68
ANDREAS PLESCH
real objective, even a highly corrected Plan-Apochromat, will exhibit some
residual optical aberrations. (2) Light scattering will occur within the
specimen. (3) Particularly for small, faint signals, focusing may be difficult. Each factor will contribute to signal broadening. The real optical resolution limit is about 50% below the diffraction limited theoretical resolution.
Figure 9 shows the results of two different CCD cameras (Fig. 9a,c:
standard CCD M300 with 11 /lm pixels; Fig. 9b,d: mega pixel CCD M1
with 6.7 /lm pixels) and two different objectives (Fig. 9a,b: Plan-Neofluar
100x/1.3 Oil; Fig. 9c,d: Plan-Apochromat 63x/1.4 Oil). The raw images
were digitally zoomed to approximately the same size (upper row of
Fig. 9). Zooming was performed without interpolation in order to recognize the individual pixels. Intensity profiles of the signals are displayed for
regions A (middle row of Fig. 9) and B (bottom row of Fig. 9), respectively.
For better comparison of their shape the profiles have been normalized.
Consequently, they range from 0 (which corresponds to the fluorescence
background intensity outside the spots) to 100 (which corresponds to the
maximum intensity of the spots). At 100x magnification (Fig. 9a,b) both
cameras resolve the two signals of region A (middle row of Fig. 9a,b),
although the mega pixel CCD provides slightly better contrast. This is reflected by the more pronounced dip of the profIle. Due to the poor signal
intensity of one of the two signals of region B (bottom row of Fig. 9a,b) the
darker region separating the signals is at the very limit of visibility.
With the 63 x Plan-Apochromat and the standard CCD, the centers of
the signals of region A are approximately 2.5 pixels apart. While the signals
should be resolved if they were of approximately equal intensity, they appear as one larger signal (middle row of Fig. 9c) due to the low intensity of
one of the spots with only 40% of the intensity of the second spot. The
fluorescence background further reduces the contrast. This situation corresponds roughly to the simulation of Fig. 10. The mega pixel CCD resolves the spots of region A (middle row of Fig. 9d), while the signals
of region B are again at the limit of resolution due to their low and uneven
intensity (bottom row of Fig. 9c,d).
Although intensity profiles are helpful to appreciate subtle differences
of signal intensities and very valuable to explain theory and compare theoretical to real results, they do not exactly correlate to the visual interpretation of the image. The information content of the image as a two-dimensional scene is higher than a one-dimensional profile. The human visual
system performs a lot of image analysis and pattern recognition that the
human observer is not aware of. For example, the presence of two spots is
quite evident from the morphology of the signals in region A. From their
ANDREAS PLESCH
real objective, even a highly corrected Plan-Apochromat, will exhibit some
residual optical aberrations. (2) Light scattering will occur within the
specimen. (3) Particularly for small, faint signals, focusing may be difficult. Each factor will contribute to signal broadening. The real optical resolution limit is about 50% below the diffraction limited theoretical resolution.
Figure 9 shows the results of two different CCD cameras (Fig. 9a,c:
standard CCD M300 with 11 /lm pixels; Fig. 9b,d: mega pixel CCD M1
with 6.7 /lm pixels) and two different objectives (Fig. 9a,b: Plan-Neofluar
100x/1.3 Oil; Fig. 9c,d: Plan-Apochromat 63x/1.4 Oil). The raw images
were digitally zoomed to approximately the same size (upper row of
Fig. 9). Zooming was performed without interpolation in order to recognize the individual pixels. Intensity profiles of the signals are displayed for
regions A (middle row of Fig. 9) and B (bottom row of Fig. 9), respectively.
For better comparison of their shape the profiles have been normalized.
Consequently, they range from 0 (which corresponds to the fluorescence
background intensity outside the spots) to 100 (which corresponds to the
maximum intensity of the spots). At 100x magnification (Fig. 9a,b) both
cameras resolve the two signals of region A (middle row of Fig. 9a,b),
although the mega pixel CCD provides slightly better contrast. This is reflected by the more pronounced dip of the profIle. Due to the poor signal
intensity of one of the two signals of region B (bottom row of Fig. 9a,b) the
darker region separating the signals is at the very limit of visibility.
With the 63 x Plan-Apochromat and the standard CCD, the centers of
the signals of region A are approximately 2.5 pixels apart. While the signals
should be resolved if they were of approximately equal intensity, they appear as one larger signal (middle row of Fig. 9c) due to the low intensity of
one of the spots with only 40% of the intensity of the second spot. The
fluorescence background further reduces the contrast. This situation corresponds roughly to the simulation of Fig. 10. The mega pixel CCD resolves the spots of region A (middle row of Fig. 9d), while the signals
of region B are again at the limit of resolution due to their low and uneven
intensity (bottom row of Fig. 9c,d).
Although intensity profiles are helpful to appreciate subtle differences
of signal intensities and very valuable to explain theory and compare theoretical to real results, they do not exactly correlate to the visual interpretation of the image. The information content of the image as a two-dimensional scene is higher than a one-dimensional profile. The human visual
system performs a lot of image analysis and pattern recognition that the
human observer is not aware of. For example, the presence of two spots is
quite evident from the morphology of the signals in region A. From their