsored Links
Advertisement
 
Reply Digital Hi-End Projectors - $3,000+ USD MSRP

Tags
eshift , Jvc , native 4k , projector , uhd

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off