MTF Data

Geometric MTF average of sagittal and tangential response. The parameters are:

Samp: Higher sampling yields higher accuracy at the expense of computation time. To confirm the computation has acceptable accuracy, start at 1, and increment the sampling until the results change by less than the desired accuracy. Note that extreme precision is not required for good optimization results; three significant figures is usually adequate.

Wave: The wavelength number to use (use 0 for polychromatic).

Field: The field number.

Freq: The spatial frequency in MTF units (see "MTF Units").

!Scl: If zero, then the diffraction limit will be used to scale the results (recommended) otherwise, no scaling is done.

If Grid is zero, a fast, sparse sampling integration method is used to compute the MTF. The fast Geometric MTF algorithm is only accurate for systems with circular or elliptical pupils with modest or no apodization. For systems that violate this assumption, set Grid equal to 1. The fast sampling method used by GMTA, GMTS, and GMTT is not directly related to the Geometric MTF Analysis feature. Because only a single spatial frequency is required, the method of computation used by the MTF operands is different, and generally much faster, than the algorithm used by the analysis feature. To select the alternate grid-based algorithm used by the Geometric MTF analysis feature, set Grid equal to 1. The grid-based algorithm is usually slower than the default algorithm if the MTF is reasonably good (greater than 5%), but the grid algorithm converges faster if the aberrations are very large and the resulting MTF is very low.

If both the tangential and sagittal MTF are needed; place the GMTT and GMTS operands on adjacent lines and they will be computed simultaneously. The Geometric MTF, though approximate, can usually be computed more quickly than the Diffraction MTF, and is therefore useful for optimization. See "Performing an optimization".

Moore-Elliott Contrast, average of sagittal and tangential. This operand uses the Moore-Elliott Contrast method to optimize the MTF at a given spatial frequency. See "Optimizing for MTF" for more information on the Moore-Elliott Contrast method. The parameters are:

Wave: The wavelength number to use.

Field: The field number.

Freq: The spatial frequency in cycles per millimeter in focal systems and cycles per afocal unit (see "Afocal Mode Units") in afocal systems.

See "Hx, Hy, Px, and Py". Also see related operands MECS and MECT.

Diffraction modulation transfer function, average of sagittal and tangential. The parameters are: Samp: Higher sampling yields higher accuracy at the expense of computation time. To confirm the computation has acceptable accuracy, start at 1, and increment the sampling until the results change by less than the desired accuracy. Note that extreme precision is not required for good optimization results; three significant figures is usually adequate. There are two algorithms available for computing the MTF. If Grid is zero (recommended), a fast, sparse sampling integration method is used to compute the MTF. The fast sampling method used by MTFA, MTFS, and MTFT is not directly related to the MTF Analysis feature. Because only a single spatial frequency is required, the method of computation used by the MTF operands is different, and generally much faster, than the algorithm used by the analysis feature. To select the alternate grid-based algorithm used by the MTF analysis feature, set Grid equal to 1. The grid-based algorithm is usually slower than the default algorithm if the MTF is reasonably good (greater than 5%), but the grid algorithm converges faster if the aberrations are very large and the resulting MTF is very low.

Wave: The wavelength number to use (use 0 for polychromatic).

Field: The field number. To extract the diffraction-limited MTF use field equal 0 and Grid parameter equal 1. Similar options available for for the MTFS, MTFT, MTFN, and MTFX operands.

Freq: The spatial frequency in MTF units (see "MTF Units"). If the sampling is set too low for accurate computation of the MTF, then the MTF operands all return zero.

If both the tangential and sagittal MTF are needed; place the MTFT and MTFS operands on adjacent lines and they will be computed simultaneously. See "Performing an optimization".

Data Type: Specifies the data to be returned. An input of 0 will return the modulation amplitude; an input of 1 will return the real part; an input of 2 will return the imaginary part; an input of 3 will return the phase in degree. This input is only available for the MTFA, MTFS, and MTFT operands (and not the equivalent square-wave operands).

Discussion:

Diffraction limited MTF Tangential (MTFT) and sagittal (MTFS) are same for rotationally symmetric system. MTFT operands value can be obtained by setting Field=0 and Grid=1 at single Spatial Frequency (Freq) in Merit Function Editor. For non-symmetrically rotated system MTF Tangential (MTFT) and sagittal (MTFS) could be different at single Spatial Frequency (Freq).

Huygens Modulation transfer function, average of sagittal and tangential. This operand computes the diffraction MTF using the Huygens method (see "Huygens MTF"). The parameters are:

Samp: The pupil sampling, where 1 yields 32 x 32, 2 yields 64 x 64 etc. The sampling is assumed to be the same for both pupil and image.

Wave: The wavelength number to use (use 0 for polychromatic).

Field: The field number.

Freq: The spatial frequency in MTF units (see "MTF Units"). If the sampling is set too low for accurate computation of the MTF, then the MTF operands all return zero.

Pol?: Set to 0 to ignore polarization and 1 to consider it.

All Conf?: Set to 0 to use the current configuration (defined by the last CONF operand preceding this operand), and 1 to sum over all configurations. See "Huygens MTF" for a full discussion of this option.

Ima Delta: The image delta in micrometers used for the computation. If zero, the default image delta is used.

If both the tangential and sagittal MTF are needed; place the MTHT and MTHS operands on adjacent lines and they will be computed simultaneously. See "Performing an optimization".