Jump to content

mstark

Members
  • Posts

    552
  • Joined

  • Last visited

  • Days Won

    1

Everything posted by mstark

  1. Unless I'm missing something, the 2nd image shouldn't look like that. An integer factor scaling doesn't cause blurriness, it'll look just the same as the last picture. You are missing something If you got access to a high DPI device, do have a look at the demo posted above . Editing the demo page with a third example that will further illustrate the issue!
  2. The great thing about image sampling is that you can downsample images with virtually no loss of quality. This means that, as long as they prepare the original set of assets to be high-DPI friendly, they can easily be downsampled to a standard resolution and still look as good, or, more likely, better. Yes, it would increase rendering time, and marginally increase post-production, but I doubt rendering times are their biggest worry, since it doesn't actually use up any manpower. That way, they could ship the game with standard resolution imagery, not to take up too much disk space, and offer a retina (2x) download pack. I doubt it would affect production time much. Another scenario is that they render the game only for standard resolution monitors at first, but include a DPI toggle that's hidden, so that they could release high dpi imagery at a later point. Maybe as a patch. I actually have a hard time thinking they wouldn't, at some point, release high dpi versions of the renders, because the shift is happening, whether people want to acknowledge it or now . DPI scaling demo available here now: http://martinstark.net/dpi/ (YOU WILL ONLY SEE THE EFFECT ON HIGH DPI SCREENS). Look at it without zooming the page. On a standard DPI monitor you won't actually see any difference between the images, you need to view it at a device with 2.0x DPI scaling (currently, that means a few Android flagship phones, though most of them are 1.5x, and iPhone 4, iPad 3 and Retina MBP). The effect will be most apparent on a retina iPad 3 & MBP because of screen real-estate. Landscape mode for side-by-side comparison recommended.
  3. Ah right, monitor manufacturers are happy that consumers don't need to change their monitor and buy a new one, they prefer to ignore the replacement bussiness that might fill their pockets. For the average consumer not all is good though. If they fall for the "hype" they need a 4 times stronger graphics card. And the monitor and the card will drain more power naturally. It's always cheaper for manufacturers to cling to what they are doing for as long as is possible. They've already got the factories that can produce standard resolution monitors, and can churn these out virtually at parts cost. To produce high DPI monitors they'd need to build new factories, that can do so... a major investment, that many are reluctant to take. Samsung was one of the first companies to invest in factories for producing high DPI screens, thanks to this, they are now a market leader, together with LG/AU (Apple's screen manufacturer). For the average consumer it is always good when the tech world moves forward. Yes, it'd introduce new, more expensive hardware on the marked, but this would greatly drive down the prices of today's 1080p monitors, in addition to bringing higher-cost, high DPI monitors to the market. Graphics cards designed to drive today's 1080p monitors would become cheaper, and the tech world would move forward.
  4. Downsampling is fine, if the imagery for Project Eternity is prepared at retina level and then downsampled for standard dpi devices, this would be the same thing as I tried explaining in my first post. What I'm trying to get at is that we'll need [at least] 2 sets of graphical assets, one generated for high dpi monitors, and one generated for standard dpi monitors. A single set of assets is OK it you are okay with them being either upscaled, making them appear blurred, or not not scaled at all, which would make them appear tiny on a retina screen. But why should the PE team accept that, when all it takes is to generate a second set of assets, and include a DPI toggle option? What I desperately want to avoid is upsampling, or game areas & characters appearing so tiny (as is happening with IE games now, unless you play them at lower resolutions on your shiny high resolution monitor, which looks terrible). This is why there's a need for more than one set of graphical assets. If they render everything at retina resolution, and then downsample all of it for regular monitors, it'd have the same effect. As long as there are two sets of assets, and a way of toggling between them. I'm preparing a website that'll demo the effect of using different assets for a retina device. It might help you visualize the the issue if you imagine playing Baldur's Gate on an iPad 3. At native resolution, because BG only has a single set of graphical assets, the characters and UI would be so small that you could barely make them out. You could, of course, play it at 1024*768, but that would make it look very blurry compared to anything else you view on the device. If you had access to a second set of assets generated at twice the resolution of the original assets, you could still play the game at the iPad 3's native resolution, without it looking blurry, or tiny. A great thing about what I'm suggesting is that it wouldn't take up much extra time at all. It'd mean some additional rendering time, but the great thing about rendering is that it can be done over night, and without anyone attending to it. They's simply need to tell their machine to render it at a sufficiently high resolution, then leave it for the time it takes it to do so.
  5. Yeah, the water effects weren't the best, but the texture on the renders are higher quality than those of BG2/IWD2, in my opinion. It looks like PE will be have an option for using real time rendered water, which would look far better than sprites based water. Adding to that, there were sliding doors, moving lamps, area lighting effects, the FlaK cannons were animated 3D obejcts... etc. Unfortunately, they didn't use any anti-aliasing for the 3D props, making their edges look far sharper than the rest of the map.
  6. Do you have an iPad 3 right there with you? I do, along with a couple of different high DPI devices, and I will gladly prepare a demo for you that you can view on it to see that it will, indeed, look terrible. Or tiny. Terrible or tiny, unless a 2x resolution version is prepared of the same area/image. I, too, didn't care about this at all until I saw it in action... it truly makes a difference. If you look at screen 2 & 3 in my example, this demos the two options than an iPad 3 uses for image scaling. It will either upscale the image to not become so tiny that you can't make out details in it, at the same time making it blurry... or it will show more of the image on the screen, making details tiny. The 4th example demos what happens if the iPad 3 has access to a 2x version of said image, it will load that image instead, resulting in an unbelievably sharp image, while retaining the scale the image creator intended for it to be seen at. (This article, though quite technical, and completely aimed at the web, will give you a primer on why http://coding.smashi...rds-retina-web/)
  7. There are far too many helpful people on this forum
  8. Sorry for double post, but here's a quick graphic I made explaining the effect, that is understandable when viewed on standard DPI monitors: To achieve this, all bitmap graphics would have to be produced at both their "regular", and 2x resolution.
  9. I'm just thankful that the wave of smartphones finally brought High DPI screens to the attention of a broader market, that way monitor manufacturers can't continue to ignore this issue as they have been doing for ages (to great profitability for them). Now they have to up their game in more ways than meaningless "1:2,000,000 dynamic contrast ratio (an outright lie)" or "1ms response time (input lag nullifies this achievement)". Imagine when screens reach true print quality... 16000dpi screens anyone? It'll take some time before graphics cards & cables can handle that, though. If the PE team takes dpi scaling into consideration, their game will look amazing on this new generation of screens that are now entering the market, leaving current-gen 3D games far behind, simply because there aren't many affordable graphics cards on the market that can render good looking 3D in real time at those resolutions. The very reason why the original IE games looked so much better than the 3D games of its time. The PE team has the chance to achieve the same thing, for a new generation of hardware, because we're experiencing a paradigm shift now, too, that they can take into consideration.
  10. I really hope they'll be featured on all mainstream blogs & news sources on their final day, and that they have some really nice new screenshots available for those features... that's what'll attract a whole lot of new people
  11. You are right, and I'm an idiot. Thanks for correcting me . I much preferred Commandos 1/2 to number 3, now that I looked them up. Too long since I played them!
  12. This is exactly the point I'm arguing against. We shouldn't look at today's market if we want the game to be future proofed. Yes, "retina" or "high dpi" screens are rare, because they have just entered the market. 1080p HDTVs were rare when they were first introduced too, but today no one would ever consider watching a VHS over a BluRay. The IE games have aged well because their areas were, for the time, rendered at very high resolutions, and DPI hasn't changed much since then. But it is changing now. Rapidly. At the end of this month, Windows 8 and an army of reasonably priced 1080p 11" screens are being unleashed on the market. By 2014, it'll have started to be widely adapted. It's hard to realize how terrible old 2D games can look at these screens without having actually seen it. When you have, you will spend two hours, like me, writing a post like this, to try to convince people of the importance of this . Which is exactly why I'm suggesting it to be downloadable extras. Release it at a resolution that's good for the majority, but also make sure that the game has the capacity of handling DPI scaling. Like I said, for 3D games it's not such an issue, because you can simply upgrade the textures. For a 2D game, you need to consider it from the start. As much as I would love this, lossy audio is much less of a problem . And on that note, the songs they've released so far were lossless wave files!
  13. This is not true, a 400x400 image scaled up to 800x800, as it would be on an iPhone 4 (2.0x dpi scaling), or 600x600 as it would be on a Galaxy S2 (1.5x DPI scaling), still looks terribly blurry if you compare it side by side to an image that has been rendered at that resolution and not up-scaled. The bad thing is that I can't show you this in action, since we're all probably reading this on "normal" DPI monitors right now. I would prepare a quick example on a website for you if you have a high DPI device ready to view it on (most smartphones, or a retina iPad/MBP, Nexus 7 tablet). You will see the difference Went looking for an article on DPI scaling that would explain better than I can why this is the case, but I can't find it... will post it if I can!
  14. You're probably right, I almost only read this board, so I never thought about that! Could someone please move it for me?
  15. This issue is related to the "resolution" of PE, and is a suggestion for not rendering it to any particular modern day standard, but in a more future-proof way. It's a long and fairly complicated post, but I'm hoping that the devs of PE are considering DPI scaling, or their game risks looking outdated before it's even been released. As a web designer, the importance of DPI scaling is a very real issue to me today, and will become even more important in the next few years. As the web is rendered in 2D, these issues are immediately translatable to Project Eternity. For 3D games it's not much of a problem, because they scale very well with higher resolutions, but since 2D games are, as a rule, only rendered at one resolution, this will become a very real problem for the PE team soon. Let me give you some examples. For the sake of simplicity, let's pretend that all areas of PE are being rendered at 400px by 400px resolution. On common desktop screens today (with a DPI of 100 - 120), this will appear as a 2-3" (5-7.5cm) square, and will be made readable with that in mind. In the past 10 years, the DPI of screens have barely changed, most screens during IEs days of glory being 96 DPI. For such a long period of time, this is a relatively small increase, which is why these games are at all playable on modern day monitors. There's never been any reason to consider DPI scaling, because High DPI screens have never existed before. However, today we're already seeing an explosion in the popularity of High DPI screens, which started with phones. Around 90% of smartphones released today have a DPI somewhere between 260-420. They look incredibly sharp compared to the chunky pixels of most monitors today (did you look at the PE screenshot on a Galaxy S3? It's tiny, but Oh. My. God. so pretty). And this is the direction that computers are moving in, too. We're already seeing 1080p 11" laptops at ~220dpi, or the 15" Retina MacBook with ~220dpi. High DPI desktop screens are just about to enter the market, too. In a matter of years, they will be the new standard. Think of it as television jumping from 640*480 to 1920*1080 within the span of a few years, but even more extreme. Think of how terrible old TV programs look when scaled up to 1080p. So why's this a problem for PE? With screens already entering the market that are 2-4 times higher DPI than what's remained a standard for near 15 years, this will need to be considered during the development process. The old IE games are completely unplayable on such devices, unless they're played at non-native resolutions, but you wouldn't want to do that. Upscaling is terrible. The 400px x 400px area render I mentioned earlier that would usually appear as 2-3" appears 0.5"-1" on such a screen, making it unreadable. The web has been wrestling with this problem for years, how to make websites readable on devices that would render text and images too small to be seen? What's the solution that the web industry has gone for? Enter DPI scaling. The iPhone 4 has a DPI scaling factor of 2.0, meaning it will render everything 2x as big. For vector graphics, such as text, it's easy enough, since it's infinitely scaleable without losing quality. But bitmap images, as the 2D renders of PE will be, are stretched. For the image quality, this has the same effect as playing BG in 800x600, but stretched to fit a 1080p screen. It. Looks. Terrible. So what we do, as a web designers, is to serve the same 400px image, but rendered in 2x the resolution (800px) scaled to half the size. This makes the image appear as large and readable on the phone, as it would on a 24" screen, with the benefit of it being twice as sharp without being unreadably small. What does this mean for PE? It means that PE, as a 2D game, needs to allow the user to select DPI scaling. Normal scaling on devices today, for the web, is 1x, 1.3x, 1.5x and 2.0x. For vector graphics and 3D props this is easy enogh. But it also means that all bitmap images: area renders, character portraits, UI graphics, etc. needs to be rendered at not only "normal" resolution, but 1.3x, 1.5x, 2.0x in order to remain playable on High DPI screens that are already on the market. Adding DPI scaling means that on super high resolution (high DPI) monitors you would still see the same amount of area on the screen, but it'll appear far, far sharper, rather than what's happening with current IE games, where areas appear smaller and smaller, until they're so tiny you can't play it. The issue stems from old IE games only having their area files rendered at one resolution. Ever tried playing BG2 at the native resolution of a Retina MacBook pro? For your own sanity: don't. Ever looked at the PE screenshot on an 11" 1080p screen? It looks gorgeous, but it's also tiny. Really. Tiny. If the same area was also rendered at 2x the resolution, then scaled to a DPI factor of 2.0, the game would remain playable, and still look incredibly sharp and beautiful. As a bonus, allowing for DPI scaling would indirectly allow users to zoom, without the backgrounds becoming 'pixelated'/upscaled. The drawback to this is that it would make the game's image files take up an insane amount of space, so my suggestion would be to add DPI scaling packs as optional downloadable content. I should probably mention that the team behind BG:EE discussed re-rendering all the area files to better fit modern day resolutions, but apparently much of the original work has been lost, making this an impossibility . Hope this makes sense to you all, and that the PE dev team will consider implementing solutions that allows the game to remain playable on high DPI monitors! Since the PE team has mentioned that they are intending to allow limited zooming, I'm guessing they're already thinking along the lines of rendering their areas at a higher resolution than "needed" for "normal" monitors today, this post is suggesting that you take it a step further, and allow the user to select DPI scaling to best fit their monitor. I suppose this could be achieved indirectly by letting the user simply zoom to a comfortable level, but that leaves the issue with the UI and character portraits appearing tiny. I'd like to end this post with saying how incredibly excited I am about this game, and that I love everything you guys have revealed about it so far Regards, M (for the sake of sanity, DPI = PPI in this post)
  16. It only looks like "Sim Ant" because you're watching it in a tiny youtube window. I'm playing BG2 on both a 27" (2560*1440) and a 24" (1920*1200) monitor, and it really isn't appearing much tinier than it did years ago on my 15" 1024*768 monitor. You're just seeing a lot more on the screen. When I play it on my 1080p 13" laptop screen it's a bit hellish, though. Allowing for DPI scaling is far more important than what resolutions they render anything in. If you render everything to a particular resolution that's popular today, it will date as quickly as the IE games. If you render it with 2x, 4x, etc. DPI scaling in mind, it'll look amazing forever. I doubt their hardware could feasibly render all their areas at 4x in a reasonable time, though. Anyway, that brings me to the best looking top down 2D game I've ever played: Commandos 3 (2003) ... and this game was rendered from 4 different angles. It also had a lot of animated 3D bits, and water was animated, driveable vehicles, mood-setting weather effects... (massive picture, do look at full version!)
  17. We *will* make it to 3.5! We have to! Wheee!
  18. ^that, plus: http://en.wikipedia....phic_projection
  19. A lot more awesome! But that being said, the Underdark was a pretty damn fully realized area to begin with! (IMO) The mindflayer dungeon of the Underdark is one of my strongest memories of the game... it took me forever to figure out a way to beat it!
  20. You might want to try out some of the Pirates of the Caribbean games? I think there are a few "medieval naval sims" out there newer than that game, though. Seriously though, it'd be great, but what you're describing sounds like an entirely different game of its own, how would it integrate with the original idea of the game? If you think of it as a top-down, tactical combat cRPG, in the vein of the original IE games? If there's a solution good enough, I'm sure Obsidian will consider integrating it. BG2 had some naval action, but it was part of the storyline, rather than a game mechanic of its own. You were taken along, as opposed to controlling it. I liked this, if only because it brought you to a very different environment. I don't think this should be taken as literally as a "naval component" being vital to the game, though!
  21. Limited zoom has already been mentioned, it'll all depend on at what resolution they end up exporting the final game graphics at. The higher the resolution, the longer the generation of the environments will take... if they have enough time to generate it at extreme resolutions, I'm positive we'll get the ability to zoom in (assuming PEs default camera view is fully "zoomed out"). You'll be happy to know there'll be no such thing as camera angles. Possibly limited zooming (for a better strategical overview of a battlefield), but that's it. Actually, from my experience across different game genres, zooming OUT is always favorable for a tactical view of the field. Zooming in is only useful for either precise aiming or LOS. And ego shots. (Of course, I'm assuming that the default PE camera view is fully zoomed out with the option to zoom in a bit. That's not confirmed, though.) This, is what I wanted my post to say. Feel like I left OUT the word OUT.
  22. The world is full of griefers, allow them to grief themselves to oblivion by killing anything and everything out there! On a serious note though, agreeing with the OP. I'd love if Obsidian kept the game realistic, in the sense that no one, short of godlike beings, is unkillable.
  23. Being a martial artist myself (Hung Ga, Yee lineage, 5th generation from Wong Fei Hung) I can't stress how important the points made by hideo kuze are... please do allow your concept artists, as well as class developers, to study the actual real-world background of the characters you create! As much as I disliked the game, the Diablo 3 monk, in its own unrealistic ways, got a lot of this right.
  24. Hello, Obsidian Order! http://www.kickstart...ofile/612573107 Having pledged $80 so far, I'd love to be Oblivious of the Obsidian Order
×
×
  • Create New...