FAQS: Rendering

Improving RenderZone rendering performance: Why does my file render so slowly?

Hardware: Rendering speed is controlled primarily by the speed of the processor. The faster the processor, the faster the rendering is. If you do not have enough RAM, this will force your computer to use virtual memory, which will slow you down with OS X, and slow you down a lot with Windows. If you are regularly working with a file of this size, you should consider adding more RAM. The video card will only affect the Wire Frame and Interactive Shaded displays, and will not have any effect on rendering speed.

Shadow maps: One frequent cause of insufficient memory is too many, or too high resolution shadow maps. See the Shadow map memory FAQ for further details.
Geometry: The more geometry is contained in your scene, the longer it will take to render, and the more memory it will require. In general, you should only model what you will see. For example, if you are working for a hardware manufacturer and making a presentation model of a doorknob, you can model all the details, curves, keyhole, etc. But if you are making a model of a subdivision, you should create extremely simple doorknobs (if any), and you should never use the "presentation doorknob" in a "subdivision model" -- even if you have it handy from a previous job, or found it somewhere on the web. Inspect your polygons. If an object looks like it is shaded in a Wire Frame display, then most likely its resolution is too high and it should be modified or recreated.

Surface styles: The more surface style effects you select, the more time it will take to render. Therefore, if you are concerned about rendering speed, you should only select the options that you will need. You can also get an idea of how complex a surface style is by watching how long it takes to draw the small preview of the surface style icon. Changing the preview to sphere will take a little longer to draw the preview, thus giving you a better relative idea of its complexity. The type of preview icon will not have any affect on the time it takes to render objects with this surface style. You should also generally avoid "double transparencies." See the Glass surface styles FAQ for further details.

Rendering options: The more rendering options you select, the more time it will take to render. The main option to avoid (if possible) is Super Sampling. See the Super Sampling FAQ for further details.

Lighting: Advanced lighting effects such as Atmospheric lighting, Area or Line lights, Accurate Glow, and Environment lights will add (perhaps considerable) time to your renderings. Use the basic lighting suggestions from the Lighting FAQ for fastest results and see the other tips for optimizing more advanced lighting.

Hidden Line rendering (or Surface Render) performance: What is it?

Hidden Line rendering is somewhat of an intensive process because, unlike most other programs, form·Z will derive the intersection lines between objects. To do this, form·Z actually executes Boolean operations between all intersecting objects in a file. Thus, if you have lots of intersecting objects, this can add considerable time to the rendering. You will probably save time by either executing the Boolean operations in advance (in case you need to generate more than one Hidden Line) or otherwise build or modify the geometry so that it is free from intersections.

To see if Intersections are an issue with your file, start the Hidden Line rendering process. The first portion of the progress bar will be for Forming Surfaces, and the second portion will be Resolving Intersections. Compare the number of polygons at the end of each of these and if they are the same, then your file is free from intersections. If the number is a bit larger after Resolving Intersections (which is normal) then this is probably not slowing the rendering down too much. But if the number of polygons is more than 50% larger, then this will be slowing you down quite a bit, and the rendering will be much faster if you eliminate some or most of these intersecting objects.

Also, the number of lines to be plotted on screen can affect the performance. Generally small objects relative to the scene that are highly detailed should be avoided when performance is a consideration. Very large objects relative to the scene (such as a huge ground plane) could also adversely affect performance and the extents of a ground plane should generally be limited as much as possible. The Hidden Line Options dialog also contain an option to Hide Edges based on the interior angle between surfaces. Checking this option will reduce the number of lines that need to be plotted for the Hidden Line, generally reducing the time necessary to complete the rendering.

Lighting: What lighting options does form·Z offer?

(In addition to the following brief summaries, see section 6.5 of your User’s Manual for more complete details.)

Basic lighting: If you find that your shadows are too dark, or that the "back side" of their model is too dark, you can brighten them by simply adding a few extra strategically placed light sources. Please follow this link for a discussion on this, including a simple solution using 4 distant lights. This discussion also includes a Basic_Lighting.fmz file you can download that demonstrates this lighting setup.

Light direction from Sun: form·Z allows you to position a Distant light source based on Geographic Location and Time of Day / Year. To set this, edit the Distant light by double-clicking on it from the Lights palette. Click on Light Direction from the Sun button and select the desired options and settings. Note that, if you do not find your location from our extensive database, you can either enter the Latitude and Longitude, or you can add your site to our list of cities. Click OK and the sun will be positioned automatically. Note that you can also animate the sun over time using the Sun Animation command from the Extensions menu.

Atmospheric lighting: In reality, sunlight bounces off the atmosphere of the earth and illuminates objects from all sides. form·Z allows you to simulate this effect using a "sky dome" of lights. To set this, edit your Distant light, select Accurate Intensity, and click Options. You need to reduce the Output Power of the sun to somewhere between ~0 - 1.0 Watts / SqFt, and the Sky Luminance should be set between ~5 - 25 Candelas / SqFt. For a starting point, you might try 0.08 Watts / SqFt and 10 Candelas / SqFt. Note that the higher the Quality of the Atmospheric light, the more light sources will be used and thus the more subtle the shadows, and the longer the render times. Version 5.0 has also added a Simple Atmospheric Light option that will allow you to specify the exact number of light sources (samples) used, as well as allowing exact control over the color of the Atmospheric light.

Area lights: Area lights allow you to turn any object into a light. Each face of the object will then become its own light source. It is generally recommended to use objects with the fewest possible polygons for best performance. Increasing the Quality Slider will generate additional samples per face, resulting in better quality renderings (and increased render times). Common uses include setting a plane above your objects to provide a nice general illumination, or placing a plane in a window opening to simulate the light coming in from outside. Note that light will only be emitted in the direction of the normal of the surface. Note also that Area lights do not work well with Soft Shadows, so if you need shadows, use Hard Raytraced Shadows and, with a reasonable quality setting, you will get soft looking shadows, because light is emitted from the entire surface of the Area light, not just from a single point. See section 6.5.6 of your User’s Manual for further details.

Custom lights: Custom lights can emit variable intensities in different directions about a light source. You can either define this distribution manually, or you can load IES data that has been preset for a particular fixture. Since many lighting manufacturers provide this data as free online downloads, Custom lights allow you to reproduce the light from a specific lighting fixture from a specific manufacturer. See section 6.5.7 of your User’s Manual for further details.

Environment lights: An Environment light uses a standard (TIF or JPG or HDRI) image to define the intensity, distribution, and color of the light. The light is wrapped around the entire scene, so you can effectively light the entire scene with a single light source (or perhaps combine this with a single Distant sunlight to create harder directional shadows). The # Of Samples field controls how many individual light sources are used to interpret the image, and you should be able to get a quick idea of how this will look using 50-100 samples. For a final rendering you probably want to use ~300 - 1000 samples (typically more samples are needed for interior renderings because the light will be illuminating objects through smaller window openings). You might also want to reduce the Color Factor to ~10-20% for more realistic results. See section 6.5.8 of your User’s Manual for further details.

Light glow: Light glow simulates the scattering of light as it passes through a smoky or foggy area. Simple Glow will not be blocked by other objects (even if the light is casting shadows) but Accurate Glow for a shadow casting light will be stopped by shadow casting objects. Thus you could use Simple Glow for spot lights on a stage, or Accurate Glow to simulate the streams of light coming through a cathedral window. Accurate Glow will also add considerable time to the rendering. See section 6.5.5 of your User’s Manual for further details.

Glass surface styles: What is the best ways of applying them?

Glass reflections should generally not be used with another type of transparency, as this creates a "double transparency" which will take considerably longer to calculate.

Glass Reflection includes a Refraction coefficient which controls how light will bend as it passes through a material. Thus this surface style should only be used for solid objects and not surfaces or surface solids, which do not have any thickness through which the light can bend.

Because Glass Reflections refract the light, they are not included in an Alpha Channel Background (because you can't "bend" something that is not there at the time of the rendering). If you want to use an Alpha Channel Background and do not need Refraction, you can use a Mirror Reflection combined with a Simple Transparency and then these transparent objects will be included in an Alpha Channel Background. Glass Transparencies should be included in Alpha channels.

Transparent shadows can be generated so that light will pass through glass materials based on the Transparency / Transmission factors. Transparent shadows are set in the Shadows section of the Light Parameters dialog (which can be accessed by double-clicking on the light from within the Lights palette.)

Transparent shadows will also pick up the color of the transparent material so, for example, the colors of a stained glass window could be projected onto a wall or the floor. Transparent shadows do take a bit more time to calculate, so another option might be to select the glass objects and turn off the Attribute: Casts Shadows. This will allow light to come through your windows without extra calculation times.

Moire patterns: Why do I sometimes get radial patterns in my images, especially when viewing a grid or a brick surface style in perspective?

Moire patterns can occur when there is insufficient resolution to show the necessary detail in a perspective image. The easiest / quickest way to fix / improve this is to edit the surface style that is responsible for the moire pattern (usually a Grid or Brick shader), click the Color (or other) options, and check the Area Sample option.

If this does not sufficiently resolve the problem, the next best solution would be to render at a higher resolution. If you cannot increase the resolution of the rendering (ie, you are generating an image for the Web, or an animation), see the section below.

Super sampling: What is it and how is it used?

This is a rendering effect that can be used to reduce or eliminate aliasing or moire patterns. To achieve this, the image is rendered at 2, 3, or 4 times the specified resolution (for Low, Medium, or High Super Sampling) and then sampled down to the desired resolution. Because rendering at 2, 3, or 4 times the resolution generates 4, 9, or 16 times the number of pixels, the rendering times will also increase up to 4, 9, or 16 times, when using Low, Med, or High Super Sampling. Because of the amount of additional time that super sampling adds, it is generally recommended to try other methods for reducing moire patterns first, such as enabling the Area Sampling option from the Surface Styles Options dialog, or rendering at a higher resolution.

Super sampling can also be accomplished manually, by rendering at a higher resolution and then reducing the resolution of the final image by exactly 2, 3, or 4 times. This option may be desirable because in this way you will be able to use the higher resolution image as well, if you like. For example, you could generate a high resolution rendering for print and then manually super sample it to generate a high quality / lower resolution image for a web site.

Rendering problems: Why am I having problems rendering a particular model or view and how do I fix it?

Rendering problems could occur for a number of different reasons.

While modeling calculations are incredibly accurate, rendering calculations can break down if your objects (or lights or views) are too far from the origin. In general you should keep your objects (lights and views) within 5 - 10 miles, or 10 - 15 kilometers of the origin. If you find that your file is farther than this, make sure that all your objects and layers are unlocked, Select All objects (including the ghosted and invisible objects), select any lights or views that are in the same location as your objects, activate the Move tool and Point Snap, and place your cursor on a part of your file that you wish to move to the world origin. Make sure that the AWC boxes are checked in the Prompts palette, and make a note of the distance that is shown there. (This can be useful for aligning additional data if you need to import / export to another program again). Then click in the screen to start the Move, enter 0 in the Prompts palette, and press the enter or return key. All your objects (and lights / views) will then be moved to the origin.

It is also possible that most of your objects are located reasonably close to the origin, but some objects (or lights or views) are far from the origin. To eliminate this as a cause of the problem, check the parameters of your lights and views, and try to Fit your file from a Top AND Front view. If it does not fit properly, then you have some objects that are out in space. In this case, unlock all objects and layers, frame pick half of the screen and select the Cut command from the Edit menu. Fit your project again. (You can get a New Model and Paste what you cut to see what you have removed from the file.) Repeat as necessary until your file fits properly.

If your objects are reasonably close to the origin and your file fits properly, then the problem could be caused by a bad object or attribute. To find this, turn off half of your layers and render again. If the problem goes away, turn off these layers and turn on the other half, save the file under a new name, and repeat as necessary. Once you have simplified this down to a single layer, then frame pick half the objects and Cut them out of the file. If the problem goes away, Select All and Paste (this will replace the selected objects with the copied objects). In this way you can identify the object(s) that are responsible for the problem. Most likely if you delete or recreate this bad object you should be able to resolve the problem.

Note also that you can try to simplify your rendering options -- either before or after following the steps above. For example, if you are generating a RenderZone rendering, you could turn off Shadows, Textures, Reflections, and Transparencies. Noting which options are required to cause the problem can help identify the cause of the problem, or rendering with as few options as necessary to reproduce the problem can speed up the identification of the bad object.

After completing steps 1-3, if you have identified an object that is causing a problem, or if you have a file with a problem that you are unable to simplify, please send this to support@formz.com. We can then find the cause of the problem, report it so that it can be fixed, and help you find a work-around if necessary.

Soft shadows: Why can't I see my soft shadows, or why are they so fuzzy or rough?

Soft shadows utilize shadow maps to create the soft edges. In certain conditions the shadow map resolution can be too low, which can result in the pixels of the shadow map being visible, or in extreme cases, it can cause the shadow to disappear altogether. (Think of zooming in really close into a low resolution image.) To solve this issue, you need to know a little more about how shadow maps work:

By default, the size of a shadow map is the smallest bounding box that encompasses all of the objects that cast shadows. And by default the shadow map resolution is equal to the resolution of your renderings (which you specify via Image Options from the Display menu). If your view contains all of the objects that cast shadows, then you should not have a problem with shadow map resolution, but as you zoom into your scene you are effectively zooming into your shadow map.

Most often users encounter problems with shadow map resolution when they introduce a large ground plane into their scene. To remedy this situation, simply Query the Attributes of the ground plane and set its Casts Shadows Attribute to Off. Leave Receives Shadows ON so that it will still receive shadows from the other objects in your scene. If you would like to change this (or many other attributes) for multiple objects simultaneously, you can do this with the Set Attributes tool.

If turning off the Shadow Casting Attribute for the ground plane (and outlying objects) is not enough, then you can also change the Quality of the Soft Shadow to High. In extreme cases you may also need to increase the resolution of the shadow map or limit its extents from the Shadow Map Options dialog. Edit your light and click the Map Options button from the middle of the Light Parameters dialog. Here you can increase or decrease the resolution of the shadow map. Note that doubling the "Times Image Size" number will quadruple the memory requirements for this light, which will be displayed at the bottom of this dialog. You can also use the Limit Map: option to constrain the size of the map to All Objects (default) or All Completely Visible Faces or Objects. The smaller the size of the map, the less resolution you will need to get the same quality shadow, but note that limiting the map may cause shadows to be dropped from objects or faces that are not completely in the scene.

Hard Raytraced shadows do not have this limitation and will always produce shadows with crisp edges.
See the FAQ below for further details.

Shadow map memory: Why do I get an insufficient memory message?

Shadow map memory requirements are determined by the size and resolution of the shadow maps, and by the types of lights that are casting shadows. If you edit your light parameters and click on Map Options, you can see and control how much memory your shadow maps will use. By default this will be set to 1 times Image Size. Thus as you increase the resolution of your image (from the Display Menu/ Image Options), the memory requirements for your shadow maps will also increase. If you double the resolution of your image, it will double both the X and Y size of your shadow map and thus this will require 4 times more memory for the shadow maps. Alternately, if you enter 0.5 or 0.7 times Image Size in the Shadow Map Options, you will use one quarter, or one half the amount of memory, respectively. The resulting memory requirements for each light will be shown at the bottom of the Shadow Map Options dialog.

Thus, to check how much memory your shadow maps will require, first set your desired image resolution and then multiply the memory requirements for one light by the number of lights that are casting soft shadows. Note also that Point lights will cast shadows in 6 directions and thus will require 6 times more memory than distant or cone lights. If you do not have enough memory for this, you can either reduce the number of lights that are casting soft shadows, use cone lights instead of point lights to cast the shadows, or reduce the resolution of the shadow maps.

form·Z version 5.0 and later will automatically detect when you do not have enough for full resolution shadow maps and give you the option to automatically reduce their resolution.

Image resolution: My image is too low resolution or too grainy. How do I fix this?

You can set the resolution of your window from the Image Options command in the Display menu. Of course the higher the resolution you set, the larger an image file you will create and the longer it will take to render. Doubling the resolution will approximately quadruple the image file size and render times.

In general, final "standard print images" should be in the ~3000 - 4000 pixel range for good results. Thus if you are rendering an 8 x 10" image, you might want to select 300 dpi (which would result in a 2400 x 3000 pixel image). Alternately, if you are creating a 24 x 36" image, you might want to select 100 dpi (which would result in a 2400 x 3600 pixel image). This rule works in most cases because you will typically inspect a smaller image more closely, while people will generally stand farther away from larger images. Of course it is possible that you may need a higher resolution rendering, but this is a good final resolution to start with.

Also note that the Set Image Size rendering option (where you specify a portion of an image to render) will only render at screen resolution. You need to disable this option to render at the resolution you specified in the Image Options.

Image map resolution: What size and resolution should I use for my image map, color maps, or transparency maps?

In a perfect world, you should size your image maps such that you have at least 1 pixel in the image = 1 pixel in your final rendering. Thus, you should size these so that they are not too small (or you will see the pixels of these images in your rendering) and not too large (or you will increase memory usage and render times). For example, if you have a picture on a wall, and the picture occupies 2" x 2" out of an 8” x 10" rendering @ 300 dpi, then your picture image map should be 2” x 300 dpi = 600 pixels (square). If you are using a marble tile for the floor, and the largest tile in your image is 1" x 1", then at the same 300 dpi, this should be 1" x 300 dpi = 300 pixels.

As another example, if you have an oriental rug on the floor (where the pattern is not repetitive) and the rug fills the entire bottom of the image, then this should be 10" x 300 dpi = 3000 pixels. Alternately, if you are rendering the same image at 24 x 36" @ 100 dpi, then you would want this image map to be 36" x 100 dpi = 3600 pixels. And then if your rug is bi-laterally symmetrical, you can use only 1/4 of the image at 1/2 the pixels in each direction.

By sizing your image maps accordingly, you can get the best rendering results using the least amount of time and memory. Of course if you are unsure, you might want to do a quick calculation as outlined above, and then make the image map just a little larger, just in case. Also note that if you are concerned with the qualilty or performance of an image map in Interactive Shaded mode, you should size your image maps based on powers of 2- such as 256 x 256, or 1024 x 1024.

Imager: How do I generate multiple renderings and save them as images, one after the other?

form·Z includes the “Imager," which is a batch rendering utility. You can access the imager by either running the stand-alone application from within your form·Z folder, or by selecting New Imager Set from the File menu within form·Z. This utility allows you to load a list of views (or scenes) from existing projects and render these using any of the available render modes at any possible image resolution. As soon as form·Z completes the first image, it will save it in the desired format, and continue to the next rendering. Once the Imager Set is complete, you can double-click on the Complete Status of each image to see its render time, or click the View button to see the resulting image.

While form·Z version 4.x and 5.0 will only allow you to run either form·Z OR the Imager, version 5.5 and later allow you to run both applications at the same time on the same computer. Version 5.5 also allows Imager Sets to be network rendered, so you can use the power of multiple render clients to process the jobs. See the Network rendering FAQ for further information.

Video cards, drivers, and rendering speed: Will a video card speed up my rendering?

Better video cards will speed up the Wire Frame and Interactive Shaded display modes, but all other display modes (such as Hidden Line or RenderZone) will not be affected by the video card. The speed of other display modes are controlled by the processor, so if you need to speed up rendering, you should first make sure that you have enough RAM, so that you are not using virtual memory (check the Memory Meter at the bottom of the form·Z window or check the Task Manager on Windows, or the Activity Monitor on OS X). Using virtual memory will slow form·Z down (somewhat on OS X and a lot on Windows) and thus you might consider adding more RAM so that form·Z has enough actual memory. If you still need more speed, then you should upgrade your processor(s). Note that most render modes will only use a single processor, but the Raytrace mode will use as many processors as are available. Thus if you have a computer with more than one processor, you should generally render with the Raytrace render mode.

It is also generally recommended to use the latest video driver for your card. On OS X, simply running Apple's Software Update will do this automatically. On Windows, go to the web site of the manufacturer of the video card (ie, www.ati.com, www.nvidia.com, etc.) and download and install the latest driver for your card and operating system. Note that most manufacturers recommend downloading, uninstalling, rebooting, turning off virus detection, and then installing the latest driver.

See also Improving RenderZone Performance FAQ.

Network rendering: How do I set up a render farm and what are the requirements for a render client?

Here are the basic steps to set up a render farm:

Install the Render Server on any computer that will be ON on a regular basis, has a good network connection, and has enough disk space to store all the files. Then launch the Render Server.

Install the Render Client on as many computers as you like as long as they are "powerful enough" to render the jobs by themselves (ie, they are reasonably fast and have enough RAM). Putting an old slow computer as a Render Client may actually slow down the network rendering process, but any modern computers should be fine as long as they have enough RAM. Run the Render Clients and add the IP address of the Render Server the first time you connect. (The IP address of the Server will automatically be displayed at the top of the Render Server window.)

Launch form·Z, open the project file, select the view you wish to render, select the desired resolution (from Display / Image Options), and select Network Render from the Display menu. (Add the IP address of the Render Server the first time you connect.) Select the desired Renderer and Render Options, specify either Manual or Automatic Retrieval, and click OK. With version 5.5 and later, you can also network render Imager sets. Simply create an Imager set and click the Network Render button. See the Imager FAQ for more info.

Note that all programs (form·Z, Render Server, and Render Client) should be of the same version (ie, 5.5.2, not some 5.5.0, some 5.5.1, and others 5.5.2...). Note also that you can install all applications on the same computer, but you will need to have enough RAM for all applications, and if the Render Client is processing in the background, you will notice it. And, of course, all the computers need to be networked together. If there is a Firewall between the computers, you need to open a port to allow form·Z and the render components to communicate.

All the network rendering commands and functions are described in Section 3.6.8 of your User’s Manual.

Rendering: Why do my reflections disappear when I set the reflection value below 10%?

You need to set the Minimum Ray Contribution to a value lower than your reflection value. This setting is found in the Raytrace Options from the top of the RenderZone Options dialog. Lower values will produce better results, but also add to the render time.

Rendering: Why do I get a circular shape in my windows when rendering with an accurate glass reflection in certain views?

This artifact is caused by the Minimum Ray Contribution. Lower values will produce better results, but also add to the render time.

Rendering: Why does my stack of glass panes look black?

This can be caused by a setting too low in recursive rays, which can cause a dark glass effect if you have more layers than recursive rays. Increase the number of Recursive Rays found in the Raytrace Options from the top of the RenderZone Options dialog, or reduce the number of glass layers. (Note that higher values may require more time to render.)

Rendering: I see holes in my surfaces when i use raytracing that are not visible when I use z-buffer?

You have non-planar faces in your object. Triangulate these or use the Decompose Non-Planar Surfaces checkbox the RenderZone Options dialog.