This is the interface for the LEAD Video Stabilizer Filter.
typedef [v1_enum] enum lmvstab_enumClippingOption
{
lmvstab_CLIPPINGOPTIONS_OUTLINE,
lmvstab_CLIPPINGOPTIONS_CROP,
lmvstab_CLIPPINGOPTIONS_CROPANDSTRETCH
} lmvstab_enumClippingOption ;
Describes possible cropping options.
Constant |
Description |
lmvstab_CLIPPINGOPTIONS_OUTLINE |
Frame the resulting video with bands of the specified color. The video size remains the same as that of the original. |
lmvstab_CLIPPINGOPTIONS_CROP |
Crop the video. Video size is smaller than that of the original. |
lmvstab_CLIPPINGOPTIONS_CROPANDSTRETCH |
Crop the video but then resize it so it is the same size as the original. |
typedef [v1_enum] enum lmvstab_GlobalEstimationCoefficients
{
lmvstab_GLOBALESTIMATIONCOEFFICIENTS_X_Y,
lmvstab_GLOBALESTIMATIONCOEFFICIENTS_X_Y_THETA
} lmvstab_GlobalEstimationCoefficients ;
Describes the possible stabilization options.
Constant |
Description |
lmvstab_GLOBALESTIMATIONCOEFFICIENTS_X_Y |
Fix the shaking that occurs along the x and y axes. |
lmvstab_GLOBALESTIMATIONCOEFFICIENTS_X_Y_THETA |
Fix the shaking that occurs along both axes as well as fix shaking from rotation. |
typedef [v1_enum] enum lmvstab_GlobalEstimationQuality
{
lmvstab_GLOBALESTIMATIONQUALITY_1,
lmvstab_GLOBALESTIMATIONQUALITY_2,
lmvstab_GLOBALESTIMATIONQUALITY_3,
lmvstab_GLOBALESTIMATIONQUALITY_4
} lmvstab_GlobalEstimationQuality ;
Describes quality options
Constant |
Description |
lmvstab_GLOBALESTIMATIONQUALITY_1 |
Highest quality but slow processing |
lmvstab_GLOBALESTIMATIONQUALITY_2 |
More performance, lower quality |
lmvstab_GLOBALESTIMATIONQUALITY_3 |
Slower performance, better quality |
lmvstab_GLOBALESTIMATIONQUALITY_4 |
Fastest but worst quality |
typedef [v1_enum] enum lmvstab_PathSmoothingLevel
{
lmvstab_PATHSMOOTHINGLEVEL_1 = 4,
lmvstab_PATHSMOOTHINGLEVEL_2 = 8,
lmvstab_PATHSMOOTHINGLEVEL_3 = 12,
lmvstab_PATHSMOOTHINGLEVEL_4 = 16,
lmvstab_PATHSMOOTHINGLEVEL_5 = 24,
lmvstab_PATHSMOOTHINGLEVEL_6 = 32,
lmvstab_PATHSMOOTHINGLEVEL_7 = 64,
lmvstab_PATHSMOOTHINGLEVEL_8 = 128,
lmvstab_PATHSMOOTHINGLEVEL_COUNT = 8
} lmvstab_PathSmoothingLevel ;
Describes the amount of path smoothing. Many factors determine which amount is best for a particular video, including:
Input video frames per second. Higher frame rates need to use higher levels (lmvstab_PATHSMOOTHINGLEVEL_8).
Whether a stationary camera was used. If so, use a higher level value (8).
Whether a moving camera was used. If so, use levels (6-4).
Whether you want to minimize the amount of clipping. If so, use the lower levels (1-4).
typedef [v1_enum] enum lmvstab_InterpolationType
{
lmvstab_INTERPOLATIONTYPE_NEAREST,
lmvstab_INTERPOLATIONTYPE_BILINEAR
} lmvstab_InterpolationType ;
Describes possible interpolation options when shifting and rotating input frames to get stabilized output. Two interpolation methods are available, as follows:
Constant |
Description |
lmvstab_INTERPOLATIONTYPE_NEAREST |
Simple and fast interpolation |
lmvstab_INTERPOLATIONTYPE_BILINEAR |
Better quality interpolation but slower |
Type |
Name |
Description |
VARIANT_BOOL |
Enabled |
Enables or disables video stabilization. If set to VARIANT_FALSE, the filter passes the video samples intact. |
VARIANT_BOOL |
AutoReconnect |
If set to VARIANT_TRUE, the filter can apply changes that affect media types while the graph is running (without having to stop and manually reconnect).This includes options such as clipping. |
VARIANT_BOOL |
ROIEnabled |
If set to VARIANT_TRUE, the filter processes only the region of interest. Otherwise, the entire frame gets processed. The region of interest restricts the area considered when reducing shaking, but the reductions are applied to the entire frame. Limiting the filter to a portion of the video increases performance, allowing the filter to be run in real time. |
long |
ROITop |
Top coordinate, in pixels, of the region of interest |
long |
ROILeft |
Left coordinate, in pixels, of the region of interest |
long |
ROIBottom |
Bottom coordinate, in pixels, of the region of interest |
long |
ROIRight |
Right coordinate, in pixels, of the region of interest |
lmvstab_GlobalEstimationCoefficients |
GlobalEstimationCoefficients |
Gets or sets the lmvstab_GlobalEstimationCoefficients value to be used. This describes the type of stabilization to be performed. |
lmvstab_GlobalEstimationQuality |
GlobalEstimationQuality |
Gets or sets the lmvstab_GlobalEstimationQuality value to be used. This value describes the balance between the speed and the quality of the processing. |
lmvstab_PathSmoothingLevel |
PathSmoothingLevel |
Gets or sets the lmvstab_PathSmoothingLevel values to be used. |
OLE_COLOR |
BkFillColor |
Color to use to fill any empty areas resulting from processing. |
long |
LeftRightCropMargin |
Left and right crop margin |
long |
TopBottomCropMargin |
Top and bottom crop margin |
lmvstab_InterpolationType |
InterpolationType |
Gets and sets the lmvstab_InterpolationType value to be used. |
lmvstab_enumClippingOption |
ClippingOption |
Gets and sets the lmvstab_enumClippingOption value to be used. |
typedef [v1_enum] enum
{
lmvstab_APILEVEL_1 = 0,
lmvstab_APILEVEL_Latest = lmvstab_APILEVEL_1
} lmvstab_APILEVEL ;
Describes possible interpolation options when shifting and rotating input frames to get stabilized output. Two interpolation methods are available, as follows:
Constant |
Description |
lmvstab_APILEVEL_1 |
Uses the defaults from the first version of the filter. |
lmvstab_APILEVEL_Latest |
Uses the defaults from the latest version of the filter. |
HRESULT ResetToDefaultsEx(lmvstab_APILEVEL ApiLevel)
lmvstab_APILEVEL ApiLevel
Call this method to restore default options for all properties.
The ApiLevel determines which default values to use. Current applications should set ApiLevel to lmvstab_APILEVEL_1 = 0.
The filter might change in the future and have different properties or default behaviour. Calling this method ensures the filter will have the same default values as they were at the time you developed your application.
S_OK always.
Notes:
Currently, this filter stabilizes video in the vertical, horizontal, and angular directions. Shake from other directions can cause unexpected results.
For best results, make sure any moving objects inside the video are smaller than 15% of the video frame size.