455 replies to this topic

[hrontos]

KielDaMan, on 28 August 2012 - 05:58 AM, said:

I've experienced this recently with my last render attempt. After I thoroughly enjoyed my Batman Tumbler render, I became overly ambitious and tried to render this "nightmare" of POV-Ray:

After more than 30 hours, the rendering stopped at around 35%, the monitor changing to a blue screen with the text "physical memory dumped" (that's what I remember) then the computer shut down. I know the model is composed of A HUGE amount of trans pieces, but I was ready to be patient even if it took days for this to finish, though guess I reached the limit of my computer's capabilities (Intel Core i3-2310M CPU @ 2.10GHz, 2.00GB RAM, 64-bit OS) so the computer just "gave up". [Default LDD2PovRay settings (no changes), 512x384 No AA, 1366 x 768 image resolution]

It's a pitty, but you still can restart the rendering. If you used the ini file add line with Continue_Trace=On to the ini file. If you did not used ini file, put +c into the text box which is next to the resolution dropdown list in POV-Ray GUI. POV-Ray will continue from the momement when it was aborted.

I had similar problems, without blue screen, but with sudden shutdowns, due to processor overheating during hot days. I used power management to reduce the maximum frequency of the processor.

[hrontos]

For this rendering I did not used any traditional light source. I only added 5 emissive panels, so whole lighting is based on radiosity calculation. It produces nice soft shadows. Shading is a bit flat, since I used 10% ambient light value. May be I should stay with 0.



1366x768, 1803x1014, 2732x1536

Model was not built by me, I took it from the official set in LDD topic and made some correction of colors of transparent parts. But some other colors are also wrong (at least the version I have at home does not have black base plates and also some other pieces are different).

[Bobsy]

I know I'm inviting mod-wrath by not posting this in the minifig customisation board, but it pertains to LDD2PovRay directly, so:

Can anyone set me up with png format templates for faces, torsos and the like in the precise proportions that LDD uses? I've done some errant fumbling with the decals others have made but with no real luck and as far as I can tell LDD does not allow you to copy/edit/view the decorations files at all. A file with a basic face to help with positioning eyes and mouth, as well as the correct dimensions to ensure it doesn't squish or warp when used with LDD2PovRay, that'd be grand.

[hrontos]

Bobsy, on 08 September 2012 - 10:00 AM, said:

I know I'm inviting mod-wrath by not posting this in the minifig customisation board, but it pertains to LDD2PovRay directly, so:

Can anyone set me up with png format templates for faces, torsos and the like in the precise proportions that LDD uses? I've done some errant fumbling with the decals others have made but with no real luck and as far as I can tell LDD does not allow you to copy/edit/view the decorations files at all. A file with a basic face to help with positioning eyes and mouth, as well as the correct dimensions to ensure it doesn't squish or warp when used with LDD2PovRay, that'd be grand.

Original LDD decorations are png images having dimensions only 128x128 pixels. They may contain transparent areas - brick color is then visible in these areas. LDD2POVray displays the original decorations in their original size, with checkers pattern background which can be seen "through" transparent areas. As a starting ppoint, you can make a printscreen of the LDD2POVray window displaying the desired decoration (the same can be done by aplying decoration to 2x2 tile in LDD and positioning the camera to look directly at the tile, so I do not consider this little trick as a violation of anything).

Actually LDD is using uv mapping to map triangles from decoration to triangles on brick. This makes idependent of the size and aspect ratio of the decoration. This means, that you can make decoration having the same aspect ratio as target surface and use it. It is easier to draw it like that. It is quite dificult to draw it square and try to estimate how it will look when stretched to some rectangular surface. It is easier to make it rectangular. You can put your existing real minifig on the scanner, scan the body, crop away the unwanted area and use it directly. This test is quite quick. If you will be satisfied, add those transparent areas and you will have nice new minifig body decoration.

[MarkusOSX]

Here´s my render of a 2-Speed Gearbox i´ve build.

http://flic.kr/p/d8xh6A

Greetz

Markus

[ADHO15]

I'm not sure if it's acceptable to request here, but does anyone fancy rendering my Jedi Temple MOC and sending it to me? I couldn't get LDD2PovRay working on my laptop.
The LXF File can be found in the linked topic.
Thanks a lot.

[Bobsy]

ADHO15, on 10 September 2012 - 04:00 PM, said:

I'm not sure if it's acceptable to request here, but does anyone fancy rendering my Jedi Temple MOC and sending it to me? I couldn't get LDD2PovRay working on my laptop.
The LXF File can be found in the linked topic.
Thanks a lot.

Check the thread.  



(actually earlier this morning while I was at work. But still)

Edited by Bobsy, 10 September 2012 - 04:18 PM.

[Seluryar]

After doing a reformat, I lost my quick way of rendering, I think I have found my sweet spot, is there a file where I can backup my new default settings?

[hrontos]

Seluryar, on 10 September 2012 - 05:11 PM, said:

After doing a reformat, I lost my quick way of rendering, I think I have found my sweet spot, is there a file where I can backup my new default settings?

Currently the only way is to run RegEdit.exe and export key HKEY_CURRENT_USER\Software\LDD Tools\LDD to POV-Ray Converter to a reg file. This reg key contains the last saved default settings so do not forget to save you current settings as default (by clicking "Set as defaults" menu item in the Settings menu in the LDD2POVray.)

[Seluryar]

Thats what I did, set them as default, My sweet spot may have included checking 'Don't Generate Includes'

[hrontos]

Seluryar, on 10 September 2012 - 06:09 PM, said:

Thats what I did, set them as default, My sweet spot may have included checking 'Don't Generate Includes'

The only purpose of the "Don't generate includes" checkbox is to avoid problem with LDD exclusive use of db.lif file. When model is converted for the first time, it is necessary to generate missing includes (if any). When the same model needs some small adjustments like change of colors or camera view, it is annoying to open/close LDD over and over to change, convert, try in POV-Ray. So it is better to check it, access to db.lif will not be necessary and LDD can be running during the conversion,

[hrontos]

1366x768, 1803x1014, 2732x1536

[Seluryar]

hrontos, on 10 September 2012 - 06:33 PM, said:

The only purpose of the "Don't generate includes" checkbox is to avoid problem with LDD exclusive use of db.lif file. When model is converted for the first time, it is necessary to generate missing includes (if any). When the same model needs some small adjustments like change of colors or camera view, it is annoying to open/close LDD over and over to change, convert, try in POV-Ray. So it is better to check it, access to db.lif will not be necessary and LDD can be running during the conversion,

Then thats not what i did to hit my sweet spot...

[bbqqq]

Custom decors divide idea: 11 custom decors and 4 post process decors. (Not sure the wings color should be white or grey?)


LEGO 4000012 piper airplane by Nachapon S., on Flickr

Edited by bbqqq, 12 September 2012 - 04:31 PM.

[hrontos]

bbqqq, on 12 September 2012 - 03:54 PM, said:

Custom decors divide idea: 11 custom decors and 4 post process decors. (Not sure the wings color should be white or grey?)

Image

At first sight I did not recognized what is the "idea" but it is very nice trick to make larger custom decoration by splitting into many smaller and applying to multiple tiles.

[bbqqq]

hrontos, on 12 September 2012 - 05:37 PM, said:

At first sight I did not recognized what is the "idea" but it is very nice trick to make larger custom decoration by splitting into many smaller and applying to multiple tiles.

Thanks. Could you please add an option check box to apply all same LDD decors to same custom decors in next version? (only have to select the custom decor once when you have hundred of same kind of replacements.)


In some case, post process focal blur is faster, and can adjust strength of blur in real-time. Quick render the blur mask by delete non-blur parts, disable all light, LDD geometry, +ua.

Post-process focal blur mask by Nachapon S., on Flickr

Edited by bbqqq, 15 September 2012 - 01:46 AM.

[hrontos]

bbqqq, on 14 September 2012 - 04:38 PM, said:

Thanks. Could you please add an option check box to apply all same LDD decors to same custom decors in next version? (only have to select the custom decor once when you have hundred of same kind of replacements.)

Actually it behaves that way already now. It is not possible to replace one LDD decoration by multiple custom decorations. It is really picture for picture replacement. This means, that if one decoration is applied to multiple bricks, it is enough to specify new decoration picture once and it will be used on all of those bricks.

bbqqq, on 14 September 2012 - 04:38 PM, said:

In some case, post process focal blur is faster, and can adjust strength of blur in real-time. Quick render the blur mask by delete non-blur parts, disable all light, LDD geometry, +ua.
Image

Yes, I have read some articles on this kind of focal blur comparing advantages of both aproaches. It works well when there is one sharp area in the center of the picture. Problem may appear, when there are more objects, that should be sharp and each should be surrounded by blurred area. It requires bit more manual work to get it realistic when doing the post processing. But it is probably still faster than to calculate it using focal blur. Imagine a view from the clown position in your picture. Clown sees 2 or 3 horses at once and they should be sharp. People in the audience visible "through" the horses is further and should be less sharp.

[bbqqq]

hrontos, on 14 September 2012 - 08:28 PM, said:

Actually it behaves that way already now. It is not possible to replace one LDD decoration by multiple custom decorations. It is really picture for picture replacement. This means, that if one decoration is applied to multiple bricks, it is enough to specify new decoration picture once and it will be used on all of those bricks.

Yes, I have read some articles on this kind of focal blur comparing advantages of both aproaches. It works well when there is one sharp area in the center of the picture. Problem may appear, when there are more objects, that should be sharp and each should be surrounded by blurred area. It requires bit more manual work to get it realistic when doing the post processing. But it is probably still faster than to calculate it using focal blur. Imagine a view from the clown position in your picture. Clown sees 2 or 3 horses at once and they should be sharp. People in the audience visible "through" the horses is further and should be less sharp.

Maybe I did something wrong so I though it didn't replace all the same decors.

I added two more focal blur circus pics above (rendered one time only). I agree with you that the blur trick requires more manual work and may not realistic. I will try to divide more different layers of blur by LDD group function on top view. Hope it will looks more realistic.

I tried to find a way to speed up trans parts rendering without losing quality too much.
Speed up trans parts render trick,Total rendering time 28 min 1600x1200 by AMD Phenom 9600 @2.3 ghz
(Total rendering time 10 min for 800x600)
The final pic (#1) is result of 5 min quick post-process of 2+3+4.(1600x1200 no AA, all)
2= all parts/ ldd geometry/ 11 min
3= trans parts deleted/ all bevels/ 16 min
4= trans parts chang to black/ no light/ 1 min (use as selection mask)
Is there a way to render trans and non-trans with different quality automatically to speed up the rendering?

Speed up trans parts render trick_Total render time 28 min 1600x1200 by AMD Phenom 9600 @2.3 ghz by Nachapon S., on Flickr

Edited by bbqqq, 16 September 2012 - 07:14 PM.

[hrontos]

bbqqq, on 16 September 2012 - 05:22 PM, said:

Is there a way to render trans and non-trans with different quality automatically to speed up the rendering?

You mean to add a checkbox to LDD2POVray GUI which will disable beveling of transparent parts when checked? Just like you did - you rendered transparent parts without bevels. It should be easily possible without any significant effort.

[bbqqq]

hrontos, on 16 September 2012 - 06:54 PM, said:

You mean to add a checkbox to LDD2POVray GUI which will disable beveling of transparent parts when checked? Just like you did - you rendered transparent parts without bevels. It should be easily possible without any significant effort.

Maybe you have to see the full size pic (1600x1200) on flickr to see the different.The final pic is trans parts without bevels but others parts full bevels with Lego logo.

Edited by bbqqq, 16 September 2012 - 07:25 PM.

[hrontos]

bbqqq, on 16 September 2012 - 07:13 PM, said:

Maybe you have to see the full size pic (1600x1200) on flickr to see the different.The final pic is trans parts without bevels but others parts full bevels with Lego logo.

Yes, I think we speak both about the same - when this option will be selected transparent parts will not be beveled and will be rendered only using LDD geometry (but with LEGO text when level of detail with LEGO text is selected).

[bbqqq]

hrontos, on 16 September 2012 - 07:32 PM, said:

Yes, I think we speak both about the same - when this option will be selected transparent parts will not be beveled and will be rendered only using LDD geometry (but with LEGO text when level of detail with LEGO text is selected).

Are there any features of pov-ray or ldd2povray to allow: trans parts=minimum quality while non-trans parts=maximum quality by render just one time only?

Edited by bbqqq, 16 September 2012 - 07:43 PM.

[hrontos]

bbqqq, on 16 September 2012 - 07:41 PM, said:

Are there any features of pov-ray or ldd2povray to allow: trans parts=minimum quality while non-trans parts=maximum quality by render just one time only?

No, not at this moment. Transparent and non-transparent parts differ only in material definition and transparent parts are using merge and non-transparent are using union operation when enhancing geometry. All other handling and scene processing is the same.

[bbqqq]

hrontos, on 16 September 2012 - 08:39 PM, said:

No, not at this moment. Transparent and non-transparent parts differ only in material definition and transparent parts are using merge and non-transparent are using union operation when enhancing geometry. All other handling and scene processing is the same.

Then the trick just do what software can not do yet.  
I mean is it possible to add a checkbox to LDD2POVray GUI which will disable "ONLY" beveling of transparent parts when checked?
1600 x 1200 rendering time 40 minutes only. The idea is to minimal only the quality of trans parts is not very noticeable, but can save a lot of rendering time.
My link

jedi bus by Nachapon S., on Flickr

Edited by bbqqq, 17 September 2012 - 12:27 AM.

[hrontos]

bbqqq, on 16 September 2012 - 11:44 PM, said:

I mean is it possible to add a checkbox to LDD2POVray GUI which will disable "ONLY" beveling of transparent parts when checked?

Yes, as I said, this can be done without significant effort.