I’ve been doing this same thing myself for a about 5 months now, in fact I’m in the middle of a job right now. So far I’ve done just over 80 photo composites. Here are some of the things I have learnt. Your client is not really interested in the “how” of what we do, they are just interested in the end result and how much it is going to cost them. The real estate agent knows that a furnished sales picture is better than a vacant picture. Their options are either stage a room with hired furniture and have a photographer take pictures, or use 3D composites, however the objective is to create a 3D composite that not only looks real but is appealing and inviting to prospective purchasers. If you can’t do that then you won’t get any return business.
3D composites are certainly cheaper so you need to pitch a price that is competitive to similar services alreadybeing offered in the market place. How much you charge is really up to you but keeping things simple is generally the best option. Whatever price you decide on make sure you also quote on the charge for revisions such as a change of colour on a lounge suite, a different picture in a picture frame, or a rearrangement of furniture, etc. Anything more than this and you should charge the full amount for the composite again. The jobs that I do usually involve at least four pictures per property, a bedroom, living room, bathroom and kitchen. Each type of room presents its’ own challenges but rooms with windows and mirrors take the most amount of time. Bathrooms with multiple reflections off mirrors and glass showers screens are often the most difficult and time consuming both in creation and in render time. What you have to ask yourself is how much your time is worth when you take into consideration the time it takes to camera match the photo, add the furniture, rendertime and compositing.
If you can get your workflow as efficient as possible it will help to maximise your return. Some of the things that I do include having a template job folder with sub folders for your blends, source images, renderlayer outputs, and final images. Create a startup blend with a room, camera and base light setup, and correct render settings. Compile an extensive texture library and model library. What saves a huge amount of time is to have furniture combination blends already to append into a scene such as dining room sets that include the table, chairs, place settings, glasses, table decorations etc. Lounge settings with side tables, lamps, coffee table, rugs etc all ready to go and bedroom settings setup the same way. Another thing I have found really usefull is to have two makehuman models, one in a standing position and one in a seated position included in my startup blend. This is very useful in scaling everything in the room properly. For example once your camera is matched to the photo add the layer with the standing human model on it and scale it to the correct size whilst it is standing in the doorway of your backdrop photo. This way when you add beds, lounge suites etc to a scene you can then scale these objects relative to your human model.
Last tip is don’t get caught up in trying to make your composites “physically correct”, they just have to “look” right. Most times a HDRi for reflections and ambient occlusion for shadows provide convincing (but not accurate) results. Unless an ouside light source creates a distinctive inside effect or the photographer used a flash when taking the source photo, more complicated light sources are probably not necessary. Someone looking at a real estate ad is not going to notice that there are no caustic effects from the wine glass on the table so don’t make more work for yourself than is needed.
I hope you find some of these tips useful.