Kdenlive/Manual/Clip Menu/Stabilize: Difference between revisions

From KDE Wiki Sandbox
m (typo in kdenlive fixed, change formatting)
(Marked this version for translation)
Line 10: Line 10:
This feature applies image stabilization algorithms to the clip which can reduce the shakiness of a shaky bit of footage.
This feature applies image stabilization algorithms to the clip which can reduce the shakiness of a shaky bit of footage.


=== stabilize ( vstab) ===
=== stabilize ( vstab) === <!--T:9-->


<!--T:4-->
<!--T:4-->
Line 21: Line 21:
The "shutter angle" setting that the narrator says he does not know what it does (1:35) ... Shutter angle is an angle that images could be maximum rotated for stabilization. see [http://www.mltframework.org/bin/view/MLT/FilterVideostab here].
The "shutter angle" setting that the narrator says he does not know what it does (1:35) ... Shutter angle is an angle that images could be maximum rotated for stabilization. see [http://www.mltframework.org/bin/view/MLT/FilterVideostab here].


=== stabilize (transcode) ===
=== stabilize (transcode) === <!--T:10-->


<!--T:11-->
[[File:Kdenlive stabilise transcode.png]]
[[File:Kdenlive stabilise transcode.png]]


<!--T:12-->
Based on the tooltips from this screen and the docs [http://public.hronopik.de/vid.stab/features.php?lang=en here] this what all the options mean:
Based on the tooltips from this screen and the docs [http://public.hronopik.de/vid.stab/features.php?lang=en here] this what all the options mean:


<!--T:13-->
'''optzoom''' = use optimal zoom  (calculated from transforms). Causes video to zoom until 90% of transformations are hidden.  default - checked.
'''optzoom''' = use optimal zoom  (calculated from transforms). Causes video to zoom until 90% of transformations are hidden.  default - checked.


<!--T:14-->
'''accuracy''' = accuracy of shakiness detection. Should be  >=shakiness factor. 1: low (fast processing) 15: high (slow processing)  - default 4
'''accuracy''' = accuracy of shakiness detection. Should be  >=shakiness factor. 1: low (fast processing) 15: high (slow processing)  - default 4


<!--T:15-->
'''algo''' (= algorithm used)  unchecked =brute force. checked= small measurement fields - default = checked
'''algo''' (= algorithm used)  unchecked =brute force. checked= small measurement fields - default = checked


<!--T:16-->
'''maxangle''' = maximum angle to rotate - in radians. default = -1  which means no limit
'''maxangle''' = maximum angle to rotate - in radians. default = -1  which means no limit


<!--T:17-->
'''relative'''  unchecked = absolute transform, checked = relative transform. Default is checked.  
'''relative'''  unchecked = absolute transform, checked = relative transform. Default is checked.  


<!--T:18-->
'''zoom''' - Additional zoom during transform. Percentage to zoom > 0 = zoom in, <0 = zoom out.  The zoom specified here is additional to the optimum zoom calculated by the program when optzoom is checked. default = 0
'''zoom''' - Additional zoom during transform. Percentage to zoom > 0 = zoom in, <0 = zoom out.  The zoom specified here is additional to the optimum zoom calculated by the program when optzoom is checked. default = 0


<!--T:19-->
'''shakiness''' - how shaky is the video. and how quick is the camera? 1: little (fast processing) 10: very strong/quick (slow processing) Default = 4
'''shakiness''' - how shaky is the video. and how quick is the camera? 1: little (fast processing) 10: very strong/quick (slow processing) Default = 4


<!--T:20-->
'''invert''' - invert the transform. default = unchecked
'''invert''' - invert the transform. default = unchecked


<!--T:21-->
'''sharpen''' - sharpen transformed image.  amount of sharpening: 0: no sharpening. uses filter unsharp with 5x5 matrix Default = 0.8
'''sharpen''' - sharpen transformed image.  amount of sharpening: 0: no sharpening. uses filter unsharp with 5x5 matrix Default = 0.8


<!--T:22-->
'''stepsize''' - step size of search process, region around minimum  is scanned with 1 pixel resolution- default = 6
'''stepsize''' - step size of search process, region around minimum  is scanned with 1 pixel resolution- default = 6


<!--T:23-->
'''show''' 0= draw nothing, 1 or 2 = show fields and transforms.  Use 1 or 2 to preview what the process is going to do.  [http://public.hronopik.de/vid.stab/files/skiing_veryshaky_visualized8_short.ogv example] default = 0
'''show''' 0= draw nothing, 1 or 2 = show fields and transforms.  Use 1 or 2 to preview what the process is going to do.  [http://public.hronopik.de/vid.stab/files/skiing_veryshaky_visualized8_short.ogv example] default = 0


<!--T:24-->
'''mincontrast''' below this contrast - the field is discarded.  Range 0-1 default =0.3
'''mincontrast''' below this contrast - the field is discarded.  Range 0-1 default =0.3


<!--T:25-->
'''smoothing''' = number of frames for lowpass filtering.  Used to calculate number of frames to be used for lowpass filtering  used for stabilizing. Number of frames = smoothing * 2 + 1.  default =10. A larger smoothing means that we transform away more camera movements. The resulting clip has a lower change in camera speed. Demo of the difference  [http://public.hronopik.de/vid.stab/files/skiing_veryshaky_short_vs_longsmoothing_above.ogv here]
'''smoothing''' = number of frames for lowpass filtering.  Used to calculate number of frames to be used for lowpass filtering  used for stabilizing. Number of frames = smoothing * 2 + 1.  default =10. A larger smoothing means that we transform away more camera movements. The resulting clip has a lower change in camera speed. Demo of the difference  [http://public.hronopik.de/vid.stab/files/skiing_veryshaky_short_vs_longsmoothing_above.ogv here]


<!--T:26-->
'''maxshift''' maximal number of pixels to translate image.    (default = -1 no limit)
'''maxshift''' maximal number of pixels to translate image.    (default = -1 no limit)


<!--T:27-->
'''crop''' unchecked = The border of the transformed frames contains the pixels from previous frames. Checked = black background. default = unchecked
'''crop''' unchecked = The border of the transformed frames contains the pixels from previous frames. Checked = black background. default = unchecked




<!--T:28-->
This video shows a side by side comparison between vstab and transcode stabilize methods.
This video shows a side by side comparison between vstab and transcode stabilize methods.


<!--T:29-->
{{#ev:youtube|86SUhodScDE}}
{{#ev:youtube|86SUhodScDE}}



Revision as of 11:03, 7 January 2013

Other languages:

Stabilize

This menu item is available from right click on a clip in the project tree or under the project menu when a clip is selected in the project tree.

This feature applies image stabilization algorithms to the clip which can reduce the shakiness of a shaky bit of footage.

stabilize ( vstab)

This tutorial shows how to use the stabilize feature - vstab version:

{{#ev:youtube|SWCoqZLJNgY}}

The "shutter angle" setting that the narrator says he does not know what it does (1:35) ... Shutter angle is an angle that images could be maximum rotated for stabilization. see here.

stabilize (transcode)

Based on the tooltips from this screen and the docs here this what all the options mean:

optzoom = use optimal zoom (calculated from transforms). Causes video to zoom until 90% of transformations are hidden. default - checked.

accuracy = accuracy of shakiness detection. Should be >=shakiness factor. 1: low (fast processing) 15: high (slow processing) - default 4

algo (= algorithm used) unchecked =brute force. checked= small measurement fields - default = checked

maxangle = maximum angle to rotate - in radians. default = -1 which means no limit

relative unchecked = absolute transform, checked = relative transform. Default is checked.

zoom - Additional zoom during transform. Percentage to zoom > 0 = zoom in, <0 = zoom out. The zoom specified here is additional to the optimum zoom calculated by the program when optzoom is checked. default = 0

shakiness - how shaky is the video. and how quick is the camera? 1: little (fast processing) 10: very strong/quick (slow processing) Default = 4

invert - invert the transform. default = unchecked

sharpen - sharpen transformed image. amount of sharpening: 0: no sharpening. uses filter unsharp with 5x5 matrix Default = 0.8

stepsize - step size of search process, region around minimum is scanned with 1 pixel resolution- default = 6

show 0= draw nothing, 1 or 2 = show fields and transforms. Use 1 or 2 to preview what the process is going to do. example default = 0

mincontrast below this contrast - the field is discarded. Range 0-1 default =0.3

smoothing = number of frames for lowpass filtering. Used to calculate number of frames to be used for lowpass filtering used for stabilizing. Number of frames = smoothing * 2 + 1. default =10. A larger smoothing means that we transform away more camera movements. The resulting clip has a lower change in camera speed. Demo of the difference here

maxshift maximal number of pixels to translate image. (default = -1 no limit)

crop unchecked = The border of the transformed frames contains the pixels from previous frames. Checked = black background. default = unchecked


This video shows a side by side comparison between vstab and transcode stabilize methods.

{{#ev:youtube|86SUhodScDE}}