[Kdenlive-devel] My experience with Blender

[Florian Cramer]

Quick footnote:

> - Cinelerra: Editing simply doesn't work, crashes and hangs frequently,
> looks abandoned.

Actually, it is not abandoned, and it might be just your perfect tool,
among others, for the Ken Burns effect. There are just a few caveats:

- Have you made sure that you installed the community version,
  http://cv.cinelerra.org ("Cinelerra CV" - some people wrongly think it
  stands for an unstable CVS version)? The community version is
  sufficiently stable (crashes occasionally, but less so than kdenlive)
  and has a cleaned-up interface.

- Have you read the Cinelerra CV manual? It is actually a very good
  documentation: http://cv.cinelerra.org/docs.php . To get started,
  read Rob Fisher's quick Cinelerra tutorial:
  http://www.robfisher.net/video/cinelerra1.html  
  In short, it is essential to have understood the concepts of
  armed/disarmed tracks, copy-paste vs. drag-and-drop editing, the
  difference between previewer and compositor, the function of gang
  faders and key frames to use the program.  It has a high initial
  learning curve (similar, although not quite as high, to vi or Emacs),
  but is actually an efficient and powerful editing tool.

- Cinelerra's strength for something like the Ken Burns effects lies in
  its fully integrated compositing, particularly in the concept of the
  "virtual camera" and the "virtual projector" in conjunction with key
  framing 
  <http://cv.cinelerra.org/docs/split_manual_en/cinelerra_cv_manual_en_8.html#SEC129>.
  You can drop an image to the editing timeline, adjust its handles for
  the desired running time, set a key frame at the beginning and then go
  into the virtual camera settings of the compositor window, set a zoom
  factor and the part of the image you want to be visible; then you set
  a second key frame at the end, and adjust the camera settings a second
  time (for example, with the same zoom factor, but with a different
  visible part of the image). As a result, Cinelerra creates a smooth
  virtual camera pan (and, possibly zoom) between the two selected parts
  of the image. By setting additional key frames, you can create even
  more complex pans/zooms in the image. 

Here is a sample video I created in five minutes using the image
http://www.dcresource.com/reviews/nikon/d300-review/nightshot.jpg,
Cinelerra for editing/compositing, and ffmpeg for encoding from yuv4mpeg
to mpeg4/h.264: 
http://data.plaintext.cc/cityscape.mp4

Florian

-- 
http://cramer.plaintext.cc:70
gopher://cramer.plaintext.cc