TheKitchenAbode

Hi Mark, Thanks for the more detailed explanation concerning your issue. Think I understand now. Have had similar situations where I need vertical dividers in a cabinet. Say for cookbooks or a tray divider. As I usually only have to do this once in a while I just float a partition(s) into an open cabinet to create the look. Would definitely be much better if we could control all the items that make up the box. As you mentioned in your original post most manufacturers offer both 5/8" & 3/4" box construction, it would be nice if this could be spec'd and shown. Graham

Ok Mark try this. Framed Cabinet Inset Separation "0" in under framed. Open up the front in the cabinet DBX Delete the bottom separation Select "Add New" "Separation" The default "0" should work. If you just go on a separation and try to change it manually to "0" it reverts to 1/16". This should give you the look. Does this help?

Mark - Do you mean that you want the bottom to extend out to be flush with the door face? The frame is around all edges of the door. Graham

For this I set the box construction to framed and full overlay. Now you can control the default side & separation widths down to 1/16" Graham

Hi Dennis, I change all point lights to either spot or off. Say for example you have a wall sconce point light, if you aim a 3D spot light with shadows off at the wall sconce light and adjust the angle & drop off rate you can simulate the lighting effect on the wall behind the point light. To make the wall sconce appear on increase the fixtures emissivity to make it look bright. That's what I did in the Riverstone bathroom renders. Also used 3D spot lights between the ceiling beams to create the downward light effect. From my testing you have to turn-off all point lights in the plan, even having one turned on will really slow down your Raytrace. Best is to just use a very simple plan with one or two lights. Raytrace with all lights off. This is likely as fast a trace as you will every get. Then make the light(s) a spot and run the Raytrace. Then change it to point light(s) and run a Raytrace. I think you will see a big difference in times. No need to run more than a few passes as for a given scene every pass takes the same length of time. Let me know if this helps. Graham

Doug, From a programing perspective I understand enough and appreciate the complexities involved. What I am driving at is why there appears to be a lack of any definitive information as to the relative impact on Raytrace times for a given set of parameters. For example, I just ran the following comparison. 1630 x 791 @ 300 pi = 8:47 minutes, 15 passes 1630 x 791 @ 1200 pi = 8:57 minutes, 15 passes 1630 x 791 @ 4800 pi = 9:46 minutes, 15 passes I was expecting the times to have increased according to the p/i density count increase. Took these pics into Photoshop and realized that the images were not really being generated at a higher p/i for the set size. The pics were being resized downward as the p/i increased and as such there was no purpose in increasing the p/i, which explains why the times were not effected. This may have been the programmers intent but I am certain that most users do not interpret increasing p/i in this manner. My assumption was that the resolution for that size was being increased and as such the image could be enlarged without suffering significant loss in quality. Just a simple example that only took about 1/2hr to investigate. I am certain that other users have done similar investigations/comparisons, still baffled as to why the answers seem to be so elusive. Graham

Doug, I have mentioned in this and in other postings that point lights for some reason have a very negative effect on Raytrace times. Even one in a plan will degrade performance significantly. I never use point lights for this reason, only spot lights. For me they are the most significant item concerning Raytrace time. I have also found that some materials can cause problems, not sure why but some polished items and some metals. I always change polished to reflective and alter metals to general type material. The other factor concerns shadows turned on with lights. I try to minimize how many lights have this turned on. In most cases it is not necessary to have every light generating a shadow. One or two or just a 3D light will usually do the trick. You can see from the Raytrace times I have posted above that by adjusting these items alone I was able to drop the Raytrace time from 54 minutes per pass down to under 1 minute per pass. I could never get that improvement level no matter how much hardware I threw at it. What surprises me the most is that I am not hearing from any other member as to their experience concerning these items or whether they are able to confirm this on their systems. Raytrace times are for many of us a real issue and as such I believe it is important to nail this down once and for all. Keep in mind that potential new users of Chief run the trail version and scan this forum, if I was evaluating this software and ran a Raytrace on the Riverstone example I would not be impressed at all. These renderings are extremely important to my clients and my sales, and the ability to generate them as rapidly as possible is equally important to the sales of Chief Architect. Graham

Hi Scott, I would not under estimate the value of your system. From the preliminary results when Raytracing this plan as is you did 10 passes in 1 hour, 45 minutes. On my system it took 54 minutes to do just 1 pass. Your pushing this through at least 6 times faster. Also, there is no way I could run such a scene on my system, it would take about 10 hours. There is also another important advantage with your system. As my designs reach completion I have to refine color & texture, as we all know there is often a significant difference as to how these appear in a regular camera view versus a Raytrace. To get these right I need to run a Raytrace, usually about 2 or 3 passes. Even with my tweaks this takes about 2 to 3 minutes. In most cases it may take 4 or 5 adjustments per color or texture, combine this with maybe 8 different important materials to adjust and you can see how this time can really add up. For my work these lighting tweaks are more than acceptable, however there are members who may not have this option. I believe Jon is a prime example of this as it appears he is producing images for print/publication. This would require the utmost accuracy and quality and a system that allows this to be achieved in as efficient manner as possible. Graham

Just ran again. This time 30 passes, 26 minutes 28 seconds. 57 seconds per pass. Was doing some other work at the time, likely a smidge slower due to multi-tasking. Graham

Here is my last run. See screen capture below. All lights are off, just using 8 3D Spot cameras to light the scene. This did 10 passes in 8 minutes 56 sec, that's 54 seconds per pass. Probably about as good as I will ever get on my machine. Graham

Hi Scott - I have the Raytrace settings as follows: Use Camera View Settings "OFF" Use Ambient Occlusion "On". Min .3, Max 1 Direct Sunlight 5.0 Enable Environmental Light "Off" Photon Mapping "Off" Size is 1630 X 791 300 pixels/inch Graham

Hi Scott, Over the years I have found two things that really impact on Raytrace times. One is the number of lights and more importantly the type of light. The tweaks I performed was to turn of all lights that were not important to the scene, this reduced the lights on to about 8. The other more important one was to get rid of those point lights, even one will slow down rendering 8 to 10 fold. The sconces beside the mirror are points, changed them to spots. There are also several 3d cameras pointing towards the mirror walls, these are also points, changed them to spots. Made some adjustment to these to create a similar effect as a point light. If you do this with your machine I suspect your Raytrace time per pass will be in the seconds. To make the sconces appear lite I cranked up the emissivity on the glass, around 70%. Graham

Thanks for the support Cheryl & Rick. Will do that. Many Thanks, Graham

Fear not ye mere mortals for we are not to be defeated today. With just a few lighting tweaks. 10 passes 10:38 minutes/sec, that's about 1 minute per pass. Intel 2 Quad Q6600 2.4 GHz 4GB ram For comparison the original without lighting tweaks took 54 minutes to make just one pass. Graham

Thanks Larry - Got it. Graham

Is there a link for this Riverstone plan?

Any chance that you could post this plan so we could all give it a try? Graham

Scott, I think the last Raytrace time is really good, that equates to 10 sec per pass. Was just wondering how you got such a significant reduction compared to your first posted result that was around 1 minute per pass.

That's how I do it. Sometimes for cabinetry glass doors I need a different grid layout than the one generated by Chief. Make up my own and float into place. Graham

I think it would be useful if there was a way to generate a list of the objects in a plan that was using a specific material. Graham

Hi Scott - is this correct, on your first run you reported 16 passes in about 16 minutes. Now you are reporting 10 passes in 1:45. That's a very significant reduction in Raytrace time. Did you make other changes? Graham

Suggest using the Windows task manager. It will display a graph for each core and show % usage. May be better indicator than core temp. Graham

Don't we already do this when using Chief, we copy symbols & objects from the library catalogs, download from 3D Warehouse, etc. The design is the arrangement of objects and not really the object in itself. Certainly if I saw something unique in someone's plan that I would like to use I would ask them if it would be ok to use it. The question however would be if it is truly unique or novel. For example I recently posted a symbol/object with a described technique that allows a molding/casing to be recessed. From what I can gather this could be considered unique & novel. However from my perspective it really has no monetary value, unlikely to become a millionaire by taking out a patent or copyright on this. Please feel free to copy this, my reward is to see it used. Graham

Chief should open up another forum category "Good Plans" then we can post them for other members to use. Maybe if they are really "Good" we might be able to sell them to each other.

Agree 200% - When people start copying your design work then you know you've done something right.