Korrektur-Parameter

EstimateGeometry (/EG=1) ⇧ Range of values [0-5] Correction according current or earlier error estimation: 0: The correction of the geometric image errors takes place not on the base of the values, which were estimated from the image, but with the constant values UseHeight/~Size/~Rotation/~Vergence_V/~Vergence_V and WindowOffset. 1: The geometric errors are estimated from the image and used for the correction. 2: For the correction of the geometric errors, the entries of the ListofImages will be used. 3: The geometric errors are estimated from the image and used for the correction. With jpg- and tif-images, the correction values are written to the commentary area of the original images (only with license key). These values will be later used, if EstimateGeometry is set equal 4. 4: For the correction of the geometric errors, the entries of the commentary area of the jpg/tif-images are used. If these values are not valid for any reason, a fallback to EstimateGeometry = 3 will take place. Interactively made changes of the stereo window are also be stored. 5: The correction values will be read out from the file ListofImages and written into the commentary area of the original images. The file ListofImages will updated with the new values. -3/-4/-5: It will be read and stored only into the left images (normally both images left and right are used).

ApplyLeftRight (/AL=0) ⇧ Range of values [-1,0,1] Correct left, right or both images symmetrically: 0: The stereo correction will be applied symmetrically to the left and right image. -1: Only the left image will be equalized, the right one keeps untouched. 1: Only the right image will be equalized, the left one keeps untouched.

EstimateVergence_V (/EV=2) ⇧ Range of values [0-2] Activates the correction of vertical vergence error (sensor rotation around vertical axis). Vertical vergence errors result from convergent optical axes. 0: Estimation of vertical vergence error deactivated. 1: Estimation of vertical vergence error activated. 2: Estimation of vertical vergence error only activated, if objects are detected in all 4 image quadrants. Hint: With exactely parallel mounted cameras, there are no vertical vergence errors. In that cases the results can be enhanced for problematic images (low contrast and less content) by deactivating the vergence estimation.

EstimateVergence_H (/EE=2) ⇧ Range of values [0-2] Activates the correction of horizontal vergence error (sensor rotation around horizontal axis). Horizontal vergence errors can occur when using mirror rigs. 0: Estimation of horizontal vergence error deactivated. 1: Estimation of horizontal vergence error activated. 2: Estimation of horizontal vergence error only activated, if objects are detected in all 4 image quadrants. Hint: With exactely parallel mounted cameras, there are no horizontal vergence errors. In that cases the results can be enhanced for problematic images (low contrast and less content) by deactivating the vergence estimation.

UseHeight (/UH=0.0) ⇧ Range of values [-1000, 1000] pixel If EstimateGeometry = 0 is set, the value given here is used for the height error correction. A positive value shifts the right image downward and the left image upward.

UseSize (/US=0.0) ⇧ Range of values [-5.0, 5.0] % If EstimateGeometry = 0 is set, the value given here is used for the size correction. A positive value makes the right image smaller and increases the left image.

UseRotation (/UR=0.0) ⇧ Range of values [-5.0, 5.0] degrees If EstimateGeometry = 0 is set, the value given here is used for the correction of the rotation. A positive value rotates the right image in the clockwise direction and the left image just otherwise.

UseVergence_V (/UV=0.0) ⇧ Range of values [-5.0, 5.0] degrees If EstimateGeometry = 0 is set, the value given here is used for the correction of the vertical vergence. The vertical vergence is the angle, over that the shooting axes are tilted against the parallel adjustment a) inwards (convergent shooting axes, vergence angle < 0) or b) outwards (divergent shooting axes, vergent angle > 0). Thus, a postive value of UseVergence_V corrects convergent axes. Note: UseVergence_V is a virtual angle on the assumption of a standard focal length. The vergence angle actually used with the photo cannot be determined afterwards, but this is not absolutely necessary for the correction.

UseVergence_H (/UE=0.0) ⇧ Range of values [-5.0, 5.0] degrees If EstimateGeometry = 0 is set, the value given here is used for the correction of the horizontal vergence. The horizontal vergence is the angle, over that the shooting axes are tilted against the parallel adjustment upward or downward. That error may exist only for misadustments of two cameras, e.g. when using halfcoated mirrors. However, sometimes it also helps to overcome lens distortion errors. Note: UseVergence is a virtual angle on the assumption of a standard focal length. The vergence angle actually used with the photo cannot be determined afterwards, but this is not absolutely necessary for the correction.

WindowOffset (/WO=5) ⇧ Range of values [-512, 512] Additional offset for the stereo window. Independent of EstimateWindow, an additional shift of the stereo window takes place. A postive value shifts the stereo scenery into the background. WindowOffset may be given in pixel, in promille of the image width or in promille of the projection image. The combobox is named WinOffReference. Attention: WindowOffset is only active when EstimateGeometry = 0 (preset of all correction values) or = 1 or = 3 (estimation of the correction values from the input image).

WinOffReference (/WR=1) ⇧ Range of values [0-5] This parameter sets the unit for WindowOffset. 0: Pixel units 1: Promille of the image width. 2: Promille of the projektion width with an assumed width-to- height ratio of the projektor of 4:3. 3: Promille of the projektion width with an assumed width-to-height ratio of the projektor of 16:10. 4: Promille of the projektion width with an assumed width-to-height ratio of the projektor of 16:9. 5: Promille of the projektion width with an assumed width-to-height ratio of the projektor of 2048:1080. Attention: This parameter WinOffReference is also used for the units of FarPointLimit (value in the box nearby EstimateWindow) and TargetFarPoint (value in the box nearby AutoFrameFloating.

EstimateWindow (/EW=2) ⇧ Range of values [0-5] Method for setting the stereo window: 0: There is no automatic shift of the stereo window. Only for test purposes - the result won't be a very useful stereoscopic image 1: The stereo window will be set about in the middle of the space. Only for test purposes - the result won't be a very useful stereoscopic image. 2: The stereo window will be set into the near point. 3: The stereo window will be set into the near point, isolated objects in the middle of the image will be ignored. 4: The stereo window will be set into the far point. 5: Near point framing, as long as the given far point disparity (in the box nearby) is not exceeded (divergence rule should then be fulfilled). If the deviation is larger than this value, the far point distance will be fixed. So, the images can always be projected without pain, but with the cost of a possible violation of the stereo window rule. Attention: After the automatic setting of the stereo window an additional shift takes always place with WindowOffset. The unit for the far shift will be set using the parameter WinOffReference (combo box in the Gui).

AutoOutputContrast (/OC=3) ⇧ Range of values [0-4] Controls the contrast-normalizing and the contrast- and color-balancing with the output image. With the contrast-normalizing the histogram will be streched to it's full range, with the contrast-balancing the histograms of the left and right images are matched according to their shape. With the color-balancing the histograms are matched for all three base-colors individually. 0: Contrastbalancing/normalizing and color-balancing switched off. 1: Contrast-balancing switched on, color-balancing switched off (the histograms of the grey values are matched, no color balancing). 2: Contrast-normalizing switched on, color-balancing switched off (the histograms of the grey values are streched and matched, no color balancing). 3: Color-balancing and color-balancing switched on (the histograms of the 3 base-colors are matched). 4: Color-normalizing and color-balancingswitched on (the histograms of the 3 base-colors are streched and matched). With Cosima, the geometric image correction, the adjustment of color and contrast as well as the scaling to the final image size will be done in one single interpolation step. This guarantees minimal image losses and the best reachable quality! With Cosima no gaps result in the histogram from the contrast-normalization!

AutoGammaAdjust (/AG=0) ⇧ Range of values [0-3] Gamma correction. The gravity center of the histogram will be shifted using an exponential function (only one degree of freedom). 0: No change of the image. 1: The gravity center of the histogram is shifted a bit toward the middle (recommended for prints). 2: The gravity center of the histogram is remarkebly shifted towards the middle. 3: The gravity center of the histogram is shifted excately to the middle. As opposed to AutoOuputContrast, here only a simple gamma correction takes place using one single parameter. This method is improper to balance out between left and right image and is therefore adopted only commonly to both images.

AutoGhostBust (/AB=0) ⇧ Range of values [0-30] Suppression of ghosting images. The parameter value is the precompensation intensity of the projection crosstalk between the left and the right image (in percent). Common values are 4% or 5%. If the value is set other than 0, a test image cosima_alignment.bmp will be created. This test image includes precompensation values from 0% to 9%. The optimal value can be found using this test image. The test image will be created in the PathofImages automatically with the image size OutputHeight x OutputWidth.

ColorLeftRight (/CL=0) ⇧ Range of values -1,0,1] Defines the reference image for the contrast- and colorcorrection: -1: Applies contrast- and colorcorrection only to the left image. 0: Applies contrast- and colorcorrection symmetrically to both images. 1: Applies contrast- and colorcorrection only to the right image.

ExtColorManagement (/EC=0) ⇧ Range of values [0,1,2] External color management. This activates a color adjustment, defined by an external control file. This color adjustment is separated into three steps: The 1.st step is a gamma correction, the 2.nd step is a linear weighting of the RGB values using a 3x3 matrix and the 3.rd step is another gamma correction with an inverse gamma factor. This all will be done differently for the left and the right images. 0: Without external color adjustment. 1: External color adjustment using a file with the format of .\helpers\cosima_colors_1.txt. This file provides the gamma correction factors as values. 2: External color adjustment using a file with the format of .\helpers\cosima_colors_2.txt. This file provides the gamma evaluation function in the form of a 12 Bit table. E.g. with the application of a userdefined color profile, the color shifts of the Infitec method can be compensated.

ECM_FileName (/EF=cosima_colors.txt) ⇧ Range of values: characters The name of the control file for the external color adjustment. Details are given with the two example files .\helpers\cosima_colors_1.txt and .\helpers\cosima_colors_2.txt. The file must be located in the path PathofImages.