In pursuit of perfect panoramic In pursuit of perfect panoramic The creation of panoramic photographs can arise from several points of view, we can find a picture covering a viewing angle greater than what our lenses can be captured or simply want a higher resolution picture of what our camera can get. A well-executed panorama allows arbitrary resolution photographs at a cost far less than what a camera would come out equivalent to do in one shot.
A high resolution photo prints allows arbitrarily large without loss of detail, in many cases a picture may even exceed the threshold of human eye resolution. It is true that a good impression words less megapixels than you think but it is never more than on resolution.
A panoramic photo is making virtually enlarge the size of the sensor, a panoramic photo of 3 photos at 100mm equivalent to taking a picture at 100mm with a sensor 3 times greater than we have. It is not equivalent to taking the same photo in a shorter focal length as the outlook remains which still corresponds to a 100mm lens. About
FOV, the number of photos and the resolution The first step in a panoramic photography is to decide the focal length to use, the number of photos to take and the final resolution we want. Let us assume that the viewing angle fov will call by its acronym in English (field of view) is fixed as it depends on our composition and composition, we assume, can not be changed. Based on a fixed fov there are two scenarios:
1. Set the focal length and based on the viewing angle to determine how many photos to take.
2. Set the number of pictures to take and based on the viewing angle determine the focal length. The first scenario
makes sense when we have a single lens available, we set the focal distance and determine how many pictures can be drawn to the fov raised.
The second scenario applies when the target is a photograph of a certain resolution (megapixels) and then for the resolution sought and the fov is fixed lens determined that we should use based on the focal length to give us the number of pictures wanted .
To determine the number of photos, focal length and / or may use a calculator fov of scenarios, there are several and I recommend that of Frank van der Pohl
.
We may also use a table like the one below in which each line represents a focal length, the X axis represents the fov from 30 to 360 degrees and the Y axis gives the number of photos to take. This little board is useful that we can print and take with us and save us the use of the calculator.
Table designed for 20% overlap and camera vertically (portrait)
We then until now:
- The fov or angle of vision of our landscape
- The focal length to use
- The number of photos to take
The next step is to take the pictures, looking generally between 20% and 30% overlap, the planchette accounts presented are based on a 20% overlap, the calculator allows you to vary the overlap.
If the scenario requires more than a "row" of photos resulting in a "mosaic" no problem, the vertical FOV can be calculated as the horizontal. Fusion
focus and exposure in many scenarios each of the photos that make up the picture is not one picture but they are several, there are generally two scenarios where the need arises to take over a photo of each portion of the panoramic.
1.
exhibition Fusion
may be that in a high contrast scene one photo fails dynamic range needed for the scene, the sky could be burned or certain parts of the foreground too dark. The use of fill flash is highly dangerous because the light would not be uniform in each portion of the panorama.
In these cases we have to do is to get multiple exposures of each of the panoramic picture, as many as are necessary to achieve the desired exposure.
2. Focus Fusion
may also occur that due to the focal length used part of the panoramic combine elements in the foreground with background elements and the selected aperture can not have both things are properly in focus, in these cases must take two or more photos in which each one has the correct focus different levels we observed in the panoramic.
As conclusion we can have a panoramic view of 4 photos for which we have to take, for example, 12 photos.
the workflow, the workflow! The recommended workflow is to first work to resolve every frame of the panoramic (each stack of photos) and then assemble everything.
To make the exposure and focus blending stacking can be used
tufuse is free software that lets you blend both the focus and exposure. Other options for fusion of focus are CombineZM HeliconFocus and to fuse enfuse exhibitions. Tufuse is the only software that does both.
We have for example: Focus
foto1.jpg deep (exposure for the sky)
foto2.jpg (exposure to buildings)
Focus on the front foto1.jpg (normal exposure)
used to mix all
tufuse.exe -p 2-o fotograma1.tif foto2.jpg foto1.jpg foto3.jpg
The parameter-p 2 indicates two passes to create the fusion a showcase and one for focus. The result is a tif to be part of our scenario in which the focus and exposure are correct for both the foreground and the background.
This process is repeated for all the "batteries" in our scenario, it is necessary that all batteries have the same amount of pictures where there is only 1 plane does stacking lack focus and where there is large exposure differences only takes one exposure, ie the amount of photos for each part of the picture may be heterogeneous.
Assembling the panorama is not the purpose of this article discusses how to assemble the panoramic but I would like to point to the three most recommended programs for the assembly of the panorama that are PTGui, Hugin and PtAssembler. Hugin is free, PTGui is commercial and is the friendliest interface, PtAssembler is the most advanced and is not suitable for beginners for its interface a bit harsh.
PtAssembler If we can directly do the stacking in PtAssembler and you can call or CombineZM tufuse for the stacking, loading the images on the screen 1 PtAssembler can select a group of images and given a stack ID, so we define the stacks to use. Then in Step 5 must be secured in "processing options" that the output is tiff_m and this enabled the option to make stacking. For more information visit the online documentation PtAssembler. Summary
The workflow is then presented as follows.
1. Determine FOV, number of photos and focal length.
2. For each picture:
2.1 Determine number of images taken by outbreak (1 to N)
2.2 Determine number of exposures for each picture (1 to N) 2.3 Remove the photo
3. Move the tripod
4. Return to 2 until the end of the panoramic
5. Merge each cell with tufuse
6. Joining the result with Hugin, PTGui PTAssembler
7. Crop, final adjustments in Gimp or Photoshop Practical Example
We have a scene taken from a balcony where the panorama consists of 3 pictures, because of the contrast and the difference in focus between the rail and the buildings were taken 3 or 4 photos for each portion of the picture as follows:
Stack 1: Left Frame
These three images are combined into one using
tufuse-p 2-o izquierdo.tif izq_1.jpg izq_3.jpg izq_2.jpg
Stack 2: Central Frame
These four images are combined into one using
tufuse-p 2-o central.tif c_2.jpg c_1.jpg c_3.jpg
Stack 3: Frame
law
These four images are combined into one using
tufuse-p 2-o derecho.tif derecho_2.jpg derecho_1.jpg derecho_3.jpg
Once loaded simply stacking izquierdo.tif, cen.tif and derecho.tif in Hugin and PTGui to pan. As we have lost Exif indicate focal length and crop factor of the camera is looking at the Exif easily get any of the original photos.
The end result is:
Como podemos ver tanto la baranda como los edificios estan en foco y usamos F2.8 como apertura para todas las tomas, ademas el cielo esta correctamente expuesto y no quemado al igual que los edificios.
Y por si alguien tiene curiosidad tanto el proceso de apilar primero, stitchear despues con PtGUI o Hugin como el proceso de Stitchear primero y apilar despues de PtAssembler producen resultados identicos.
