Perspective, correction in two steps
Wide-angle lenses make the lines converge in low angle. The software corrections correct this, but crop strongly. This problem can be avoided in two shots.
Software correction of the verticals of a low angle view requires a strong cropping of the image. By mixing two views, one shot parallel to the vertical lines, the other in low angle, the correction of perspective by software allows keeping the initial framing.
In the Absence of a Shift Lens
Wide angle and architecture lovers, if you don’t have a shift lens, you need to correct converging lines in post-production. DXO ViewPoint, PhotoLab, Lightroom or Photoshop, etc. are efficient. But after correction, the image is cropped with a loss of pixels and the angle of view is narrowed. The rectification of lines dilutes the details. But these problems can be solved by taking the picture in two steps. The tripod is essential.
Two Shots
The first shot is made with the camera parallel to the subject. The lines do not converge. A second shot is taken with the same focal length, moving the camera until it frames the upper part of the subject. A longer focal length can be used as long as the frame includes all the elements of the image, both in width and height. A rotation of the lens on its nodal point would be ideal, but Photoshop can cope without it.
Start With the Parallel View
The plumb image will serve as a base. In Photoshop, it will be completed with the missing part of the low-angle image using Photomerge or an automatic layer alignment. The plumb image does not pose any particular processing problem apart from the correction of geometric aberrations and the homogenization of the two views in terms of density, contrast and color.
Super Resolution
The low-angle image needs a perspective correction. But before that, in order to optimize the preservation of details in the final image, we take advantage of Lightroom’s super resolution. A right-click on the file gives access to Enhance>Super Resolution. The image size is doubled while preserving a maximum of details of the original. This interpolation is useful, because the software correction of the perspective alters the size of the image.
Perspective correction of receding lines from the low-angle image reduces the size of the usable area of the image. Lightroom offers a super-resolution mode (Photo > Sharpen…) which successfully doubles the image definition. One will not lose details for the continuation of the treatment.
Perspective Correction
The second step corrects the perspective in Lightroom (DXO or Photoshop, etc.), automatically; or manually with alignment guides. Next, the image is cropped, keeping only the top half of the image to optimize the merge with the perfectly vertical image. Then the rectified image is resampled to the same width as the base image (in Photoshop, Image>Image size).
First check that the adjustment of geometric aberrations delivers straight lines. As an alternative to Lightroom, DXO lens profiles are often more efficient. The correction is then done in automatic mode or with marks to control the parallelism.
To avoid bad connections with the bottom of the image in the final assembly, we crop. In width, the cut is aligned with the base image; in height, we keep half. Then we go to Photoshop. The size (in pixels) of the image is then adjusted to that of the base image.
Photomerge
Photoshop will stitch the two images together. Two options work well. The first is to stack them in Photomerge (File>Automate>Photomerge) in automatic mode. It does the job, but the image may be cropped on the sides. The other option, which does not crop the picture, consists in duplicating the background image to create a new layer. The corrected image is then pasted on top of the background copy. Then align the two layers (Edit>Auto Align Layers…) in automatic mode. The alignment may require the addition of a layer mask and the use of a brush on areas showing misalignment. The image quality obtained with this perspective correction mode is not to be ashamed to compare with a shift lens.
The background of the base image is duplicated (Cmd + J on Mac, Ctrl + J on PC) to create a layer identical to the background. This is necessary for the future alignment with the rectified and cropped image that will be pasted on top.
The background must be extended upwards by at least 150% to make the collage of the rectified low-angle image (Image > Canvas size).
The top of the image is added to the top layer by copying and pasting or by dragging and dropping. If necessary, you can adjust the size of the layer (Edit > Manual transformation) to better match the size of the base image. Then select the two layers.
The layers are aligned (Edit > Automatic layer alignment) in automatic mode. If there are imperfect connections, the addition of blending masks on the layers allows you to refine the transitions by masking the perfectible areas with a brush. Then the image is cropped to the desired dimensions.
Discover the photography courses at Spéos
Spéos offers various training courses ranging from simple one-week photography courses (initiation and advanced level) to 3-year courses. The long courses to become professional photographers allow you not only to master all the photographic techniques and its vocabulary (blurs, hyperfocus, sharpness zone, depth of field, backlighting, focal length, shutter release, autofocus, wide-angle, rule of thirds, etc.), but also all the stages of shooting and image processing.
Visiting the school allows you to discover the premises, the studios and the equipment, and is undoubtedly the best way to familiarize yourself with your future way of working. This is why, in addition to the open days, Spéos offers throughout the year personalized visits by appointment to come and discover the school with a member of the team.
Text and photos: Philippe Bachelier, teacher of Printing techniques at Spéos
