Jump to content

renault

Members
  • Posts

    3,696
  • Joined

  • Last visited

  • Days Won

    39

renault last won the day on August 30

renault had the most liked content!

3 Followers

About renault

  • Birthday January 1

Profile Information

  • Gender:
    Not Telling

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

renault's Achievements

Life Member

Life Member (5/6)

7.3k

Reputation

  1. It's a very fine set John, very fine indeed. But it has its downside. If I woke up and saw that the morning was like that I would go back to bed One of the joys of retirement One of the challenges , yet to be overcome, is #1 is an early riser and would have already had the coffee on, to be followed by "where are earth are you?" Sigh Cheers Pete
  2. Hi Wayne No it was just moi, playing with a slider. BTW if you set ambient occlusion to ULTRA it will really improve the perception of depth to the snow cover (example in the last screenshot with the grasses). Normally I either keep AO off or on low, because IMHO the implementation by Asobo is the worst I have seen in any modern gaming engine. I really dislike the current technology that MS2020 is using (SSAO - screen space ambient occlusion). I am really keeping my fingers crossed that they will implement soft shadows algorithms when they move up to DX12-1 This video does a good job of showing what advances in gaming engine technologies are capable of. You may have seen automobile ads made with Unreal Engine and the CGI of a lot of Hollywood movies use their technology. At 3.05 in the following video there is a good discussion on soft shadow illumination Cheers & thank you Pete Thanks so much Iain Cheers Pete Ha, thanks Graham !! You've just invented a "cottage industry" using MS2020. I love the idea Cheers Pete Thanks Adam Cheers! Thank you John. I am really pleased you enjoyed them All the best Pete Thanks John I was thinking of you when I made them and hoping that you might catch it. I'm very pleased you appreciated it! All the best Sir Cheers Pete
  3. There are so many absolutely excellent screenshots already of this wonderful scenery that I wasn't sure where to start. So here are a few with my "take" on it. Marcus and Orbx have a winner here Cheers Pete
  4. Thanks Adam! Cheers Pete Thanks Graeme See you Pete Thanks Wayne See you Pete Thanks Iain See you Pete
  5. Thanks JM Cheers Pete Thanks John. Appreciate it Cheers Pete Thanks John I sure hope so, but in my weaker moments when I was thinking I think I see light at the end of the tunnel, but maybe its another "Update" and the light is actually a train Pete Hi Ken Thank you Yes it is, coupled with a pretty good monitor designed for accurate color preproduction, that is color calibrated monthly, along with a fully color managed workflow so that what you see in a web browser is as close as I can get to what I see in its original form. As I mentioned to you , I am fascinated with how modern gaming engines deal with color reproduction and color fidelity and it is an area that I find really satisfying to muddle around in. It becomes quite immersive and relaxing after a while. I was looking at some great photos taken 2 yrs ago at KAVX that I discovered the other day at conservation societies website. I have to say that you and Tim absolutely captured real life and made it available for all of us to enjoy ! It is to put it mildly , rather uncanny how well the scenery reproduces what is there. Cheers and thanks! Pete Thanks Paul I will have to venture out and take a lot more screenshots of the Orbx world now. Another update is forthcoming , and who knows what that might bring - ha, ha Cheers All the best Pete
  6. Hey Landon Now that we are two , we can start our own "club" I think you expressed it really well. Orbx by and large does an excellent job creating quality, high resolution scenery that works really well in MS2020. It is such a downer when the pg isn't up to par. For me the scenery streaming and the overall quality of the pg result is the Achilles heel of MS2020. It is a great concept - to have the world at your finger tips without having a massive investment in disk storage and scenery, but it really requires a streaming service that is world class in all respects. To me it is somewhat like what would happen if you signed up for a 4K UHD movie streaming service and then find out that 95% of the time they can only give you 1080p. Unfortunately, there is no substitute for pixel density . All the upscaling technology in the world won't bring the immersion back. I don't know if you have played around with Nvidia DLSS but from my experience with it I would say it does work and can give you a decent enough image at 4K while your system is rendering at 1080p. But there is a critical component to all these technologies that is not immediately obvious. It is this . It is actually really simple to upscale a low resolution image to create an artificial higher resolution one. But what is darned difficult, (I would argue almost impossible) is to actually transfer the color information present in native 4K . When you look at all these techniques with a critical eye, you quickly realize that this is what is missing. 1080p has 1/4 the number of pixels that 4K has and it is the range of the color gamut that exists at native 4K resolution because of the number of individual pixels that gets lost. Coupled with this is that 1080p is usually at 8 bit color , whereas 4K is at 10 bit or in some cases 12 bit color. So not only do you have a great many more pixels but the range of color information per pixel can be 1034 or 4096 shades of colour per sub pixel at 4K compared to a color range at 1080p 8 bit which is 255 shades of colour for each red, green and blue subpixel. Can we see such a wide color gamut. Truthfully no, but what all of us can see is the fidelity and richness of the hue variation at the higher resolution. When you view it at 1080p this is where "color banding" comes from which you don't see at higher pixel density and bit depth. Its understandable too , as how would an algorithm (even an AI based one) have any idea how to determine the unique hue combination that should exist at a much higher resolution. So by and large what you end up with is 4K with a 1080p color gamut. I see this even if I take a 1440p png and convert it to a 1440p jpg for screenshot purposes. By and large when you do this, you have about 1/2 the number of individual hues and this is without changing resolution. When you look closely you can see that the variation and richness/fidelity of the color gamut just isn't there. So if you think some of my screenshots have decent color fidelity, you should see them when rendered at full native resolution. In many cases, it is a night and day difference. And for me, Orbx is the 4K color gamut and pg is the 1080p cousin. It just doesn't work well at times. You especially see this in the "greens" in streamed pg scenery. Our eyes are actually more sensitive to green than we are to blue/red. So when you see a vast forest in MS2020 , because the pg trees canopy are low resolution, they actually have few pixels to contain different green hues. As a result you end up with a forest cover that doesn't have the variation and richness in color that the real world does - instead you get a monochromatic green everywhere. This is what I mean when I say that it is the pixel density and the ability to contain the color gamut that matters. But to be fair to Asobo, they are rendering millions of the things, so it is a real challenge to provide something which l looks "reasonable" without bringing a supercomputer to its knees. And Orbx puts in really good quality tree assets, so when they are besides pg assets , especially as you say that get all "melted", I find it really hard to stay positive. But to keep all this in perspective, what Asobo has accomplished this past year is nothing less than a miracle imho. For me, and I've said this before, I have MS2020 not for what it is today (as painful as it is at times) but for what the technology can become in the future. And I truly believe that they will get there. I sincerely hope that this note doesn't come across as a negative bashing exercise. it is the furthest intent from my mind and I hope that this little note, give some insight into my expectations of what I hope that MS2020 can become. Cheers Pete When is our first club meeting by the way?
  7. Just a test shot after my little escapade with the "silly season" issues I recently had with MS2020. Cheers Pete
  8. Thanks Nick. I see that and late this afternoon, my MS2020 did a small update (about 10 MB) and my behaviour appears to be as yours now. I'm thinking that it must have detected a missing file(s) possibly. The odd thing in all of this is once I had run the sim with everything set to 0 is Usercfg.opt , the two original very poorly rendered bushes that caused me to report the issue originally, have consistently been properly rendered and have remained so. The other thing I found in researching possible issues is a description of how pg objects such as trees are created. The photogrammetric trees are 2D objects and through shadowing and parallax , they will view as 3D objects from above. The "from above" comment is important in that the design takes into account that the resolution of the tree objects is much higher when viewed from above than when viewed sideways ( a comment was that most users will only see them when flying over them, so making objects with differing resolution saves on processing time). This has always been this way since the sim was first released as to model trees conventionally would bring it to its knees. The actual control for tree placement is apparently tied to Bing maps in that the Ai algorithm detects large green areas in the map photoscenery and using that along with other rule based criteria determines that for example it is a "treed area" and then populates it with objects accordingly. From that I would speculate (I'm not a scenery designer) that if you were using customized vegetation objects with exclusion polygons to mask out the underlying data, there is a possibility that a Bing map update would allow underlying pg object data to "pop through" because the exclusion mask no longer covers the proper area. The other aspect of pg trees is that the placement and density is completely controlled by the algorithm and the settings in the graphics menu for trees do not appear to have any affect . It does seem to me , as you note that the tree models are not particularly sharp. In fact it seemed that with SU5 this became more noticeable. It did (does) make me wonder that in preparation for XBox if this was an area where they were able to save on both memory requirements and processing. I.e. the side resolution was reduced even further , while keeping the resolution in the other axis the same. I certainly don't see this as bad programming, as for 99% of the user base who are flying over scenery and viewing the tree tops from a few thousand feet they would not notice any change. In many respects there is nothing magical about code optimization and this would fall into the category of an efficient use of resources without visually downgrading the image rendering for the majority of users. However for folks like myself who enjoy not only low and slow flying, but also investigating the wonderful scenery that the Orbx designers create with their custom high resolution objects, it does become quite noticeable and to be fair for myself is a real immersion breaker. But I recognize that I am in a very small minority (maybe a group of one - ha, ha!) and I accept it for what it is... Doesn't mean I like it, but one has to look at the sim overall from a balanced perspective. Thank you for taking the time to respond and especially with some additional screenshots. It is very much appreciated. Cheers Pete PS - I really like the aircraft
  9. @Nick Cooper, @Sniper31, @Doug Sawatzky, @Ken Hall, @Jon Clarke Gentleman Apologies for not responding sooner, but unfortunately all of the suggestions didn't have any effect on what I was observing. However, the good news is that I have finally found the root cause of the issue. In that respect , everyone's comments added info and understanding to help to track down what was going on. I am particularly grateful to fellow member Craig @craigeaglefire , who was helping me by comparing default scenery rendering. It was a comment which he made to me along the lines of why does your default scenery look different than my default scenery. That caused me to start thinking a bit "outside of the box" and led to the following. In terms of MS2020 bugs, this one has definitely caused me pause It is easiest to see with screenshots. I am so very appreciative of all the suggestions and help from everyone. Thank you Every comment helped me to track down what was going on. Cheers Pete I should emphasize that this is something I observed on my system. But at the present time it is a fresh install of MS2020 So I started wondering , since the bushes were such low resolution if, the reversed settings issue with Texture resolution for internal cockpit textures that I mentioned many postings ago was somehow at play again. i.e. A setting of Ultra in graphics options is actually low. SU4 This is a pre SU5 update. It looks identical to what Nick kindly posted above. SU5/WU6 Now the bush has been replaced by a poorly rendered tree. Also off the the right, the large bushes are very low resolution. In fact the tree is a default MS2020 asset as can be seen in these two screenshots SU5/WU6 So this morning, on a "whim" I changed the tree resolution to low to check out my speculation that possibly the render setting bug , which I did report to the Zendesk is possibly at play here. And this is what I got. Now the low resolution default tree , is now replaced by a properly rendered bush kand tghe bushes off to the right now render properly And changing back to Ultra (which is actually Low) should give me a very poor quality low resolution bush But it reverts back to a poorly rendered tree . The odd part is that the bushes to the right, don't revert back to low resolution. As I said, it is what I am seeing on my system, but with this changed setting i.e. Tree resolution to Low, my default scenery now renders identically to the check screenshots that Craig assisted me with. So thank you everyone again for all your assistance. In the developing saga of bugs and fixes in MS2020 , this one is definitely one of the stranger ones I have encountered. Cheers & thanks! Pete @Nick Cooper Nick would you mind changing the status on this to resolved please. Thank you
  10. Thanks Landon "same bushes look high resolution on my end, just like Nick's do" This was helpful as it suggests that I have a setting somewhere that is not what it should be. I will keep plugging away at it Cheers Pete
  11. Hi Doug Thanks for the assistance Are running the sim in Developer mode? - No , I never use it Also, have you tried Nick's below process? - Yes, it is one of my "go to" checks that I use when I have issues. It is a real oddity at the moment, since other folks (Nick, Landon @Sniper31 and a few fellow members I've PM'd) don't seem to see what I am seeing. That seems to point to it being a setting somewhere Isn't MS2020 absolutely fascinating though?. When it behaves itself , its great, and when it decides to misbehave, its even "greater" in terms of the variety of individual issues that can manifest themselves. Keeps us all on our toes Cheers & thanks Pete
  12. Hi Landon That's a great suggestion, and I hadn't considered that, because I didn't think that there were Orbx libraries for MS2020. I just went through Orbx central and was unable to find any mention of their use in MS2020. The only other sim I have is XPlane, and indeed there are Orbx libraries for it. Do you have any idea where they might be stored. I've just tried changing tree settings and that makes no difference to the resolution. It is almost like the tree resolution is disabled on my system. I've just been to the MS2020 official site, and there are some vague references to odd tree behavior in Germany since Wu6, but nothing related to what I am seeing. Those reports seem to see a bit of an inverse correlation between ground resolution and tree resolution i.e. high ground resolution gives low tree density and low resolution which flips when ground resolution is low. Hopefully one of the Catalina scenery developers may have some thoughts, but at this point I am beginning to wonder if this is a bug in WU6. But as Nick has shown it doesn't seem to be an issue on his system. I have noticed that as you get closer to this object, it "jumps" in its resolution and becomes increasingly poorer . It's almost as if one is viewing an mip mapped object, and the mip mapping somehow has been reversed for some reason. As you can see from this screenshot, the resolution of the surrounding vegetation, the large boulder etc and the ground texture itself all seem to be quite good. KAVX is one of my most liked scenery areas, but this is a bit of an immersion "bump" when you see this. If other folks see it as well then it very well could be a bug . But if it seems to be only affecting my system, then it would seem to be a corrupted file, or a setting which somehow isn't being applied correctly. That's why I asked the first question if others saw the same type of behavior. It all and all seems to be a bit of a mystery Cheers & thanks! All the best Pete
  13. Thanks NIck I agree with you. That is exactly what I expected I should be seeing. I have now tried deleting all KAVX Textures and let Orbx Central reinstall them. Unfortunately, it has made no difference. Any thoughts on what I should try next? Thanks Pete
×
×
  • Create New...