Page 5 of 5

Posted: Tue Dec 09, 2008 12:33 am
by dougal2
@CTZn:
Sorry, SVN has been down since thursday morning because my broadband is broken :(
I hope that the engineer who comes tomorrow can fix it. Faulty modem I think.

Posted: Tue Dec 09, 2008 7:32 am
by CTZn
Aknowledged Dougal, really really no harm, thanks actually ;)

Let's hope the guy will bring your network back "fingers in the nose" (easily) !

Posted: Wed Dec 10, 2008 1:08 am
by dougal2
OK, my connection is back, but I've had to re-arrange my setup a bit, I haven't brought all the servers back online yet.
The problem was that the cabling in my house was causing noise on the broadband system, so they disconnected me without warning :(

I've had to move modem/routers/servers to a different room, hope to have completed this evening.

Posted: Wed Dec 10, 2008 3:31 am
by CTZn
so they disconnected me without warning
Roaches !

Thanks for reporting anyway... I'll use the server as soon as it comes back online ;)

Posted: Thu Dec 11, 2008 2:09 am
by Hellstorm
Heyas. Maya is doing some batch rendering and I have thought about the FStop Thing. Apperently you shouldn't use the Global Scale Factor there.
Maya Units for Focal Length are always millimeters, film aperture is always inches. Because those units don't change I think it's wrong to use the $mtiGSF_Cam variable there.

The main reason for GSF is to compensate the unit differences from Maya to Indigo, right? However, focal lenght always has a factor of 1000 because it's always millimeter.

So without GSF and According to http://en.wikipedia.org/wiki/F-number and http://www.uscoles.com/fstop.htm the formula is correct.

As long as no one wants a camera with a lens diameter of some meters (like hubble telescope) I would use the formula without that GSF. If you want a scale conversion also for the camera I would advise an additonal scale conversion factor for the camera.

Edit: You should also note that a telescope has a pretty high "sensor distance" this stuff is pretty advanced and uncommon. So I guess we should use a common and physical correct camera model.

Posted: Thu Dec 11, 2008 6:22 am
by CTZn
Edit: You should also note that a telescope has a pretty high "sensor distance" this stuff is pretty advanced and uncommon. So I guess we should use a common and physical correct camera model.
We never discussed about making telescopes with Indigo, a common and physical model was our aim... well, I'm pretty sure of that ^^

Great, I'm taking what you say in consideration as I plan a release in the 12 following hours; i'll stress that. Thanks !

No real big deal with the upcoming release, mostly I fixed meshlights (efficacy scale optional and working for all spectrum types; uniform spectrum was implemented for a time but I removed it).

Different spectrum types than RGB can be used to define materials values, too.

The main documentation source is the wiki, as that was planned before it went down. MtI wiki is wip, I'll rework it in depth at some point.

Posted: Thu Dec 11, 2008 8:00 am
by CTZn

Code: Select all

Completed: At revision: 278 
Shazoom !

Ah well I'll look into the camera before I pack it all... Lets start with some test renders.

Posted: Thu Dec 11, 2008 9:14 am
by CTZn
Heyas. Maya is doing some batch rendering and I have thought about the FStop Thing. Apperently you shouldn't use the Global Scale Factor there.
Maya Units for Focal Length are always millimeters, film aperture is always inches. Because those units don't change I think it's wrong to use the $mtiGSF_Cam variable there.

The main reason for GSF is to compensate the unit differences from Maya to Indigo, right? However, focal lenght always has a factor of 1000 because it's always millimeter.
[...]
I read you again and without further verification that sounded logical to me.

Now from the old (web) times I only can remember dougal2 saying something was maybe wrong in the camera code, then I added mti_GSF to the thing... Certainly that was not a wise fix.

Releasing soonish :D

Posted: Thu Dec 11, 2008 9:39 am
by Hellstorm
Yeah. Think about the following:

Maya units is centimeter, focal length is (always) millimeter.
So you will multiply all units by 0.01 to get indigo meters.

So 100 maya cm will be 1 indigo m. However:
We have a camera, focal length of typical 35mm and a "sunny-day-fstop" of f/16.
Focal Length is within the script always converted to meters.
So we have a FL of 0.035 [m].
The aperture therefore is 1.1 [mm] or 0.0011 [m].

However with this multiplication of (1/GSF), and GSF=0.01, we would multiply this 0.0011 [m] to 0.11 [m]. I've never seen a camera like that :D

I really think the old formula was correct. :)
May be there's a bug, but at another place. I'll go further thru the script.

Besides... I would suggest a default F-Stop of f/16 instead of f/50. A F-Stop of f/50 is pretty high. Most times I use a f-stop between f/5.6 and f/22.