TurboFLOORPLAN Public Forum

General Category - V.16 and Earlier => V12 - V16 Looking for Help? => Topic started by: dBminus on July 29, 2011, 11:13:30 AM

Title: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on July 29, 2011, 11:13:30 AM
Hi All,

I'm an amature TFP user.  We designed our retirement home with it and I'm always amazed at what it can do.  I'm at the point where we're trying to determine our recessed can lighting in a room with high vaulted ceilings and dormers that let light in.  Now I'm experimenting with the advanced render schemes and interior lighting.  I'd have to say that the rendering outcomes have got me confused.

Here is what I think I've learned from reading this forum:
V15 indoor lighting will turn off if sunlight comes in a window  (previous versions one had to scale down indoor lighting, now its the opposite?)
Using different wattages or light types in different fixtures causes unpreditable light rendering
Use radiosity setting of 3 or higher (using 4 - with anti-aliasing of 4)

For my model, I turned off daylight, set time of day to night, used nighttime sky graphics....
I inserted a bunch of recessed cans in the vaulted ceiling.  Got the 'halo' light bloom on the ceiling.  Read where Allan suggested to lower the 'bulb' (red magic light block viewable in the 'lights' tab of the interior light property dialog).  That helps...  

I ran an experiment that has me confused.  I created a simple house model with an interior 6' cube made of drywall - no windows.  I put a recessed can in the ceiling and it rendered fine with no halo.  Being very simple, the convergence factor went from 1 to .1 in much less than the 1000 default steps allowed.  In the big model, the convergence had only reached .45 after 1000 steps.  Increasing the steps to 10000, I watched the render until step ~ 6000 where convergence was .2  The halos were still there.  I don't know why my test case works fine but the complex model acts different.  

Seems like as I add more lights, the 'auto iris' of the camera keeps the exposure constant.  I don't have a lot of confidnce that I need 10 recessed cans or 50.  Am I expecting too much science from this process?  Rendering an indoor scene seems like an art form and maybe thats all I should expect.... just tweek until its a pretty picture, but don't add up the types and wattages of the lights to make it that way...

Any insight you render experts have would be appreciated.  I've learned a lot of behaviors of various 'bulbs' in my 6' box.  Did you know putting an 'up light flood' bulb in a recessed light makes a spot light on the ceiling :)   Any clues as to how to turn the red cube around?  Turning the fixture doesn't effect the bulb, apparently....

Well, Thanks in advance... and btw, the scroll box for typing a long message seems to hide the bottom text between keystrokes... at least on my machine...  Guess I've typed too long a message...
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Jack Zimmer on July 29, 2011, 01:25:00 PM
Hi Dave,
Lighting is an art form , or should I say rendering in general which takes into account both light and materials and camera angle. In addition I am not sure any program is accurate enough to actually determine light types / location and wattage.

You can get a nice rendering for viewing a scene but the rules you mentioned help, like increasing wattage with Sun on.

Because of the variables it is hard to comment without seeing and discussing your layout and materials. I in fact plan on a 'book' series of videos on the subject. But all that being said, I do have a method I have developed that is pretty much fool proof and gets good images. But even that has it own set of rules that while simple are too involved to get into here.

Jack
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on July 29, 2011, 05:33:47 PM
Hi Jack,

Thanks for your insights.  I'm still unable to get the same rendered results using the recessed cans in a test 6' cube verses my house model.  In the test cube, I can put the can inside a sealed 'tube' and no light comes out.  In the complex model, light escapes the sealed 'tube' and causes a halo on the ceiling.  The two models just don't render the recessed can lights the same.  In the simple 'box' house, the recessed can works fine and it seems to behave correctly if I contain it inside a 'metal tube'.  Not so in the complex house.  The 'polygon' count for the simple box is always in the hundreds.  For the house, its usually around 99,000.  I guess thats a much tougher solution.

Edit update:  I ran a file:repair project and using the 'stock' 4" recessed cans in the complex model ceiling, they all rendered correctly (no halo on the ceiling!)  So I'm guessing that somewhere in my edits, I created a bunch of junk that the repair fixed.  The lights seem to be working fine now.... but I guess I should'nt count my watts as a basis for the real house lighting.... :)

Amazing what this program can do :)

Thanks for your help,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on July 30, 2011, 08:50:34 AM
Hi All,
I discovered that my recessed can lights can 'break' based on whether or not I turn on the dining room light, or if it uses a recessed type bulb, or a chandelier type bulb.  Strange.  Do I have to model each light fixture with a recessed bulb for the rendering to work correctly?

Here are a few renders.  They are all adv renders setting 5 with anti-alias at 4.  They all run to the default # of steps (1500), but the convergence varies between .35 and .45 (does that mean that ~40% of the light rays are unaccounted for when it hits the step # stop?)

The pics are from a family room towards a free standing kitchen.  At night, no daylight.  #1 is with dining room light off (dining room is opening to right left of kitchen)   #2 is with dining room light on - set with 4 75w recessed type incand. type bulbs.  #3 is with dining room fixture with original 100w chandelier type bulb.  With #1, it looks like the left table lamp is hot along with the right ceiling recess can.  With the dining room 4 75W incand on, pic #2 looks like I dimmed the lights in the family room and the lower right kitchen cab light went out (but it was on).  Pic #3, just changing the 4 75w bulbs to 1 100w chandelier incand broke the ceiling recessed cans.  The kitchen cab lower right is still out.  Theres a strange light on the outside of the windows to the right lighting up the siding, but there is no light outside.  All 3 pics have exactly the same lighting in the family room and kitchen - which is what is in the viewport.  So, I can get 3 quite different renders of the same space by changing the dining room light properties...  oh well.  Its an art form...
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on July 31, 2011, 04:33:25 AM
Hi Dave,

Welcome to the forum and welcome to the joys of rendering.

As was mentioned before it is a bit of an art form but with a little experimenting some excellent results are possible. You appear to have done an excellent job with the model and even each of the renderings are good in their own way.

I discovered that my recessed can lights can 'break' based on whether or not I turn on the dining room light, or if it uses a recessed type bulb, or a chandelier type bulb.
Yes, as you have found, one light can 'break' another.  I don't quite understand the logic of that but it does happen. It is obviously baffling the programmers a but too!

Do I have to model each light fixture with a recessed bulb for the rendering to work correctly?
I have found that I have the best and most consistant results if I only use the same light bulb and the same wattage in all fittings. This way usually they all stay on.  If I need a down light I just lower the bulb to give that effect and if I need a light dimmer than the others I just colour the bulb, usually with a shade of grey but occasionally with a colour for special effect.  I have got away with different wattages and bulb types but with a lot of experimenting and often I do not have the time - so I make them all the same.

All the best with your renderings. When you get close to what you want it would be nice to see some examples in the Gallery!

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on July 31, 2011, 09:31:20 AM
Hi Allan,

Thanks for the welcome, and thanks for your help.  I don't think I would have gotten this far without reading many of your posts and your tips and tricks with easy to follow steps!

I took your advise and changed all the bulbs to one type.  That solved the dining room fixture breaking the recessed cans problem.  Unfortunately, I chose the recessed can bulb as my standard, then ran into problems scattering the light on the ceiling with other hanging fixtures.  I ended up sneaking a few pendant type bulbs in the mix, and I'd keep breaking the can lights as I added more and more fixtures  :(   I think I'll go back and use a better general purpose bulb and lower them from the cans per your suggestion.

By the way, do you know of any way to rotate a bulb?   I've tried rotating the fixture, but the bulb doesn't seem to follow.  Virtual reality has its ways :)

Thanks again, and I'll post some renders to to gallery once I'm finished.  The forum sure has a nice collection.  I just hope mine turn out half as good as the one's I've seen you do :)

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on July 31, 2011, 10:22:19 AM
Hi Dave,

I'm glad that solved the problem - I have found using all the same bulb is the quickest solution.

By the way, do you know of any way to rotate a bulb?   I've tried rotating the fixture, but the bulb doesn't seem to follow.

By "rotate the bulb" I assume you mean where you have multiple ones in one fitting?  The only way to do it is by changing the XYZ settings on the bulb using positive and negative figures to move the bulb backwards and forwards, left or right. The distances entered are in relation to the centreline of the fitting itself (which basically is irrelevent - it is just for appearances).  You can actully use the bulbs without a fitting (almost) - just set it to extra tiny or make it all transparent.  You can then put some general light where you want it for effect or to brighten dark corners without seeing the fitting.

Looking forward to seeing your final renders!

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Doug.S on July 31, 2011, 07:21:50 PM
While I cannot be completely confident, your walls may contribute to your lighting issues...maybe.

What I could see in the rendered images is uneven light on the walls.....with exact lines of change...which seems like some walls are made of pieces and they are not aligned exactly.

Look near can lights at far left and far right and see a light brightness change...light to dark...not from the bulb light but a vertical line on the walls.

Sometimes a "file fix/auto repair" will fix walls but not always. So the manual solution is to zoom in as close as you can to every wall joint and check that the walls are aligned exactly to one another.

If you cannot see what I'm talking about, let me know and I'll mark up a rendered image. If you attach a .bld file I can mark and render to show it even more dramatically and also see how you constructed the various walls.  The "best" is to use only continuous walls and cut out openings rather than build openings by using multiple wall pieces.

Doug


PS  Also since your ceilings seem to be canted (not horizontal level) the light fixtures need to be carefully aligned with the ceiling to prevent light leakage....as they may not align or be exactly flush as is typically automatic with a conventional flat ceiling at 8' and auto installed lights for height to match ceiling height.

Again, zoom in as much as possible at the light fixture to ceiling interface to verify the correct alignment to each other as well as the light bulb red cube to the fixture.
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on July 31, 2011, 08:24:08 PM
Hi All,
Allan, by rotate the bulb, I was meaning can one take a 'up light flood' bulb and rotate it 180 deg and make it a 'down light flood'.  In my 6' test cube, I put an up light flood.  Saw the spot light on the ceiling.  Then I rotated the flood 180deg so it was pointing to the floor, and the bulb still cast the same spot on the ceiing :)  Seems like the photons from these bulbs go right through the fixtures as well.   Just curious if the actual 'bulb' could be rotated in the X,Y,Z axis.

Hi Doug,  You've got a good eye.  The uneven light on the walls bugged me too.  I double checked the 'paint' for consistency.  Checked the wall for continuity and as best I can tell, it is one piece.  The shiny strip seems to correspond exactly with wall protrusions that go into other rooms.   The shiny break on the right side is where two walls meet from from different locations, but I don't know why one would reflect more than the other.  I had tried to use one wall originally, but I could not get my opening to be flush on the farthest wall, and I always had issues selecting the dormer windows in plan view, so I put the dormer wall/window on another location.   I attached just the wall view of the left side fyi.  If you'd want to look at the whole model, you're welcome to... I'd just warn you first that its ~ 13mb.  Its got a few out buildings and other stuff helping us visualize the final landscape...  Here is a pic standing back from the house...  Thanks for your help and suggestions.
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 01, 2011, 04:29:18 AM
Hi Dave and Doug,

I noticed the different shades of render on the wall and assumed (correctly it seems) that there are walls joining this wall on the other side of it. Unfortunately this does happen, even if it starts as a continuous wall - as soon as walls join to it on the other side you can get a different effect. (see my test image below - purposely set up the same way). The only way I have found to overcome this is to temporarily drag the attached walls away form the wall in question (by just a mm or fraction of an inch). As long as long as the other walls do not actually attach, the wall on the opposite side will be OK.  Still looking for a cause and fix on this one.

Re rotating the direction of a light:  If you want to use the same bulb throughout the model to overcome render issues, all you need to do is move the bulb up or down to get the effect you want from each light.  

If you want some general light on the ceiling (but not a bright flare) then place the bulb say a third of the way between floor and ceilling. To make it a down light add another bulb and move the second bulb closer to the floor to give a 'flare' on the floor or table.

For a twin flood light (as for example used in sensor security lights) use two bulbs, move one to the left and the other to the right so it mimics the effect of the twin flood lights onto adjacent walls for example (lower them if you want to give the twin light effect on the ground).

If you use these ideas any light fitting can be an up, down or sideways light and all use the same bulb.

The outside view looks good - it would be nice to see it rendered! Basic render is usually fine for oustide, especially of 'distant' views. If close up to verandahs etc then Advanced may be better.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 01, 2011, 08:59:43 AM
Hi Allan, Doug,
Thanks for the wall shine explaination.  I'll try pulling the other walls back a tad so the final renders don't have that strip down the wall.

Sounds like I need a redo of my lighting fixtures using a general purpose bulb.  I'll get this right sooner or later :)   With all recessed can bulbs, the lighting never seemed to break, but I couldn't get general light on the ceiling either...

TFP has been very helpful working with the builder.  I keep updating the model to reflect the actual builder drawings and then I show him images of the deck or interior where I have questions.  His drawing is a 2D acad model and sometimes its hard for me or my wife to visualize what he's thinking.  TFP is a great go-between for 'getting us on the same page'.

Once again, Thanks much for the advise and tips.  The project is shaping up!  (must say the virtual house is much farther along than the real one though...)

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 01, 2011, 11:00:43 AM
Hi Dave,

Just a thought if you were going to change all your lights: You only need to change one recessed light for example and save it to the Catalogue (or Add/Edit one in the Catalogue). You can then right click one in the model and choose Select All Similar, then right click and choose Replace. This will replace all the ones in the model with the new one with adjsted bulbs .

Yes TFP is great for seeing how things will look and I agree, virtaul homes are much quicker!!

Allan

Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 01, 2011, 01:43:17 PM
Hi Allan,
I pulled the joining walls back from the left side wall and the light stripe is gone!  I was able to nudge those walls back to be right on the wall outline, but since they're no longer 'joined' the render came out fine.  Great tip,  Thanks.

Here is the adv render of that.  I still need to redo the lights.  I'll try your suggestion.  Should save a bunch of time.

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 02, 2011, 12:16:11 AM
Hi Dave,

The render looks really good and the model is excellent  You have obviously put a lot of work into it as it is quite a complex design.  Congratulations!!

Looking forwar to seeing further renders later!

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Doug.S on August 02, 2011, 10:33:31 AM
Another tip for you....to save significant render time and sometimes "fixes" unknown problems...or allows you to render at higher qualities and the render time is not so bad.

Save the most current design as a "master" file for safe keeping.

Make 2 or more copies of it....for different rendering purposes...such as:

- use one file for outdoor renders....and delete anything indoors that would not be visible or contribute to the render lighting...only render outdoor views from this file

- similarly, for indoor views...delete all exterior extra bldgs and gut rooms not seen/used....should reduce render times and sometimes yields better renders as TFP no longer calc the extra objects/lights/etc.

The only down side other than extra files/disk space is if you make any important re-designs, you may need to repeat or update several files...maybe!
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 02, 2011, 03:46:30 PM
Hi All,

Allan, Thanks for the complements.   Its been a process.... and like I've said, your tips and tricks have helped along the way.   I've redone the lights to generic bulbs and now I'm just fussing with multiple bulbs on the recessed cans that reflect near a wall to give a similar flare to the recessed can types....   A bit more model editing around the lot and some more exterior lights before I go for final renders....

Doug,  Thanks for your tips on speeding up a render.  I will try that.  Currently, outdoor renders work OK on basic, but advanced seems to boog down into days instead of hours.  I wanted to get an indoor rendor during daylight, with the morning sun timed to light up the family room through the dormers.  I could get the basic render, but advanced render ran, just at a few steps an hour... so I abandoned it.  I'll try a gutted model for that an test it again...  Thanks,

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 02, 2011, 10:04:25 PM
Dave,

As Doug mentioned if you can reduce the objects in the view it will speed render times, sometime quite dramatically.  Outbuildings for example should be on a different location and you can then just turn them off in the View Filter if they are not in the sight of the camera.  Similarly you can turn off furniture and other objects (for all locations or just some of them) in the View Filter if they are not in the scene. For exterior renders just use the filter to turn off all Interior Lighting.  Filtering out objects like this for specific renders may eliminate the need to delete them.

I could get the basic render, but advanced render ran, just at a few steps an hour... so I abandoned it. 
This comment is a bit worrying - it should not take that long unless you have an ultra slow PC or there is a corruption somewhere in the model. Even some of my very large models render in less than an hour. I think Doug alluded to this earlier when we were discussing the render that was affected by the connecting walls. The connecting walls was the issue there but there could still be another issue (with a wall usually) somewhere else.  Have you used the File/Repair Project?  This will sometimes find a corruption and delete it (but not always unfortunately).  If it does not help examine all the wall joins, select them and look for stray bits sticking out or odd shaped segments inside the wall joins. Walls should connect with simple mitre joins - if there are lots of bits and pieces in the join area (must select the area to see them) then redo the join so it is clean.

Back to rendeRing:
Don't forget too that you only need to do the Advanced render once!  It renders the whole model, so all you have to do then is to reposition the camera in the 'rendered' view window (select Modify Viewpoint) and you can then walk around the model and just click the Raytrace button again to take a new 'photo'. It will then render ultra quick, just adding reflections to the radiosity scene. You can then just reposition the camera again and do another Raytrace and so on. 

You can also click the Radiosity check box and move around through the model with all the lights on, to look for the best spots to do the next view.  You can also change camera position using the drop down list (at the top) to go to previously set viewpoints. Unfortunately you cannot save the full Radiosity model which is a shame as it is ideal for viewing the model in more realistic light. Once you close the rendered view window you lose the radiosity model and you will have to do a full render again.

REMINDER:  When doing the Modify Viewpoint and Ratrace again don't forget to click options and change the file name or it will overwrite the previously saved image!

Rendering with daylight can be an issue as mentioned earlier, but you can get a daylight effect by positioning a light outside your windows to represent the sun shining in. You can sometimes get away with using Exterior Lights (eg Street Lights) without it affecting your interior lighting. Also the background image will always show, even if you do not have daylight enabled, so if windows are visible in the view the outside scene will still render even though daylight is off. You do not need to set it to a night scene.

Hope those thoughts help.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 03, 2011, 12:34:18 PM
Hi Allan, Doug
Thanks for all your tips.  I'm still back tryng to get my recessed can down light to look the same as a recessed can bulb :)  In the process, I built a simple vaulted ceiling and tried to comparion test... but my ceiling went black when rendered.  After soom forum search, I moved my indoor camera completely indoors  (had the lense inside, but the body outside - enough to break the render...)  Attached a pic fyi.  Also attached my generic bulb down light (combo of several bulbs with gray glass) thanks to Allans tips.

WRT Advanced indoor renders with sunlight...  turns out I have the outbuildings on seperate locations and had them off as well as everything outside except terrain.  Sounds like I probably won't speed up the render with a down-sized file... bummer.  Over the months, I have had multiple moves and nudges of walls and partial tear downs as I tried to stay in step with the builder.  I did see those wall fragements you mention.  Tried to delete the slivers and most wall joints would meld back together.  I do have some that just won't though.  I also know that some members interfere with walls or other components.  Is it possible that I have some bad geometry on other locations, and even though they're off, they mess up the active view?

Aside from the very sluggish sunlight on Adv renders, I'm pleased with the night light renderings.  The universal bulb fix saved me.   (Got any suggestions on how to apply this to an undercounter florescent bulb???) :)  I'll keep looking for broken walls, etc.

Thanks again,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 03, 2011, 09:21:48 PM
Hi Dave,

A coulple of clarifications that may help:

After soom forum search, I moved my indoor camera completely indoors  (had the lense inside, but the body outside - enough to break the render...)
The technique of rendering with the camera outside works only if the camera is fully outside - looking at the outside wall for example.  After the render has finished and while you still have the rendered scene in the preview window, click the Modify Viewpoint, move the camera inside, and then just press the Raytrace button, which will re-render the new inside angle in just a few seconds. This was the process I referred to in the last post but the camera must be completely outside for the initial full radiosity render.

Advanced indoor renders with sunlight...  turns out I have the outbuildings on seperate locations and had them off as well as everything outside except terrain.  Sounds like I probably won't speed up the render with a down-sized file... bummer.  Over the months, I have had multiple moves and nudges of walls and partial tear downs as I tried to stay in step with the builder.  I did see those wall fragements you mention.  Tried to delete the slivers and most wall joints would meld back together.  I do have some that just won't though.  I also know that some members interfere with walls or other components.  Is it possible that I have some bad geometry on other locations, and even though they're off, they mess up the active view?
If inside renders are working OK timewise, then it may be that even though you have bad wall joins it may not be these that are causing the slowdown (although they should be fixed becasue they potentially will). If there are corruptions in the outbuildings, turning them off should stop those corruptions interfering. Wall joins should correct themselves if they are remade. If several walls join at one corner, delete one and redo it, if possible by extending one of the existing wall - this is the best way to fix them. Members joining into a wall should not cause a problem.  If a corruption is casuing the problem then the best way is to use the View Filter and turn everything off, then turn one thing on (eg walls) and render. If it is slow then then problem is in the walls. (If there are walls on different locations then turn those locations off to find the offending one). If however all the walls render quick, then leave them on and turn roofs on also and render again. If it is slow the problem is in the roof. If quick, turn something else on and so on until you find the group of objects that causes the slowdown. Now start deleting them one by one (starting with the one that is most unusual, large or in some way suspect) and render again.  By a process of elimination you can identify the problem element.

Aside from the very sluggish sunlight on Adv renders, I'm pleased with the night light renderings.  The universal bulb fix saved me.   (Got any suggestions on how to apply this to an undercounter florescent bulb???) :)
This is no different to any other light - it does not have to have a flourescent buld - use the same one you are using elsewhere. If you want it to be 'daylight' colour and are using an incandescent bulb as your universal one, then just change the bulb colour to a shade of blue istead of the usual white, varying the darkness of the blue to get the effect you want. If the light needs to spread more along the tube to look right, then just add a few more bulbs, spacing them along the tube and darkening the bulb colour if necessary if there is too much light.

If you still have problems and would like me to take a look at the model, email it to me (alches [AT] iinet [DOT] net [DOT] au). Replace the AT and DOTs with the appropriate symbols of course - trying to avoid more spam - there is enough now. We ban about 50 or more "users" a day from the forum, so they are certainly scanning it!! >:(

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 04, 2011, 09:12:38 AM
Hi Allan,
Well, I think I found the culpret... the terrain.  I was able to get indoor radiosity Q4 - steps 2000 to complete fairly well as I turned on the inside 1st floor filters.  With bare interior and indoor lights (which promptly went off) it took 382 sec radiosity.  Adding the furniture and kitchen cabs, appliancies, etc took 712s for radiosity.  I added the basement walls which I thought would break it, but that only increased the time to 816 sec.  Turned off the basement walls and turned on the terrain (no fills, paths, etc) and I quit the radiosity at step at 155 of 2000 at 1005 sec...

The terrain is 500' by 500' and has a lot of plateaus to get close to the topology of the land.  I attached a 'big' view of the lot fyi.  So, with the terrain as the culpret, I guess it's pretty stupid having it on doing an inside render anyways :)   If you're still interested in seeing the file, let me know.   I don't know what to do to speed up the terrain... haven't tried an external adv render yet...

Thanks for all the help,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 04, 2011, 09:55:12 AM
Hi Dave,

Well that is a bit of a relief really. If the Terrain is causing a slow Advanced render then that is not a problem, as Basic render works better for outdoor views anyway. Radiosity (in Advanced render) relies on light bounced off walls, floors and ceilings, so the only benefit outside with Advanced render is under verandahs and probably that would not be noticeable. Basic should render the scene very quickly and do a good job of it. Only if I was creating an image actually on the verandah itself, would I consider using Advanced render.

However I notice you have a lot of trees etc on the terrain. Did you have those on during your test?  If you did, then they are what would have caused the time problem. Each tree is an image that is dragged from your hard drive, so they are more likely to take time than the plateaus, especially trying to calculate all the shadows they create.

There is probably not much point sending the model at this stage unless you have further problems or if you need help with the walls issues. Basically they are just a matter of patiently working through each wall intersection to clean them up. Unfortunately if you do a lot of reworking it can leave a few problems, unless you re-do them properly at the time.

Try a Basic render outdoors and see how it goes.

Allan



Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 04, 2011, 06:23:55 PM
Hi Allan,
When the terrain bogged down the adv render, I only had the terrain on, no trees or other landscape items.

Well, I tried a basic and adv render of the front.  You're right, a huge time difference.  I was able to get some good renders inside also with just sunlight.... so I'm pretty happy.  I attached a comparison of the basic and adv of the front here.  Seems like the advanced one really pops out the colors and shadows.  (I just need a faster machine :)

Got to move the garage to align with the builders actual location.. and fix the elevation.  Its up so high, the car has to jump into the garage door :)

Thanks for the help, and I'll post some pics in the gallery after a few more fix-ups.

Thanks,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Doug.S on August 04, 2011, 06:28:30 PM
1. I would agree with Allan; probably not the terrain but rather so many trees causing a very slow render...especially if they are 3D trees rather than billboard type trees.....would also make the file size large.

2. There may be difference in render time between turning objects "off" vs. deleting them from a scene....only experimenting will tell for sure. It all depends on how the s/w code is written and data handled.

3. Another big suggestion....don't always assume you must render the results exactly as you want it to look. Sometimes that is a very difficult and time consuming effort tho it leads to understanding the program better.
Often the fast/easy solution is to use a post render process....such as using Photoshop (or similar program) to "adjust" the near final render to tweak to final desired image. This technique also lends itself to compositing...i.e. render various pieces or results and blend/combine/adjust using layers and masks...often much more effective, faster/easier, and have more control over the results.

4. The biggest advantage of advanced render is the shadows...instead of basic render yielding hard edge shadows, advanced render makes blended/gradual shadows and makes shadow areas lit much more realistically.
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Doug.S on August 04, 2011, 06:35:01 PM
For these renders, I noticed the tree behind garage and directly above the car....looks odd, stretched or something. I can see it is a different tree and that is good, but something seems wrong with it??


Hi Allan,
When the terrain bogged down the adv render, I only had the terrain on, no trees or other landscape items.

Well, I tried a basic and adv render of the front.  You're right, a huge time difference.  I was able to get some good renders inside also with just sunlight.... so I'm pretty happy.  I attached a comparison of the basic and adv of the front here.  Seems like the advanced one really pops out the colors and shadows.  (I just need a faster machine :)

Got to move the garage to align with the builders actual location.. and fix the elevation.  Its up so high, the car has to jump into the garage door :)

Thanks for the help, and I'll post some pics in the gallery after a few more fix-ups.

Thanks,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Doug.S on August 04, 2011, 06:42:42 PM
You also might note a slight difference in sun/shadow position between basic and advanced renders....it is a bug in the program.

So if you need "exact" shadows....you need to play a bit with sun time-of-day settings....usually the difference is inconsequential....unless you are making a detailed shadows study or want to match a model with a real photographed scene.
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 04, 2011, 08:20:19 PM
Hi Dave,

As Doug said there is a bug in the time between Basic and Advanced and your two images show it up clearly. The shadows in Basic are much longer, lose detail and the whole scene is duller as if it was a hour or two later. Similarly you will notice that the trees in Advances appear overexposed which I notice often happens with Advanced when used for outside renders. Sometimes darkening the who render can help. Setting daylight to Partly Cloudy can lighten the shadows in Advanced but does not affect the plants - they remain too bright. Note my comparison images, Advanced, Cloudy Advanced, Basic and Basic with a simple enhanced contrast (30%).  Advanced took 3 min 23 seconds, while Basic took 20 seconds.  

In the Basic you will notice the shadows are lighter, it has ignored the transparency of the roof over the left hand bay window, and it has rendered the lace curtains. Advanced has ignored the lace curtains and rendered the drapes behind them, shadows are darker and graduated more naturally and it has rendered the transparency of the roof correctly. Interestingly the Advanced ones have an issue with the facia on the left hand gable of the roof (?).

Try rendering yours in Basic again an hour or two earlier and try enhancing the contrast. If Advanced render is too long then Basic may suffice.

Actually if you like you could email it to me - I would like to experiment a little with it and see if I can see why the terrain is causing a render time problem.

Doug,
For these renders, I noticed the tree behind garage and directly above the car....looks odd, stretched or something. I can see it is a different tree and that is good, but something seems wrong with it??
I think it is just that particular tree - there is one like that - it has long 'droopy' leaves.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 05, 2011, 07:42:04 AM
Hi Doug, Allan,
Thanks so much for the feedback.  I'm learning a bunch here.

This may seem odd, but my original indoor adv render with clear skys really slowed by just adding the terrain... no plants (of which there is a forest of them).  I don't have an adv render of the outside to compare adding just the trees.  Here is a pic of the house w/terrain only - 2400 sec radiosity.   The pic of the house and garage took 3087 sec radiosity (had garage displayed, plants, fills, fence).

Doug, I'll try some 'post processing' effects.  I have an old coral draw program, but I'm not as advanced as you or Allan.  You've done some real impressive stuff.

Allan,  I'll go ahead and send you the file.  Don't laugh too loud when you see some of the 'features' :)  I hope it doesn't take too much of your time.
Thanks,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 05, 2011, 09:15:58 AM
Hi Dave,

It will be interesting to see the file as the terrain does not look that complicated and it is really just a single surface so I would not have thought it would affect the render time very much at all.

In fact I did some tests on a blank file.

First I Advance rendered (Q5) the Terrain with no hills or valleys and a Mesh of 1000mm (39") as a benchmark - it took 51 seconds.
Second I added lots of hills and valleys, plateaus and trenches and it added only 47 seconds (to a total of 98).
Third I changed the Terrain Mesh to 250mm (about 8") and this added 85 seconds to the flat terrain or a total of 136 seconds.

The third image is the same terrian but just a different viewpoint. Changing the viewpoint made absolutely not difference to the render time.  The only thing that made a difference was changing to a smaller mesh. This results in a much smoother terrain at a cost of 85 seconds so worth it in my opinion.  What it does seem to indicate is that the plateaus in your terrain should not be extending the render time very much unless you have a very tiny mesh but I suspect that is not the case.

Anyway it will be interesting to test out your actual model and see if I can find anything odd.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 05, 2011, 02:48:00 PM
Hi Allan,

Your terrain torture test looks like a tough golf course!  I hope my model reveals its secrets without too much trouble.
Thanks,

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 06, 2011, 01:19:19 AM
Hi dave,

Well the Terrain only rendered in 70 seconds so it is not the problem.

With the trees 113 seconds so they are not the problem either.

Garage Location added - now took 213 Seconds - checking further.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 06, 2011, 10:00:03 AM
Hi Allan,
I got similar times as you with the adv render and just terrain, or terrain & plants.  I went back and started with just the house first, basement and ground and dormer walls, footers, members, columns and decks and roof and HVAC (for chimney) to get a fairly complete rear view of the house.  Changed the background to the lake view (much better choice than the sad water fill paint I had choosen :)

I ran 4 adv renders Q4 2000 stps - all at 8:30am july 1, 11 at its gps location.  Attached are my results.  Very odd.  The house rendered well, floating on the water.  Adding just the terrain then slowed it down big time, even though the terrain by itself is good...  Then added the plants.  No big time difference... so just for grins, I removed the terrain and rendered the house with just the trees... and that seemed to work fast as well.   I don't get it.    I would agree, the terrain seems relatively simple to render, but throw the house in first, and its a different story...
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 07, 2011, 09:55:32 AM
Hi Dave,

Sorry but I can't explain why rendering the terrain by itself is quick and rendering the walls by themselves is quick but rendering both together is very slow. Certain things do affect others (as you know by the lighting) but.....?

Anyway - back to your model.  

First, there is a problem wall in the Barn and I think you suspected as much from a comment earlier and it causes a total failure of the render. Note the image for the wall in question - when you select it look for the wall that creates a long spike (as on the image) when you select it.

There are a few wall issues in the house itself, but nothing major and none affect the render.

I have rendered with each of the groups of elements off and none seem to make a difference by themselves so it does not appear that there is a major problem with any of them. Interstingly in none of the renders did the photoboard people appear - only the shadow! The images were nota good angle for the car - obviously the model has no floor in it!

I think the reason it is taking so long is the total number of polygons involved (153,000+) and the very high number of steps selected, 2000. I am not sure if there is some reason you have slected that number but it renders fine on just 200 (Quality 1) which takes 19 minutes with everything turned on including the barn. Level 3 (500 steps) takes 30 minutes but visually there is virtually no difference.  Compare them and you will notice the only difference is a few more lights on in the 200 step image so it fact it could be judged as better than the 500 step image that took somewhat longer.

After sucessfully proving that the model was OK (after the barn wall deletion) I tried rendering with 2000 steps but terminated it. It took 1hour 15 min for just 150 steps of the 2000. Incidentally, you can pause the radiosity phase at any point then click the Raytrace button to complete the image. The Terrain is added in the Raytrace pahase.

From my previous test, Interior renders need Quality 3 (500 steps) to be sure all the lights are on, but exterior renders are fine at much lower step levels and extra rendering times are pointless as there is no decernable improvement in quality.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 07, 2011, 01:32:00 PM
Hi Allan,

Thanks very much for spending all that time with the model.  How did you ever figure out what wall was broken in the barn?  Now that I know where it is, I can select it and delete it, but when I draw a new section in it misbehaves again.  I used to have doors on that wall but somewhere along the line I missed something up.  I'll go through and rebuild it.  I think the barn was one of the first things I started on and my skills were not as 'refined'.

I guess I used 2000 steps when I was doing the inside renders with lights.  I could always stop it early, as you mentioned.  With the house and terrain visible for the adv render, it sure does start off slow.  After 1000 steps, it seems to speed to the finish, but as you say, there isn't much benefit.

Us users really appreciate the time and explainations you give to our projects.  Thanks for running a great forum.

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 11, 2011, 11:12:54 AM
Hi forum,
Been adding features to the model prior to the 'final render' and I may have created more issues wrt rendering.  After importing some custom light fixtures and fine tuning some features, when I open the render 3d realview, it's initial viewport is not like the interactive render view in the main program.  When I do a standard render, the sun angle and exposure looks broken.

I ran a few test cases, one with the complex model and just about everything on - where the basic render breaks, a 2nd case where most everything is off except the walls, roof and deck that are visible, along with the terrain and a forrest of trees.  The 3rd case has the same features turned on as case 2, but all the hidden features have been erased (that file is ~ 9mb, whereas the complex one is ~ 19mb).   Its interesting that the basic render appears to work correctly with case 2 and 3 (actually faster in case 3).  The adv render which seemed to work in all cases had similar times in case 2 and 3.  I'd use the adv render except the trees and some of the terrain look over exposed as mentioned in other posts.  Seeing how V16 has just been released, I wonder if its render engine has improved with the lighting issues....  I'll have to try it :)  Attached are my 3 cases fyi.
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 11, 2011, 02:16:56 PM
Well, figured out that the basic render was 'broken' by having the interior lights on.  Apparently, they made the bright spot in the house that created shadows radiating out... making the fence post cast a shadow in the wrong direction.  Also, after a repair project, the basic render viewpoint matched the interactive render view...

Downloaded and installed a trial of V16 and imported my project into it.  I rendered the same view and it had the same bright spot in the basic render until I turned off the interior lights (should have had that off anyways...)   I attached the render views with lights off.  Looks like the basic and advanced render the same as V15 with this outside view.  Haven't tested the interior nighttime ones yet.  I did notice that V16 likes to show the white PVC rail as terrain grass though.   Selected it and the properties look OK (white PVC), but once the property box is closed, the interactive render momentarily goes to the PVC, then returns to terrain green...  Don't know what thats about yet.  V16 looks like it has some nice new features :)
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 11, 2011, 06:48:23 PM
Hi Dave,

I am working off a high resolution wide screen laptop but your images are too wide to fit on the screen. Any chance of reducing the size of your images? I can scroll sideways to see the image width and read the text but the sidways scroll bar is at the bottom of the image, so you have to scroll down the page to get to it then back up again to read the rest and then repeat the process to read the next line. Generally it is best to keep image width to 1000 pixels or under and then they will fit most users screens.

If you reduce the size of the images first, (any photo editing software will do - even the built-in Microsoft Paint) you should be able to go to your post, click Modify, untick each of the present images (to say you do not want to use them) and add the new ones in the usual way and it will just replace them with the smaller images. This will make it much easier to analyse what you have done and see what is happening.

As far as I know there have been no changes to the rendering engine in V16 (or at least they have not told us, so it must be very minor) so things should render the same way as V15. I am curious about the railing colour change, although this may be just a PC display issue due to the large model size. The model size it a bit of a worry as typically 5 MB is regarded as large. 19MB is unusual and you do not appear to have enough in the model to make that size. Have you imported anything from other programs (3DS SKP etc) or any photoboards, images etc?

Yes, lights on with Basic render will cause light from the inside to be cast outside, overpowering the sun! Pretty bright lights - I am sure the power companies and street light manufacturers would like to know how it can be done!!!

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 12, 2011, 09:19:28 AM
Hi Allan,
I replaced the images with smaller pics per your instructions.

Large file size:  Well, you're right, since I sent you my project, I've added custom light fixtures, etc.  With the extrusions and features I put in the models, Tcad created a bunch of facets and when it was read into the element library, the number of vertices was much higher than the catalog models.  I tried doing the simplify verticies, but if I went below ~85% reduction, the program would hang (V15).  I've probably got some issues going on with the dxf imports.   Probably a topic for another thread...

Green fence:  I opened the file I sent you in V16 and noticed the following.  Fence in 3d interactive view initially renders white until I move the camera, then it repaints green.   If I turn off the terrain (all else on), it renders OK (white).  If I just turn on the fence and terrain (all else off), it renders green...  Don't know if this is somehow related to the earlier discussions about render times with/without terrain on in V15.  The fence continues to render white in V15, regardless of elements on/off... even works with the latest larger file (got to work on those custom lights, etc...)

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 12, 2011, 09:45:17 AM
Hi Dave,

The smaller images are MUCH better, real easy to see now without scrolling every which way!

DXF and other added files can cause the file size to increase dramatically and I suspect this could also be the problem with the green fence - it may be running out of memory and doesn't like it. Have you tried exporting them from TCad as 3DS? Usually that works better than DXF.

What is unusual about the lights?  Do they have to be so complex? If you want special shaped lights it may be better to do them in TFP itself so there wont be memory implications that you get with imported files.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 12, 2011, 02:23:37 PM
Hi Allan,
Glad the smaller images work.  Sorry about the big ones...

I tried exporting the models as 3ds out of tcad, but TFP didn't like them.  My tcad is 9.2 and pretty old.  That may be the problem.  DXF did work OK though.  It creates a lot of vertices in TFP.  I attached an example light.  The tcad model must export a lot of detail.  Haven't figured out how to simplify it yet.  Trying the simplify vertices in TFP works for small reductions in complexity, but when I try <85%, it hangs.  Don't know if that means I have issues in the dxf file or not.

Is this light fixture something I could create inside of TFP using lines and arcs?  Do I convert them to members of various profiles?

Regarding the green fence, I can certainly see being up against a memory constraint, but why would it work OK in V15, and not in V16?

As always, thanks for your help,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 12, 2011, 03:30:46 PM
Hi Allan,

WRT the green fence, I ran a little experiment.  Opened a new V16 file, added just grass terrain and a ranch style fence.  Interactive render looked fine.  Then I edited the fence properties and made a section white PVC material (like what I had in my big model).  That rendered green!   Tried painting it white, still rendered green.  Then I changed it to a wood material, and it rendered the wood material OK.   Then I edited an adjacent section to PVC and it rendered OK as PVC!   Thinking I was losing my mind, I re-edited the first fence back to PVC and then both edited sections rendered back to green!   You may want to try a fence and see if it behaves the same with you.  Of course, a white PVC fence could turn green after a few years of moss :)

I attached my green fence fyi...

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 13, 2011, 06:13:11 AM
Hi Dave,

You are not losing your mind!!   It seems to be a bug in the program that was not there in V15.  I have done tests with both and also with Envisioneer V6 and V7.

Envisioneer 6 and TFP 15 are both OK but Env 7 and TFP V16 both have the problem.  Thankfully the Paint colous work, so looks like you will have to paint the fence instead of covering it with PVC or Porcelain. Probably cheaper too! 

I have reported it so a patch will no doubt fix it in due course.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Kajvin - Robert W on August 13, 2011, 04:32:19 PM
Hi Allan and Dave!
If you us a white texture instead of a white paint the fence will be OK!  So it is a bug,  then you use paint it is not a texture it look like it is take from windows colours direct, not via a jpg- or bmp-image  of a white colour. P.V.C and even metal work as the paint is not a texture.

Hope you understand what I meant.  

Or whit other words TFP 16 and EV 7 use Windows colour palette for paint and it will not work because it is a RGB-colour without a texture, not a texture  you have to have a texture of paint. The a simply way a round this is use colour-texture from images.  

I have made one example image from left to right, texture, texture, yellow paint, texture  and yellow paint-texture.

/Robert

PS Allan! if I understand you right you have use paint on the wall in image with grass, that will work because the wall is a texture by it.
 self, maybe I'm out on deep water her, but that i think it is. DS
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 14, 2011, 07:04:51 AM
Hi Robert,

The colours in most of the Paint section of the Catalogue ("Ben Moore...") are textures (BMP files) and the Paint I used ("Simply White") is a texture and it worked on my example. The only ones of the Paints that are not textures are the ones in the section just headed Paint (see image). The others (eg Paint - Historic Colours, Pain - Off White Colours etc) are all textures not Solid Colours. The Colours section of the Catalogue won't work also, as they just use Solid Colour (Windows RGB) also as do the Finishes -  PVC White and Porcelain, which also failed in my example. The sections of wall on the image were finised with the same colours/textures as an example of what the fence segment should be. So you are right - use any actual texture and it will work and MOST of the Paint section colours are textures, so they are fine.

I know you understand this but for those who don't the image below should help you identify which are colours and which are textures.

However, they all worked in previous versions, so it has been reported as a bug.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Kajvin - Robert W on August 14, 2011, 08:24:11 AM
Hi Allan! If you take a digital photo of real p.v.c then you get structures and colour at same time.  ;D

 V16 only
I have look at the Tapered columns the work fine, you  have dx1 and dy1 at bottom and dx2 and dy2 at the top dx2 and dy2 must be less than 1/2 of dx1 and dy1

than example dx1,dy1= 200 mm then the max dx2,dy2 is < 100 mm not = 100mm that give you a square dart point up.
If yo will have a bigger top then bottom the set dx2,dy2 negative number from -0 and -2000 mm, it look like 2000 is max. This can be fun tool to work with. dx1,dy1 are max 5000  dx2,dy2 is max 2000 is width on columns. If dx1=5000 dx2 can be between 0 and 2000, h can be max c.

Try this select tapered Circle Hollow dy1=5000 dy2= 2000 c=2000 h=1999 and t1=4 Have fun ::) ;D
/Robert
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 14, 2011, 09:05:44 AM
Hi Robert,

Another option with colours, is simple to go to a Photo Editing program, select a blank image (usually File / New) and set the size to 128 x 128 pixels. Then just choose the colour you want from the colour selection tool and 'Fill' the image with the folour. Save it as a BMP or JPG and you have the colour you wanted but in the form of a texture.

Yes the tapered columns provide some interesting possibilities. I can see a number of uses for them in making other objects.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 14, 2011, 06:07:26 PM
Hi Robert and Allan,

Thanks for clarifying the colors/textures issue.  I reached into my Benjamin Moore cans of paint and painted over the moldy PVC rails and they're all good now in V16... and like Allan said, paint is cheaper than PVC anyways :)

I've been wrapped up in simplifying my custom light fixtures, etc.  When I created them, I guess the default # of facets was too high and I wasted a lot of memory on little details that you can't really see.   Its been a learning experience... Just don't know how many more lessons I need before I get done!

The tapered columns look like fun.  A bunch more possibilites.

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 14, 2011, 06:56:05 PM
Hi Dave,

I have actually nearly completed your light fitting in TFP, just for the challenge of doing it. All the curly bits are done so it just a matter of assembling it but the negative, in the way I have done it, is that there are dozens of materials to set. No problem in that you can just select them all and make them the one texture but to change the 'candles' for example to a different colour is hard because you do not know which of the numbered materials are the ones to change. I trying to see if there is a way t do it without so many material settings to do.

I will post it when it is done, (even if it is the present multi-materials version, just to see if it makes any difference to your project size. Have just got a few other things on at the moment so may be a day or two.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 14, 2011, 07:55:54 PM
Hi Allan,

Thanks for working the light fixture.  Thats a lot of work.   I had gotten the 'max 65000 points exceeded' for some of the components the tcad dxf import had created.  Turns out the 3ds export from tcad worked for sketchup import and I've been trying to learn that tool.  I tried building some of the fixtures features within TFP first but I kept stumbling.

I'd be glad to try your model in my project when you're finished.!

Thanks,
Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 14, 2011, 09:38:19 PM
Hi Dave,

The hard bit is done - creating the main 'scrolls'. Just thought I would show you how it looks so far. Will finish it shortly.

Allan.
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 15, 2011, 07:53:58 AM
Hi Allan,

Wow, Looks nice!  Are they 'walls' textured with metal?

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 16, 2011, 02:40:23 AM
Hi Dave,

I did it with walls originally but it ends up with too many materials. Each wall has 4 faces and avery curve of the wall means another set to 4 surfaces and textures so there are hundreds and trying to pick out the candles etc from them is almost impossible.

This model was done using a Member Profile set to that shape. This Profile shape much easier to do in Envisioneer so I did it there and it reduces the number of materials considerably.  

At present I am having a little issue adding multiple lights to the fitting - it is causing a light artifact for each bulb and they are showing in the render (white specks under the bulbs - see image), so I am trying to get rid of them. It works OK with one bulb but I will try a bit more before posting the BIM file, as it really needs the extra lights.

Whether it will help the size of your project though I don't know. It still has many fascets becasue of all the curves and the 3DS itself is 1.5mb. It does import OK though without compressing it, so presumably it is smaller than your TCAD one.

One thing it does demonstrate however is that it is quite easy to create some quite comlex objects within TFP.

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 16, 2011, 11:04:42 AM
Hi Allan,

The light looks really good.  Regarding the light artifacts, I recall seeing them in my tcad imported model as well.  Don't know how to fix that.

Regarding the profiles, I've turned straight line segments into members for crown molding, etc, but when I tried that with arcs, my only choices were walls or retaining walls, footings and edgings, so I tried it with walls but I didn't get too far.  I've had some success taking my tcad models into sketchup and replacing some of the features with much fewer facets.   I'm learning to draw arcs with few polygons and revolve 2D figures with few polygons...

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Allan Chesney on August 24, 2011, 09:30:23 PM
Hi Dave,

Sorry about the delay posting the light fitting, I have been somewhat busy lately but the BIM file is attached now if you want to play with it. If you import it through File>Catalogues>Import BIM File it should put it straight into your Interior Lighting> Ceiling Fixtures catalogue.

It is big as I mentioned before so may NOT help with your size problem.

The first try was done with walls but as I said it caused too many surfaces. For this one I created a new Custom Profile. The curly 'S' shape is just a cross section of the profile just 10mm long as is the curve.

The second image shows the two profiles I created and these are in the second BIM along with a few other useful shapes you may be able to use at some time. In fact the whole firring was made with these shapes: The 'candles' are the cylinder with an extened Sphere on top for the 'flame', the 'saucer' under them is a squashed Dome and the chair is the Chain (obviously) done with 'squashed' bits of the hollow Cylinder strung together. All the curves in these make it memory intensive of course! Once assembled it was exported as a 3DS and imported again as one fitting into Interior Lighting.

The added lights artefacts mentioned before are still in this (as all lights with added bulbs) - the programmers are investigating it

I will get around to splitting this topic sometime too as it has become quite divergent!!

Allan
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on August 30, 2011, 12:07:11 PM
Hi Allan,

Thanks very much for creating the light fixture.  It looks really good.  I've downloaded it and the useful shapes as well.  I'll be experimenting with it over the next week or so.

We've been a bit distracted lately with Hurricanes along the east coast, USA.  (had to evacuate for a bit, but everything survived :)  The house in the model is being built in North Carolina, a magnet for Hurricane strikes... (What were we thinking :)

Thanks again,

Dave

Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Jack Zimmer on August 31, 2011, 08:38:32 PM
Hi Dave,

I tried to review all comments but may have missed some on the long rendering time discussion.

I don't have your file but do note that your original interior posts had a number of doors and windows. The Dining room images.

I would check all the windows in the house for Specular setting (shine). Most likely you have them set higher than 10.  With high shine settings you can expect rendering time easily double those using settings between 5 and 8.

On exteriors, high Specular settings must calculate reflections of any and all trees / plants etc + thier shadows. Even if you don't see the reflections because the camera is distant your computer is still calculating them.

If you haven't try lowering shine yet, I'd do that and see what happens. Maybe you can set them to 7 and still get a tad bit of reflection you want.

Jack
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: dBminus on September 04, 2011, 07:50:32 PM
Hi Jack,

My windows had a default specular setting of 41.  I'll try setting it to 7 and check the render times.

Thanks for the tip,

Dave
Title: Re: V15 Indoor render to verify type and quantity of lights - feasable?
Post by: Jack Zimmer on September 07, 2011, 09:57:30 PM
Hi Dave,

That lower setting should help, but I did read where you had interior lights ON and that was your biggest problem for the speed issue.

Say, you mentioned your building location and high wind situations. Have you considered an ICF solution. I am currently working with a local distributor of Amvic ICF. He also has whole house solutions for heating and cooling that are fantasitc. And, a patent on a Trikin Floor system that increases spans up to 28'.  The homes result in a 60% reduction in energy. If you have any wind above 2 mph, you can get up to 85% with the new Honeywell Wind Generator Turbine (only 6' high/ 195 lbs)

Drop me an email if you want to look into it. I'd be happy to share what I know.

Jack