This panel contains options for controlling montage acquisition. It opens when montaging is initiated and closes when a montaged file is closed.
Start
The button will start a montage acquisition.
Prescan
This button will start the relatively fast acquisition of binned down images that are used just to compose an overview of the region being montaged.
Current Z
The value shown in the text box is the Z or section number that will be used for the next montage acquisition. You can enter a new number or use the up or down arrows to increase or decrease the value by one. If you set the value to a section number that has already been acquired, you will overwrite the stored data for that section, which the program will ask you to confirm. If you increase the value you may leave a gap in Z values.
Bin Overview
Ordinarily, an overview image will be subsampled so as to make the overview size comparable to the size of a single frame image. This subsampling is referred to (inaccurately) as binning. You can adjust this binning to a lower number to see more detail when you zoom up the overview image. When the Navigator sets up a montage, the overview binning will be set to 1, under the assumption that you will want to see all the detail in a map montage.
Bin Prescan
By default, a prescan montage is taken at as high a binning as possible to take the least amount of time. You can set this binning to a lower value if you want to see more detail in the prescan. Note that the overview binning is always 1 for a prescan montage, so you need not adjust that binning as well.
Correct drift
If this option is selected, the program will use the misregistration between a pair of successive pieces to predict and correct for the error in position of the next acquired piece. When there is significant drift, this option can reduce the mean error before shifting pieces into registration, and it can prevent very large errors between successive columns of pieces.
Change focus with height
If this option is selected, focus will be changed for every piece that is acquired from a tilted specimen. This focus change will compensate for the differences in Z height at the centers of the pieces. If you are processing the data in IMOD and using corrections ofr magnification gradients, be sure to check the option that focus was adjusted between pieces on the eTomo setup page.
Show overview at end
If this option is selected, the overview image in buffer B will be shown at the end of every montage acquisition.
Align pieces in overview
If this option is selected, each piece will be shifted into alignment with preceding pieces as it is added to the overview image, based on the shifts in the overlap zones that it shares with preceding pieces. This may give a seamless overview, or it may not if there are errors in measuring shifts in the overlap zones. When a montage is read from a file, the pieces are read in from the center outward so that interior pieces will not be misaligned by errors in the shifts between pieces at the edges of the montage.
Treat as very sloppy montage
If this option is selected, SerialEM will analyze the overlap zones with the same parameters that are invoked by default with the VerySloppyMontage option in Blendmont. The large correlation areas and other parameter settings will give much better results with montages where the errors are large and where there may be strong edge effects from the edge of a cell or a grid bar. In addition, correlation coefficients will be computed in real space at the shifts corresponding to the 16 highest peaks in the cross-correlation, thus eliminating the edge effects that contaminate Fourier-based correlation. The option is turned on or off depending on whether you choose to montage by moving the stage in the Montage Setup dialog box , but you can then toggle it as desired when reading in a montage from a file. This option is not the default for two reasons: there is significantly more computation time with the larger correlations, which will be noticeable when reading from a file if previously computed shifts are not coming from an mdoc file; and if there are problems in getting the right correlation, the greater shifts allowed can give much bigger piece displacements and a much worse-looking overview.
Assess multiple peaks with weights
If this option is selected, correlation coefficients will be computed in real space at the shifts corresponding to the 16 highest peaks in the cross-correlation, The correlations will be weighted by the deviation from the expected peak location. This weighting will help the program find the true shift when there are periodic structures (holes or grid squares), but it may be counterproductive in other cases. In general, then, this setting should be appropriate for montages with cryo-specimens on grids with regularly spaced holes, unless the stage calibration is bad enough that the error is comparable to the spacing between periodic structures. Make sure there is a high-defocus magnification calibration to prevent this problem in grid square maps.
This option will be disabled when a montage satisfies the size criterion for applying grid map parameters in the correlations to find the shifts; in that case 50 peaks will be analyzed (or 100 if Treat as very sloppy montage is selected and weighting by distance from expected shift will always be used. See Set Params for Grid Maps in the Camera - Cross-correlation submenu.