Help needed on massive landscape
nereus541_f392973b9f
Posts: 24
I will admit that I'm a bit of a noob dealing with Bryce. I'm trying to develop a relatively accurate representation of a coastal area with offshore islands. The area is massive (1024 sq km), but the LOD is important to the scenes.
What is the best method of creating the landscape for use in a self-directed app that includes fly-overs, views from shore and views from sea? Take into consideration that 58% of the landscape is ocean with land backing the shore and sea backing the islands. Cliffs, a range of slopes and inland volcanic lakes are all part of the landscape.
I'm not asking much, but HALP!!!
Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Bryce can import DEMs (digital elevation maps). Just find the appropriate one on the Internet. Bryce can also use PGMs (portable greyscale maps), also many free on the Internet. High resolution ones usually are not free. You can also draw the height map in a graphics application. Black is low, white high. It is important to save the height map as 16-bit greyscale TIFF (not 16-bit color).
I have ASCs, DEMs, BILs, TINs, TGAs, BTs, PGMs, BMPs, JPEGs, plus another 32 formats available. That still doesn't answer the question. Based on most of the tutorial recommendations, I have between 64 and 256 segments to consolidate into a cohesive landscape.
Due to the need for an aerial view plus accuracy to the horizon in the shore-to-sea views, I'm asking for guidance on dealing with massive landscapes where the level of detail must match the real world. Any recommendations?
I understand what you are asking. Your questions doesn't really have a simple answer. I uploaded a few hours ago some samples of a project I am working on regarding an island of sorts. Is this the type of thing you are looking for?
http://www.daz3d.com/forums/discussion/7547/
Yes. Plan your shot sequences first. There's absolutely no point trying to create a 'one model fits all' solution, dragging your computer's performance down, your scene navigation down, and your ability to organise your scene down if you're creating stuff that will never be seen.
There's a saying in video production: "It doesn't have to BE right; it just has to LOOK LIKE it is." If your first shot is an aerial view above the cloud layer where atmospheric haze reduces clarity, designing a multi-gigabyte scene with all the detail in the world is flushing your time and CPU power down the toilet. A 512-resolution mesh and a cloud layer might be all you need for the first 5s of your video: render that out, and while it's rendering, model the next shot in another instance of Bryce, which will be a cut from the aerial view to... whatever the next scene is, only closer, and facing west. Render that out, and then load up the eastward facing scene. Whatever.
DO NOT, UNDER ANY CIRCUMSTANCES, think a 'power-drop' from 20,000—2 m is going to be worth waiting for. Don't try to 'eat the elephant' all in one gulp. Plan your shots first, then frame and model accordingly.
Yes. Plan your shot sequences first. There's absolutely no point trying to create a 'one model fits all' solution, dragging your computer's performance down, your scene navigation down, and your ability to organise your scene down if you're creating stuff that will never be seen.
There's a saying in video production: "It doesn't have to BE right; it just has to LOOK LIKE it is." If your first shot is an aerial view above the cloud layer where atmospheric haze reduces clarity, designing a multi-gigabyte scene with all the detail in the world is flushing your time and CPU power down the toilet. A 512-resolution mesh and a cloud layer might be all you need for the first 5s of your video: render that out, and while it's rendering, model the next shot in another instance of Bryce, which will be a cut from the aerial view to... whatever the next scene is, only closer, and facing west. Render that out, and then load up the eastward facing scene. Whatever.
DO NOT, UNDER ANY CIRCUMSTANCES, think a 'power-drop' from 20,000—2 m is going to be worth waiting for. Don't try to 'eat the elephant' all in one gulp. Plan your shots first, then frame and model accordingly.
Another reason for that...Bryce is a 32 bit app with a hard memory limit. Monster shots are going to eat RAM and if the scene gets too big...it will crash (sorry it doesn't exit gracefully). Doesn't matter how much RAM/CPU power you've got, Bryce has a 2 gig hard limit.
Another reason for that...Bryce is a 32 bit app with a hard memory limit. Monster shots are going to eat RAM and if the scene gets too big...it will crash (sorry it doesn't exit gracefully). Doesn't matter how much RAM/CPU power you've got, Bryce has a 2 gig hard limit.
Yes but there is also the Large Address Aware solution that allows you to access up to 3.4gb if you have at least 4gb of total system ram. That nearly doubles the amount of useful memory so it shouldn't be overlooked.
Check these out:
http://www.techpowerup.com/forums/showthread.php?t=112556
I believe the utility even works with Bryce 5 and the like. Isn't that amazing!
Edit: One of the links I included no longer functions. It lead to a Daz thread about Large Address Awareness where it is discussed in detail. I will find it for you but know that the utility is safe and super effective.
I'm not so concerned about the aerial, except I keep having difficulty separating land from sea on the world surface. (That's just a matter of getting my head around rendering techniques.)
What scares me is the widescreen view of islands in the distance that are key to the scene. I'm dealing with almost a dozen islands that range in distance between 2km and 21km from the shore. (The horizon is 24km from my viewpoint.) With the interactive nature of the app, it's a blur-jump to one of those islands or a point out to sea, looking back to the point you just departed.
It does, even with Bryce 4.
@Nereus541 - have you had a look at the tutorial thread http://www.daz3d.com/forums/discussion/2839/
Good grief, where are you: Tahiti?
OK, so you're about 15-30 m above sea level (2-5 m above sea-level gives you a horizon at about 15 km, give or take). You have an app that allows travel between images from one island to the next using a combination animation and blur effect, staying at the same altitude.
Sounds like you need several sets linked to the 'fly-to' set. The 'fly-to' set is a navigation master: this is the archipelago, laid out using low-res islands on an infinite sea plane, accurate in relative position and scale, but not in color or detail — they're too far away from each other visually to worry about the nitty-gritty.
Another way of looking at this is in terms of designing your own personal HDR background; that's basically seascape on the ground, islands on the horizon and a lovely day up top. You'll need as many HDRs as you have travel locations however.
ANOTHER another way of thinking is using QTVR exports of your scene at each destination location, and using a repeated 'teleport' sequence to get to the next point of view. So it works like this: You have one scene per island, each with great detail AT the location, and fuzzy, low-res representations of your destination islands sitting on the horizon. QTVR technology allows you to set hotpoints in your image. While at your start location you're able to use the mouse to rotate your view on-the-spot. When a hotpoint is clicked, the 'fly-to' animation fades up, the new QTVR movie is loaded, and the viewer looks at a new island location... with hotpoints going to the other islands.
Using the QTVR method, you create the low res archipelago first to scale, light it, and save it as a reference master.
Then you copy this scene file, set your camera to one of the islands and build a detailed set just for the island you're at. Then create a QTVR movie for that location.
Reload the low res master again and do that for the next island, and so on. Like Street View for Google Maps or Google Earth.
I have ASCs, DEMs, BILs, TINs, TGAs, BTs, PGMs, BMPs, JPEGs, plus another 32 formats available. That still doesn't answer the question. Based on most of the tutorial recommendations, I have between 64 and 256 segments to consolidate into a cohesive landscape.
Due to the need for an aerial view plus accuracy to the horizon in the shore-to-sea views, I'm asking for guidance on dealing with massive landscapes where the level of detail must match the real world. Any recommendations?
What he's suggesting, (the use of DEM's) is the answer. A good DEM of a particular landscape is made from an aerial view and is real world accurate because it's made from the real world landscape. He's telling you Bryce can read those and convert them into a terrain in Bryce. You're not going to get any more "real world" then that.
I'm kind of scratching my head though as to why a self described noobie to Bryce is trying to tackle what sounds like a project that would challenge hardcore experienced professionals? Then expecting everyone else to solve it for him and arguing with the ones that try to?
Basically this project is an elephant dinner...you need to break it down in small enough bites to eat, then before you know it, the elephant will be eaten.
Here are a few examples made using DEMs in Bryce:
http://www.bryce5.com/details.php?image_id=4133
http://www.bryce5.com/details.php?image_id=1453
http://www.bryce5.com/details.php?image_id=577
http://www.bryce5.com/details.php?image_id=571
http://www.horo.ch/raytracing/gal/img/t0164.jpg
http://www.horo.ch/raytracing/gal/img/t0118.jpg (Venus Beta Regio)
http://www.horo.ch/raytracing/gal/img/t0117.jpg (Venus Aphrodite Terra)
INORITE?
I'm ok with animating in Bryce, but the thing I've learned working with multimedia for the last 15 years is: use the tool that's right for the job.
Seriously: get into a helicopter kitted out with a gyro-stabilised camera rig. That will cost you about US$1-2000 an hour, with a cameraman extra for $500 a day. Fly between your islands for 3 hours, collect your footage and YOU ARE DONE, dude. That's $6000 expenses, with minimal editing time, almost zero setup time and you can deliver the photorealistic product to a happy client inside 5 days for $9000.