Severe need for documentation

Started by PabloMack, April 18, 2010, 01:25:43 PM

Previous topic - Next topic

jritchie777

Quote from: dandelO on May 24, 2010, 07:04:10 PM
Quote from: jritchie777 on May 24, 2010, 01:41:08 AM
What are the limits on the input fields???

JR

Well, there's the thing. There are no limits for most of the input fields. You have pretty much complete and unrestricted control of your creativity.

Only problem when the program crashes due to not liking some of the values I've put in, then my creativity becomes a guessing game...

neuspadrin

Quote from: jritchie777 on May 24, 2010, 07:46:09 PM
Quote from: dandelO on May 24, 2010, 07:04:10 PM
Quote from: jritchie777 on May 24, 2010, 01:41:08 AM
What are the limits on the input fields???

JR

Well, there's the thing. There are no limits for most of the input fields. You have pretty much complete and unrestricted control of your creativity.

Only problem when the program crashes due to not liking some of the values I've put in, then my creativity becomes a guessing game...

My general rule of thumb is if I don't know what it does, add/subtract the current value divided by 10.  See how much it affects it, and then modify again depending on how much more i feel i need.

Example, the default value is 1000.  So 1000/10 = 100.  So I pick a value between 900-1100.  Then change depending on that.

One exception is values between 0-1.  These often in TG2 will represent a % value of 0-100% (with 1 being 100%).  So if its a 0, a 1, or any decimal between I'll try out other decimal values or switching the 1 to a 0 or a 0 to a 1 first.  Then depending on how that worked out move on to bigger values.

Kadri

#62
I am more of a hobbyist so it doesn't much matter for me.
But every time i see such topics i remember what my friend said years ago .
( I hope i didn't wrote this before  :) )
I think this idiom is in many languages " By trial and error " !

My friend asked me something about a program. He had to meet a deadline for work .
I said " You can try ".

He said " I have time for trying , but not for error! .




dandelO

Aye, I do see your point, Jritchie777, it isn't ideal at all.

I don't think I've ever crashed TG myself because of any extreme inputs, though. And I really do use some ridiculous settings sometimes. What parameters do you find these problems mostly occurring?

I could imagine some really extreme settings that might cause crashes, if you were to input, say, 1,000,000,000m as a displacement multiplier on a near-to-camera terrain then, yes, the renderer would most likely crash the program due to impossibility of rendering such extremes, if the 3D preview doesn't crash it beforehand, that is. I couldn't really see anyone using such extreme settings as this, though.

In another thread yesterday, as an example on extreme inputs, Seth wrote a post on using, I think it was a value of 10,000 for transparency, on a water shader applied to a sphere. The transparency slider's limits here are '1', this doesn't mean you can't input higher.
I've experimented on this myself since then, and it appears that if you raise transparency above '1'(the logical maximum) then, sure, it isn't physically correct but, Terragen will do as you tell it, most of the time. This results in a physically impossible calculation, of course, how can you possibly go beyond 100%('1') in the real world? But, TG doesn't really discriminate in this way, it'll take that value and do the mathematical sum of your inputs, in this case, some new way to create luminosity that will cast off from the given surface onto surrounding elements.

'10,000% Pffffft! No problem. What else have you got?' Says Terragen.

dandelO

#64
Neuspadrin has some good advice there, too. ^^

Increments of the value in a given field's defaults are the way I usually experiment. It's extremely difficult to see the difference between cloud internal scattering values of, say, '0.25' and '0.3'. Try changing things logically, maybe from 0.25 to 1. This will let you see better what this setting is actually doing.
I often use internal scattering of 5+, I've went even further, of course. I think my volumetric fire file in the sharing forum might use about '8' in this field, I'm not sure now.
Regardless of the exact number I used, a value of '5' here is 10 times that particular slider's limit...

A lot of applications will not let you increase such settings, embrace it! ;)

EDIT: In fact, the actual value used here is '10' so, it's 20 times the 'maximum'. http://forums.planetside.co.uk/index.php?topic=9039.0

jritchie777

I find that the area that gives me the most trouble for crashing is using negative values for manipulation of terrain.  Either into the power fractal, or one of the other displacement shaders.  I start getting interesting results until I touch one - then bam, T2 crash...

I can't help my obsession with negatively displaced terrain.
JR

Henry Blewer

I am going to have to try the negative displacement you mention. I am curious what is going on. It could be the negative displacement is forcing a division error of some kind (divide by zero?)
http://flickr.com/photos/njeneb/
Forget Tuesday; It's just Monday spelled with a T

Tangled-Universe

The b*tch with settings like "fake internal scattering" is that the outcome of the setting heavily depends on:

the lighting of your scene, sun's strength e.g.
chosen cloud color
chosen scatter color
chosen density
chosen edge sharpness to a lesser extent
chosen light propagation and propagation mix
and to even make it worse, also the fake ambient settings affect how changes to the fake internal scattering setting look...

All in all, with or without documentation, it is really a matter of experience on how to use them in your advantage.
Given that you might have all these settings well documented, you still need to find out yourself how the settings are balanced and how to balance them for every scene you're working on. These kinds of things can NOT be documented.

Martin

rcallicotte

Not to argue a clear point, Martin, but you just did document this.

Someday, it would be great, if people like Martin and whoever <make a list here> could get together to write chapters on their various strengths to enhance the basics to the place of understanding you're talking about here.
So this is Disney World.  Can we live here?

gregsandor

Has anyone else found crashing on negative displacement?  I've been getting memory errors causing crashes, but chalked it up to the large masks and populations I've been using.  I do have some negative deisplacements on my terrain, but all less than 1 meter in depth.

Henry Blewer

It seems to work for me. Sorry...
http://flickr.com/photos/njeneb/
Forget Tuesday; It's just Monday spelled with a T

dandelO

I remember, in a certain couple of updates, that negative displacement in PF's was broken for a time. It worked, then didn't, now, it does again. Apparently, there shouldn't be any problem with negative displacement values in a fractal any more.

nesthead

For now we have a plan for taking the next step and we'll see where that leads. If progress is not rapid enough, we may indeed have to hire someone on contract to help out.

- Oshyan

(April)

Sorry to nag but how's it coming?

gregsandor

It works fine now.  The only problem I've had lately is out of memory errors due to my use of dozens of 8192x maps :) at the same time.

Quote from: dandelO on May 27, 2010, 10:13:23 AM
I remember, in a certain couple of updates, that negative displacement in PF's was broken for a time. It worked, then didn't, now, it does again. Apparently, there shouldn't be any problem with negative displacement values in a fractal any more.

Henry Blewer

Try using two color images. Power fractals can be merged in to provide variation to the coverage.
http://flickr.com/photos/njeneb/
Forget Tuesday; It's just Monday spelled with a T