Jump to content

jpprod

Members
  • Posts

    12
  • Joined

  • Last visited

    Never

Contact Methods

  • Website URL
    http://koti.mbnet.fi/~jpprod
  • ICQ
    0

Profile Information

  • Location
    Espoo, Finland
  • Interests
    Game development, scifi

jpprod's Achievements

Ensign

Ensign (1/8)

  1. - Volumetric fogging (for during snow and rain storms) - Support for volumetric space weather areas (nebulae, space storms etc) Wow, these two are damn hard to do well even on a limited environment of an FPS game; every single implementation I've seen has had loads of visual glitches (certain types of rendering not fogged etc., transparent surfaces are particulary tricky) or hasn't just been fast enough. And you're planning on rendering them in interplanetary scale? Any chance of revealing any of your ideas on the implementational side? Originally posted by LordDavid: ... Just curious, you mentioned quite a few graphical improvements, and if I recall correctly, you were planning on implimenting Environmental Bump Mapping into the game but did not because too few cards supported it or some other reason. Now that GF3's, etc support it, will it be here too? Probably not. I don't want to get involved too much in card-specific features. If anything changes, I will update the feature list accordingly. But isn't EMBM on a DX8 card, in essence, just a special case of the pixel shaders (specifically: heightmap + dependant texture read), which are on the list? Though gotta say that the pixel shaders themselves aren't too widely supported as of yet.
  2. I think he actually has a point, while most of the things that go on in the BCM universe can be tracked by floating points, there is alot of room for more impressive graphics. In the end, complexity of game world processing is all about scale. It is not possible to handle spaceship universal positioning with simple x/y/z coordinates unless there are very restricted mission zones or the universe is in reality, really small. Even double-precision floating point math wouldn't be sufficient. For example, in order to achieve seamless universal processing, my game engine has two sets of coordinates to do the positioning, local (3x 64bit float) and universal (3x 64bit int) ones. As far as I know, Battlecruiser Millennium uses zoning and some sort of strategy to hide the zone boudaries. To make the game world run, this complex positioning has to be hidden from the AI by abstraction. And that's just one of the hard problems that need to be solved in freeform space sims. Among others is the inadequecy of traditional z-buffering in presenting objects of interplanetary scale. Aforementioned is one area which BCM handles really admirably. Free form game world that is persistent across the entire duration of the game also represents many non-trivial problems, like state restoration compatibility in case of an upgrade patch. In conclusion, I dare to claim that in a programming a perspective, freeform space sim is at least as complex as a flight sim. In flight sims 1) Aforementioned Z-issue does not exist. You can fog + clip the scenery at a certain distance. 2) Complex zoning or positioning scheme is not required, since distances are not that long. 3) With the (possible) exception of Falcon 4, game world is not 100% persistent. It's much easier to create a savegame format which is cross-version compatible. 4) Atmospheric effects are much, much easier to do. In a flight sim, standard fog looks fine and impressive clouds can be created with just a couple of particles. In a space sim, fog and cloud effects (ie. nebulas) would require an advanced particle system with highly scalable LOD + true volumetric fog. Feel free to disagree, of course [ 12-10-2001: Message edited by: Jukka T. Paajanen ]
  3. quote:Originally posted by Oban: While you guys are on this topic I have a question, Im buying a new system and Im torn , I am not really a tech guy so I was wondering should I go with a P4 1.7 /G3 Ti 200/512 Meg PC800 Ram or P4 1.7/G3 Ti 500/ 256 Meg PC 800 Ram Is PC Ram that much better? specifically for BCM.. which is better Thanks for any helpful advice. Personally I would not get any less than 512MB RAM today. Games are beginning to require 256MB just to be playable at full content detail, and with the next generation (all those Unreal 2 and Doom III engine based games) things will keep moving in that direction. RAM gives the memory-hog-of-an-OS Windows XP the breathing room it needs. I must also disagree with the performance delta between GF3/Ti500 and GF3/Ti200, outside very high resolutions, it really is not that significant. Furthermore; not to start a CPU war or anything, but since price is set, why not consider an AMD Athlon XP-based system? You could propably get a higher performing Athlon XP 1700+ rig wíth 512MB DDR at a lower price point than a P4 1.7GHz/256MB PC800. Today the reliability of the best Athlon platforms is equal to that of best P4s.
  4. Firing up Tie Fighter on my brand new AMD 486/80MHz rig with 1MB Gravis Ultrasound. Game's soundtrack really lifted this title from excellent space-sim/shooter to incredible one. GUS had magnificent orchestral instrument set and crystal-clear audio quality. Tie Fighter had the famous LucasArts iMuse MIDI music system, which seamlessly created interactive background music for the game. Most important of all, the music was not ripped directly from Star Wars movies, but cleverly composed by mixing some of the familiar themes with beautiful new ones. (Face it - while original SW songs are excellent, we've all heard them over for a couple of hundred times.) Defining moment in that game was experiencing just about any mission kicking into full action. Flying in friendly space the music was rather peaceful, combination of pizzicato strings, military snare drum, oboe. The suddenly, the tune changes to indicate an alarming turn of events: a fleet of hostile ships emerge from hyperspace. As you fly your ship to intercept them, the chords in the music are seamlessly turning more and more intense, and volume level is rising. New instruments join into playing the theme. As you're in weapons range and first laser shots are being fired, you'll find yourself engaged in full battle with music following all the way there! The iMuse truly was the next best thing to having an onboard London symphony orchestra I fear we've lost the seamless interactivity forever with the widespread use of digital streaming music. Sure the sound quality is potentially beyond that of even professional MIDI hardware, but the flexibility of real-time playback is gone. Since there is no control information present, about the best you can do with digistream music is to crossfade between songs. Provided the composer feels comfortable with the idea, I'm definitely going to try to develop something like iMuse for my upcoming game. Key is, I think, getting the control information ("beat") there and accurately syncronizing crossfades.
  5. quote:Originally posted by Epsilon 5: Making that several space are not collapsed into one. That's actually a HTML feature, and I'm pretty sure no UBB add-on offers this feature since it may have performance implications. Generally speaking there are two ways to achieve this: 1. Replacing all spaces with the HTML non-breaking space, &nbsp;. This approach is problematic, since non-breaking spaces interfere with text wrapping and may break HTML code, which is enabled in this forum. The replace procedure would take some CPU power or extra storage space off the server (reformatted message is either stored or message text reformatted per-request). 2. Using <pre>-tag. This makes no sense to do automatically, since you can use it manually, and there even already is a tag for this in the UBB code (the CODE tag). Sorry for the slightly off-topic post, but as a message board developer hobbyist (my own board is much faster than UBB by the way), I just can't help myself [edit: HTML tag broke the message]
  6. quote:Originally posted by Emmett.hendrick: G Force mx 64m(3d Accel) TV GeForce 2 Pro 64m 400mhz DDR AGP Creative Forget the GF2/MX, it's not even in the same class with the other GF2 chips. In performance, the GF2/MX is about on-par with original SDR GeForce, which is at least real-world fillrate wise not that much faster than the Voodoo3 3000 you're replacing. At this point, GeForce2 Pro cards have the best price/performance ratio out of cards based around nVidia graphics chips. But Creative wouldn't be my first choice of brand... 2D image quality of GeForce2 cards varies greatly from manufacturer to another, and Creative's 2D has traditionally been extremely poor. Make no mistake, GeForce2 has a superb RAMDAC capable of Matrox-like 2D. However, quality of the card's construction and the components used in the RFI filters are what make or break the sharpness of 2D on GeForce2 cards. Basing on personal experience, user comments, and Anandtech's roundup of GeForce3 cards (conclusions can be drawn from component quality though the chip's different), I believe Elsa and Gainward GeForce2 cards are among the better ones in terms of 2D image quality.
  7. Anandtech's review of Athlon XPLooks like AMD clearly took the performance crown back from Intel (if it had even lost it in the first place, depends on point of view). The new 1.53GHz Athlon XP has a surprisingly large performance advantage over AMD's previous flagship processor, 1.4GHz Tbird Athlon. Assuming XP is able to ramp up in clockspeed in the area of 1.8-2.0GHz, it may hold it's own even against the upcoming .13m Northwood P4. In my opinion the performance rating system -while being admirably conservative from AMD's part - is unnecessary and perhaps even a bit confusing. It's hard to accept the fact that the average computer-buying consumer is an idiot easily impressed by big numbers.
  8. I have the Asus V6600 Deluxe (GF1 instead of GF2), but in terms of video capture, it's identical to the V7700 Deluxe. As the video capture is not hardware based on this card, supported realtime video encoding formats and in essence the practicality of the function depend entirely on the software used. The buldled Ulead Videostudio is not great by any means, but it can capture video in both MPEG1 and MPEG2 formats. Capturing full-resolution MPEG2 video realtime is going to take a lot of CPU horsepower, however, so you may get frame drops and loss of audio sync if you can't settle for any lower quality.
  9. I was expecting the new NV25 core as well as .13 micron process with the Titanium series GeForce3's. Now it seems that the top model, GF3/Ti500, is only a slight clockrate increase due to improved - but still .15 micron - manufacturing process. Comparing the hardware of Radeon 8500 and GeForce3/Ti500 there is no question which chip is better. Radeon 8500 offers higher clockrate, same fillrate per clock, more features, and most of all, assuming vertex shader implementations are roughly equal, double vertex shader (programmable T&L) performance per clock. As I've stated before; if drivers allow it, Radeon 8500 should outperform the GF3/Ti500 with ease.
  10. quote:Originally posted by Supreme Cmdr: Demo : They are not supported and never have been. So I don't give a toss if anyone wants support for it or not. ... What'd I miss in your post? But don't you think that it's a good idea for developer to maintain a public discussion place for a released demo, even if the demo is unsupported? On a public forum gamers can help each other with known but perhaps still undocumented issues as well as provide feedback. It might be just me though, anything that reduces the amount of redundant incoming email is OK in my book
  11. If registration for this forum is denied anyone who has not purcached one of your games, how are you planning on handling future demo feedback and technical support? In my experience public forum is the very best way to do this. Gamers can help each other if they run into issues that are not listed in manuals or FAQ's, and anyone with a shred of patience can get an answer to a question for which one already exists. It is however a good idea to restrict discussion forums for a commercial game as those who pirated the title won't be able to participate.
  12. nVidia did a really good job with the nForce. They matched performance of VIA KT266A which is practically the fourth tweak to same basic core logic for: - KT133 was KX133 + Socket-A interface - KT133A was KT133 + 133MHz FSB and synchronous operation mode support - KT266 was KT133A + DDR memory controller and Vlink for southbridge interconnection - KT266A is KT266 + redesigned memory controller Also, OEM:s will love this solution. Finally a all-in-one chipset with decent performance. And since OEM market is where the money is, nForce could be a real cash-cow for nVidia. Personally I find the audio part of the chipset, nForce APU, the most interesting one. It is the first ever audio chip to incorporate full DirectSound8 hardware acceleration. And it shows in the real world too, in the tests nForce APU handily slaps SB Live around. What I'd really want to see is integration of the APU's functionality on a video chip. Sound and video go logically hand in hand: with such integration, effects like attenuation and wavetracing could be, at least theoretically, implemented using actual geometry data.
×
×
  • Create New...