Basically, I want to start doing some motion tracking but all the tutorials that I have seen don’t really go into detail about matching your own camera specifications to that of blenders motion tracking system. I want to know exactly why and how to do just that. To be more specific, how do I determine my camera’s sensor width and how does this affect my workflow when motion tracking?
how do I determine my camera’s sensor width
Look in your camera manual or search for your camera specification in google. You may have to convert the sensor type to sensor width using the table here http://en.wikipedia.org/wiki/Image_sensor_format
There is a script that can take predefined lengths and angles to generate lens and sensor width in Blender. But you have to shoot those elements first.
would this work for my mobile phone camera? as samsung are pretty closed mouthed about the sensor specs for their phones.
yes i know its dumb trying to make a digital film with a phone that only record in 640x480 format, but we work with what we have, hey ?
I may be in over my head here. I tried using the add-on but couldn’t figure it out. I mean I did set the appropriate perspective lines but when I pushed the button to calibrate the current camera it caused my camera to turn upside down. From this point am I just supposed to turn everything upside down? I figured I could just go ahead and enter the settings manually by using the camera specs found here:
[http://www.dpreview.com/products/canon/compacts/canon_sx150is/specification] but that aspect ratio makes my clip seem too elongated even though that seems to be the exact camera I am using to film. Basically, I’m just looking for some verification of whether or not I’m applying the settings correctly.
Here it can be tricky. If you adjust picture aspect ratio in movie clip editor - it only adjusts the preview. It has no influence on how picture is handled by the solver. If the image uses different that 1.0 pixel aspect ratio you have to set it properly, because the solver will otherwise work on not correct data.