My Scripts on ScriptSpot - LC/PPT Calc Updated! - OLD OLD

[studioDIM]

as it is set now, it renders one small view (its size dependant from the "precision" spinner value) to cshow the results, and it should then exit.
You're saying it keeps looping after it pops the "Exposed!" dialog?

You should avoid the vray vfb: have it off.
It's meant to work with a standard vfb.
And it's meant to work .
Seems it only works sometimes :P

Lele

[jacksc02]

yeah once you hit expose it starts looping without showing the VFB

[studioDIM]

that's correct.
It's created in memory only, not displayed.
It only displays it after it's done exposing.
You can check the maxscript listener for the iteration progress.

Lele

[jacksc02]

ah ok....

Another thing, the ESC key is still dodgy. In my scene if I try to cancel it it stops for a second then restarts.

[studioDIM]

dodgy dodgy...
1.1c, you say?
I thought i had an interrupt in every possible cycle, and checked it too...
Will look into it, i'll catch you on msn.

Lele

[studioDIM]

Bump.
Thanks to ChrisJ and DaForce for the invaluable BUGGING.
KEEP the esc key pressed to finally stop the "unmanned freighttrain without brakes" as DaForce kindly dubbed it
"Hey" from ChrisJ sorted the same effect.
He's scary a chap.

First post to grab version 1.1d of the lightmeter/autoexposure/freighttrain

Lele

[jacksc02]

good lad!
keep on truckin..

[studioDIM]

A new EARLY beta of an exposimeter is ready to be downloaded and tried out @ http://www.scriptspot.com/lele/float...meter_0.1b.mse

It's a non-iterative attempt at an exposimeter.
Just pick a cam, and press expose: in most cases it should allow for a fairly precise exposure.
Exposure bias is off at the moment, so do any correction needed by hand on the physcam ISOs (or other parameter).

I would LOVE some feedback on this.
Thanks for your time, kind people

Lele

[jacksc02]

good work lele, I will test her out in the next few days for ya and hopefully I wont find any bugs

[studioDIM]

I am noticing that for some yet unknown reason, on some scenes (purely diffuse ones) the script works perfectly, exposing correctly at the first go.
For scenes with glossies, instead, it seem to have issues exposing the first time around, *probably* due to the glossies sampling, and may need two or three iterations to expose correctly.
I'll put here a sample output to the maxscript listener for those with courage:

"#(0.0165713, 1.0, 0.361193)" these are the three brightness values: minimum, maximum averaged on the brightest 25% of pixels and average brightness for the whole picture
"56.1579" camera ISOs before change
1.0 Multiplier for the camera ISOs. 1.0 means the camera IS exposed, and no change will take place.
"56.1579" result ISOs after change (same consideration as above, when before and after change match.)

I am working on a different way of weighting the pixels, but since my math skills are weak at best, it might take a while.

Lele

[jow]

Works a treat here Lele, Though ive tested on a couple of small test scenes.
I will give it a whirl on a couple of large exteriors tomorrow.
Thanks man.

[RErender]

Hi lele,

Been playing around with Exposimeters and had some comments/questions.

It would be good to set the bucket size in the preset to a smaller size 1/4 or 1/8 of the render size would be good. This will allow multi core cpus to run at 100%

What preset/gi method are they using? seems very slow. maybe allow for loading a custom preset to allow experimenting with speed vs. quality.

[studioDIM]

good idea for the bucketsize, indeed.
As for the method, is purely qmc with 10 bounces, with a good-ish AA thrown in.
Ideally you'd use the script at the start of a scene, where you just balance the lighting on surfaces of largely correct diffuse level, to obtain optimal results.
Also, since the preview is generally very small, there is no need to keep the heavy poly objects in, let's say, meshsmoothed form: a blocking shape would do allright in most cases.
I cannot but agree it can be dead slow when the scene is complex, but the forcing of the settings it's done to have the "quickest" and most consistent way of evaluating a picture.
QMC excels at accuracy, and coupled with the AA gives my script nice, properly averaged pixels *almost* every time
I don't know how bad artifacting would be for other GI methods, in the case of a small picture (ie. at most 150x150 px. generally 50x50 is more than enough).
I can surely try and implement it later on, but bear in mind that the tiny image will put to the test any other interpolated method...

Lele

[jacksc02]

almost every time

[studioDIM]

Err, well.

http://www.scriptspot.com/lele/float...meter_0.1c.rar

Has ISO or Shutter Speed Priority mode, and allows for preservation of the user's AA and GI settings.
It now creates four buckets, regardless of image size, so keep them small and simple
It still has the same "blinding" issue with glossies and speculars in general.
If it exposes oddly, try and turn them off, re-expose, and turn them back on.

If something doesn't work, blame ChrisJ, he betatested it and approved for release...

Lele