Specific question about Irrd. map

[jedisalf]

Ive looked the manual, but the info in there is too technical, not as user friendly for a newbie like me.

Saw i was hoping if someone in here could help me.

Is about the Irradiance Map setting:

The first Irr. map setting are Min rate and Max rate, for me all this does is control the "amount of prepasses" for the GI solution, so if Min = -3 and Max = -1, that means there will be a total of 3 prepasses, right?....well heres my question:

What happens in the following cases:

TYPE A:
min : -4 max : -1 ----> total of prepasses = 4

TYPE B:
min : -3 max : 0 ----> total of prepasses = 4

TYPE C:
min : -2 max : 1 ----> total of prepasses = 4

As you can see all of those are 4 prepasses, but different min and max rates, which one owuld be better quality?, or what are the differences between them?

thanks.

[Da_elf]

let me see if i remember the technical part correctly. a sample rate of 0 means its going to take a sample for every pixel. and -1 means every 2 pixels and so on and so on going to lower quality the lower the number. above 0 means it takes its more samples per pixel.

In areas where there isnt much detail it doesnt need much samples so it will go with the lower number for those areas and for areas with lots of detail it will go for the max number. therefore irradiance maps are adaptive in this sence.

Even though the passes are equal in all of those your min and max are what determins the quality. not the amount of passes. on a scene that has very simple flat surface on the left and complex stuff on the right you can try -4 -1 and it will be 3 passes. the left side will calculate faster cuz it will use -4 in that area. now if you used max/min of -1 -1 then you will get the same look on the right as the before mentioned render. however, using -1 for the left will be a waste of time for nothing. sure you only have 1 pass not but still...you loose tonnes of time calculating the left side.

[jedisalf]

ok...i kinda lost you wit the lkeft and right stuff, cuz it looks like youre saying that MAX is for everything on the left side, and MIN is to calculate whatever is on the right side. I thinkyou didnt menat that, you were just trying to explain, right?

Well, what i understand after carefully reaidng what you said is this:

the lower the number --> more samples per pixel --> more quality

the higher the number --> less samples per pixel ---> less quality

Is that right? when i say "lower numbers" i mean negative...-4 takes more sampples than -2, but -8 takes even more samples, and so on...

Well according to this, and with my first question exmple,
the TYPE A will be better quaility than TYPE C.

i know this depends on how complex the scene is. The complexity of the scene, and the "quility on details" is what determines which MIN and MAX setting to use.

[losbellos]

actually you dont get the same look, it is very important to choose the right values for the beginning and the finish the highest quality.
for exapmple a -2/-1 will have much more detail than -4/-1.
It because most of the first pass, which makes vray to take more or less samples based on the first pass samples. If you have less there will be some areas that will sacrifice quality even if you do -4/0 or more, while -2/-1 or -3/-1 will do the need.

[jedisalf]

for exapmple a -2/-1 will have much more detail than -4/-1.

ok...now im officially lost.



I though lower -4/-1 is more samples per pixels than -2/-1, hence more time consuming, more quaility.

[jedisalf]

wait, wait..i think i know what you mean.......im doing some renders right now.....and i notice that -4/-1 is faster than -2/1, with the last setting it takes more time on the first pass than the previous setting.

I just though it was the other way around....

[J_Bug]

It's a refinement process so whatever the settings may be... the first pass gives a general idea of the lighting, 2nd gets finer, 3rd looking better...

As this goes on the irrad map becomes more dense with a smaller more detailed set of samples.

--Jon

[Da_elf]

-4 is a lower number than -2. meaning using -4 as a sample pass is less quality than -2. if you look at it when it does -4 its in big blocks then -3 the blocks are refined and are half the size of the -4 blocks and half that at -2 etc etc till you get to 0 where each sample point is 1pixel x 1pixel