In the following tutorial we will go through the process of rendering an image in separate passes using render elements and compose them afterwards in photoshop.

Although I am a vray user, the same method can be used for any other renderer that supports render elements.

1)      Rendering the scene.

In the “render scene” menu, select the “render elements” tab and add the following passes: vraydiffusefilter, vrayGlobalIllumination, vrayRAWlighting, vrayRAWshadow, vrayReflection, vrayselfillumination, vraySpecular.

Note: if you have refracting objects in your scene, you need to add vrayRefraction as well. Also, if you don’t have self illuminating objects you can disregard the vray self illumination pass.

Render the scene and save all the passes (including the main image) in a new folder using descriptive names for each pass.

We still have to render the ambient occlusion pass. There are several ways to do this but I prefer to do it the old fashion way.

Before going forward with this make sure you save a copy of the file.

Select all the objects in the scene and assign them a vray light material. In the map slot choose vray dirt and set the subdivisions to about 30.

Before hitting render, disable the global illumination and de-activate disable all the lights (from the Vray global switches menu).

Save the rendered AO pass and open photoshop.

2)      Blending the layers in photoshop.

Open the main image, double click on it and change name of the layer to “original image”.

Open all the other passes and paste them in new layers in the original image file.

Make a copy of the vray diffuse filter layer and re-arrange the layer order as seen in the screen capture bellow:

Hide all the layers except the 2 diffuse layers, the raw lighting and the raw global illumination.

Create 2 layer groups and place one diffuse layer and the raw lighting in one group and the other 2 visible layers in the other group.

In each group, make sure that the diffuse layer is at the bottom.

Change the blending mode of both vray raw lighting and vray raw global illumination to “Multiply”.

Set the blending mode of the group above to linear dodge (see image bellow).

Up to this point, your image should look like the one bellow:

Unhide the vray raw shadow layer, and invert it (Go to “image”, “adjust” and click “invert”).

Set the opacity to about 20 and choose “multiply” as blending mode.

Unhide all the layers above, except for the AO pass and the original image and set the blending mode to “linear dodge”.

At this point you can unhide the original image and compare it to the result achieved with the blended layers bellow. You will notice that the blended layers look a bit better than the original image because the vray raw shadow pass gives more contrast to the image.

You can now delete the original image layer (we have only kept it for reference so far), and unhide the AO pass. Adjust the opacity to about 15 and set the blending mode to multiply.

This last step is very useful because it adds more detail and depth to the image.

Needless to say that rendering the image in separate passes is highly recommended because this way you can easily control the lighting, the shadows, etc. without having to render again.

You can download the layered psd file that I have used for this example here.

Additionally you can use vray wire color, vray material id or object id, to adjust the colors, brightness or saturation of each object in the scene independently.

Although architectural photographers make incredible efforts to shoot their photos without people in their “scenes”, in the architectural visualization industry everybody adds them to make the renderings look more natural, alive and to give it a better sense of scale, especially for exterior renderings.
However, blending people in renderings and make them look like they “are there” can prove to be a difficult task.

Choosing the right cutouts
Before starting the blending process it is crucial that you choose the cutouts that are right for your scene.
-if you have an overcast illumination you can not blend cutouts of people that have been photographed on a bright sunny day. They will have highlights and shadows on them, and they won’t look good no matter how much you tweak them in photoshop.
-find cutouts of people that were photographed from a similar angle. Imagine trying to integrate cutouts that were shot from an aerial view into a street view rendering.
-last but not least, try to find photos of people that have been shot next to a location that is somewhat similar to the one in your rendering. For example, if you are rendering a business center, you need photos of business men, dressed in suites, carrying laptops, etc.
(this one was a little obvious, but still worth mentioning).

Scale:
Nothing can destroy a rendering like people that are not at the right scale.  In street view renderings, the scale can be estimated a little easier by placing the people next to elements such as cars, doors, etc. In aerial views though, this can be a little more difficult.
Here is a tip that I almost always use to get the scale right:
When rendering, place cylinders at desired heights (1.70m-1.80m) where you will need to integrate people in post processing. This can help you a lot with the next step as well (shadow direction and sharpness)

Shadows:
1)    Direction
If you used the tip with the cylinder mentioned above, you won’t have any problems getting this one right. It is very important that the “faked” shadows cast by the people in the scene follow the same direction as other rendered elements. If you haven’t used cylinders, look at lighting posts, fences or other similar elements to see the direction of the shadow.
2)    Sharpness and intensity
Another really important step to keep in mind; If you rendering is shot at 12 o’clock your shadows will be sharp. By comparison, if it is a late afternoon rendering, the shadows will be less sharp and less intense.
However, don’t forget that even at mid day, the shadows are never 100% sharp, as mentioned in “4 tips for better architectural renderings”

The process of actually “painting” the shadow in photoshop is quite simple. First, make 2 duplicates of the cutout you need to blend in, and drag them bellow it in the layers stack.
Use rotate, scale and skew until you have the desired shape, size and direction (like in the screen capture bellow).

Hide the one at the bottom of stack for now, and select the other one. Desaturate it and adjust the brightness and contrast until it is completely black. Add a very subtle Gaussian blur, just to “break” the edges a little. Change the blending mode to “multiply” and the opacity to 50-60%. Now hide this layer, since you’re going to work on the other one a bit.

Unhide the other one at the bottom of the stack, desaturate it as well and make it completely black. Add a more intense Gaussian blur (depending on the overall illumination in your scene) and change the blending mode to multiplier and opacity to about 70%.

Unhide the upper layer, and using a very soft eraser brush, gently delete the parts that are further from the point where the shadow meets the “object” (in our case the feet of the person blended in).

Adjust the opacities of these 2 layers until you get the desired results.

If you have an overcast illumination, you would have to manually paint the shadows, using soft brushes.
(And don’t forget to remove the shadow of the cylinder using the clone stamp tool)

Color:
1)    Color match and saturation
If you do a simple test and make everything grayscale you will notice that everything blends together much better. That is because most of the times the colors of the cutouts have different levels of saturation than the rendering. If you have a shade of red in the cutout, and another shade of red in the rendering, check the rgb coordinates (or cmyk) of the last one and try to adjust the color balance of the cutout to match those coordinates as close as possible.
A somewhat safer way would be to choose cutouts that are colored in more “neutral” shades (beige, gray, brown, etc).

2)    Contrast
Match the brightness and contrast of the cutouts to the nearby elements in the scene. Objects that are closer to the camera and exposed to direct illumination have a higher contrast than the ones in the background or in shade..

3)    Color harmony
One last thing that I try to take into consideration is color harmony. If the general color scheme of the rendering is based on shades of greens and browns, for example, I avoid using bright orange, red or other vibrant colors for the cutouts. Although this has practically nothing to do with what happens in real life, in a 3d rendering, if you don’t consider it, the people may draw too much attention.

General tips
-Don’t place just individual images of people. People also hang out in groups of 2-3 or 4, so do that as well
-Don’t place the people too close to the camera, and more importantly try not to have them look directly at the camera.
-Carefully choose the placement of the people so that they won’t mask important parts of the rendering. Remember that the main “actor” in your scene is the building you are trying to showcase; the people are just details you use to make it more “alive”.

If this happens to you, than the following tutorial might help.

Check out this tutorial.

There are several ways to prepare a tileable texture starting from photos, but from my experience the method that I’m going to show you will work in 99% of the cases.

Preparation
Regardless of what method you will chose to create your tileable texture, there are some things that you need to keep in mind.
-the first advice that I give you is that if you have the chance to take the photos yourself,  DO IT!
-when taking the photo of the material sample, try to place the camera as perpendicular as possible, so you won’t have to work more later on it to correct the verticals.
-the light must be as uniform as possible over the entire surface of the sample. Avoid highlights or shadows because these are difficult to remove.
-if the material that you need to prepare is reflective, try to avoid placing it close to objects that might cast a reflection on it. This can be quite difficult, but try to do your best because it’s really important.

Making the texture tileable.
Here is the image that I started with (click on it to view a higher res version).

Open the photo in photoshop, click on “Filter”, “other”, and “offset”.

In the fields next to “horizontal” and “vertical” type a positive value equal to half of the dimension of the photo. For example, if your photo has a dimension of 900×900, make an offset of +450 on both dimensions.

If you take a close look, you will notice that there is a visible separation both on vertical middle and horizontal middle of the image

This can be adjusted easily using the clone stamp tool, with various parameters under “hardness” and “master diameter”.
Here is the final result (click on the image to view a higher resolution texture)

Although this tutorial has been written especially for 3d artists, I think it can be useful for webdesigners and graphic designers in general, that need to create seamless textures for backgrounds, wallpapers or other similar stuff.