PPT for lightcache or irradiance map?

[rpc212]

Hi,

I am wondering if one can use a PPT saved map in lightcache or irradiance map as a saved solution? I believe this was talked about a year or so ago, but I may be totally mistaken.

Thanks!

[vlado]

You can do this, yes.

Best regards,
Vlado

[Ryan_Chiong]

welcome back vlado

/Ryan

[jonnybefree]

Re: PPT for lightcache or irradiance map?

Hi,

I am wondering if one can use a PPT saved map in lightcache or irradiance map as a saved solution? I believe this was talked about a year or so ago, but I may be totally mistaken.

Thanks!

i think you can't save an unbiased lc. so you have to leave the sample size bigger than 0 for saving it.

[vlado]

This is true, yes - the sample size must be greater than 0.0.

Best regards,
Vlado

[rpc212]

Hi,

So if I make the sample size 0.001 and save the solution I can use that solution as my lightcache or irradiance map?
I am surmising that the lower the value (eg. 0.001 as opposed to 0.9) would give a more "accurate" solution (almost unbiased solution), that could then be used in a biased rendering. Is this summation correct?

Thanks for you help!

[studioDIM]

correct, afaik.
If you want more control, try use the LC/PPT calculator script (OT sticky thread, in the pack), and set sample size to 1.01 pixels.
That's about as good a biased solution as it gets

Lele

[rpc212]

That's about as good a biased solution as it gets

What do you mean Lele exactly?

The most accurate or closest to an unbiased solution as a biased solution will get?

[joconnell]

If you're using light cache in screen space mode, the radius for the light cache samples is expressed in pixels so if you had a sample size of 1 pixel it means you'd get a sample for every pixel in the final render which is pretty much the same as using qmc or unbiased which is as high quality as it gets. I'm not sure if you can use 1.0 pixels as the sample size with this method so what lele is saying is use a sample size a tiny bit more than that like 1.01 and it'll be pretty damn sharp.

[rpc212]

Ah, thanks joconnell. I understand now!