mental ray render time

[kicks]

buddies, can anyone be my third eye?

 

i can't seem to know where/what is wrong.

i ran a preview on mr 3dsmax6 which took almost 7 hours.

>is my file too big?

>something wrong with my mr settings?

>or my pc is just slow.

 

polygon count: 860228

render time is 6:51:04

running on p4 2.4 512ram

 

tnx

[jucaro]

EEEK!!

 

Its you settings man.

From what I've learned so far with MR is to never use both GI and FG with equally high amounts of sampling rate.

 

If you do the math, (taking into consideration your current sampling rate) for every GI Photon sample cast in the scene, it will be refined 2000 times by the Final gather engine. AND if thats not enough to lengthen your render times skyhigh, your radius for GI and Photon is set to .3 which mean, there will be 2000x2000 samples in every .3 radius units in your scene (which explains the splotchiness of the result)

 

Try to uncheck PREVIEW and check Rebuild.

 

Start with a modest rate in GI, say 50 then 100 for FG, then work your GI samples.

 

Good luck

[kicks]

i'll try for the weekend. tnx very much

 

pls. checkout this tutorial settings...

http://www.evermotion.org/tutorials/rendering/mr_interior/index.htm

[abicalho]

Start LOW and go UP. Don't start with 2000 FG samples. Nor using FG. With the sampling rate you have in GI you may not even need FG.

 

Start Low (50 samples) and go up. See the quality increase. And as it increases, keep on going up until you reach the quality you want.

 

Alexander

[joske]

make sure your units AND system units are checked

turn on the 'cm' or meters in your units to see what you're doing

eg.: photon radius = 0,3 ???? 0,3 what??? meters, millimeters,....???

make sure once you got your units worked out ok, that your objects in the scene are drawn in the right scale

 

if you draw a box around one of your stairs it should add all up : the chair cannot be 5meters-2meters... you'll get the idea right?...

 

if you don't do all this, it will be impossible to achieve workable rendertimes vs quality... trust me

 

good luck

[kicks]

tnx. wil post adjusted settings & final render later.

[kicks]

'am not quite done with this.

>problem: why am I having "pixelated" result in final gathering? (see attached)Image was good and rendered fast on gi.

>i tried tweeking almost everything but could not shake it off. I was thinkin something wrong with sampling adjustments(samples/pixel,spatial contrast,etc.) even maybe the fg sample & max/min values, but nothing helps.

>hope someone? help me out on this one.

 

help is much appreciated

[Fran]

Hi,

 

I get that almost every time I use FG and have a manual clipping plane on the current camera.

[kicks]

same here, camera clipped near@3.2m, far@20m

[abicalho]

By using Fast Lookup you're throwing away the quality. You need to turn it off.

 

Then you need to adjust the Min/Max Radius so that they generate a good result. A Max radius can't be higher than the size of your bucket. A Min Radius would be around the size of the smaller edge in your rendered view, or around 10% of the Max Radius. There isn't a rule, just guidelines for starting.

 

Last, start with 100 samples to try the Radius, increase the number of samples as you figure out better quality.

 

Alexander

[joske]

you might wanna use GI in this case, then you can have a very low FG setting and have faster rendertimes... for equal results

first try to get a good feel without fgather, make your GI photons about 75 cmeters and use only 5000 to see them and to check the size.

then increase them to like 20000 or more until you get a nice overlap of them

everytime they overlap, they will average and smooth out the result

 

then turn on a final gather not bigger then 20-30 to get rid of the biggest GI spots.

then when all is looking better increase FG untill all looks crisp (2000 should be more then enough if the GI is setup right...)

this is my workflow... not the only way to go.

 

if you can't get it right after trying yourself just post your max scene so other people can have a look at it... the different workflows will be very helpull...

[kicks]

alex,

I was really thinking fast lookup has got nothing to do with quality rather than fg render time?I dunno ,here's quote from max help

Min Radius—When on, sets the minimum radius, within which final gathering must be used. Increasing this value can improve render quality but increase rendering time. Unavailable unless Max Radius is turned on. Default=0.1 (Ten percent of the Max Radius).

 

 

tnx

[kicks]

philip,

 

tnx. 'will try your workflow..haven't tried jacking up gi to that #...will try & post result. this should work, 'hope.

 

http://www.impresszio.hu/szabolcs/MentalRay/MentalRaySampling.htm

[kicks]

& still got the "pixelation"

[voltaire_ira]

By using Fast Lookup you're throwing away the quality. You need to turn it off.

 

Then you need to adjust the Min/Max Radius so that they generate a good result. A Max radius can't be higher than the size of your bucket. A Min Radius would be around the size of the smaller edge in your rendered view, or around 10% of the Max Radius. There isn't a rule, just guidelines for starting.

 

Last, start with 100 samples to try the Radius, increase the number of samples as you figure out better quality.

 

Alexander

as i have learned from others, the max radius depends on the scene extents. say the scene is about 13 meters along the longest dimension, the max radius for the fg will be 10%. that would be 1.3 m max radius then min radius will 10% of the max radius so that would be .13 m.

 

I have not tried using the max radius equal to or smaller than the buckets. I usually balance the photon radius wiith photon count so as to minimize the use of high fg samples. so far i get good results. but if the max is less than the buckets and produces good results then i'll be trying that!

 

kicks,

 

your settings are waaay to impossible to render!!! as others have mentioned, try working with small number of samples then increase little by little.

 

to remove the noise from your fg solution, either increase fg count or increase photon radius and max trace depth.i'll go with the increase in photon radius. it will smoothen out the gi but sacrifice in shading quality. its just finding the right balance between render time and quality.

[kicks]

yup 'am way to hardheaded...reworking small now...

[abicalho]

The problem is where you're adding overhead - you're increasing the number of samples, but you have a small number of phoons. Try the oposite: increase the number of photons (GI Photons at the bottom) and keep Samples small (100-500).

 

I have been successful in scenes lately using the process above (maybe even 1M photons), and not even needing FG.

 

I'll take a look at the clipping planes thing that Fran mentioned.

 

Alexander

[abicalho]

as i have learned from others, the max radius depends on the scene extents. say the scene is about 13 meters along the longest dimension, the max radius for the fg will be 10%. that would be 1.3 m max radius then min radius will 10% of the max radius so that would be .13 m.

 

I have not tried using the max radius equal to or smaller than the buckets. I usually balance the photon radius wiith photon count so as to minimize the use of high fg samples. so far i get good results. but if the max is less than the buckets and produces good results then i'll be trying that!

 

Yes, you're right. What I said is nonsense, because of the perspective view... I was thinking of something else.

 

I tend to avoid FG myself.

 

Alexander

[kicks]

alex,

>u got good point there. I am trying that one now & felt better with the results.....

[voltaire_ira]

The problem is where you're adding overhead - you're increasing the number of samples, but you have a small number of phoons. Try the oposite: increase the number of photons (GI Photons at the bottom) and keep Samples small (100-500).

 

I have been successful in scenes lately using the process above (maybe even 1M photons), and not even needing FG.

 

I'll take a look at the clipping planes thing that Fran mentioned.

 

Alexander

i agree i agree!!! i did render a scene with millions of photons and rendered a scene with no fg..it looked great! then i added 100 fg sample and it even rendered better!!!the render time was really quick.

[kicks]

[voltaire_ira]

based on your gi settings...

 

you have way too low gi/caustic photons ( 1 )...try having the same amount with the gi samples ( 200 ). also try measuring your scene with max measuring tool so that you will have an ideal fg radius size ( you didn't use one in your settings.

 

why not post your scene and maybe the guys could give you different workflows and settings and maybe better results...i'd give it a shot..kabayan.

[kicks]

thanks kabayan